daviddoji/Fastbook_Julia
1
0 Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Update and add clean

Browse Source
This commit is contained in:
Sylvain Gugger committed 2020-03-06 10:19:03 -08:00
1 parent e87f1d54e7
commit b2f1c12d4c
196 files changed
+33472 -303
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

01_intro.ipynb
+23 -34
View File
@@ -2,7 +2,7 @@
"cells": [
{
"cell_type": "code",
"execution_count": null,
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
@@ -517,7 +517,7 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 2,
"metadata": {},
"outputs": [
{
@@ -536,9 +536,9 @@
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>0.188673</td>\n",
" <td>0.012107</td>\n",
" <td>0.002706</td>\n",
" <td>0.169390</td>\n",
" <td>0.021388</td>\n",
" <td>0.005413</td>\n",
" <td>00:14</td>\n",
" </tr>\n",
" </tbody>\n",
@@ -567,10 +567,10 @@
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>0.056649</td>\n",
" <td>0.009575</td>\n",
" <td>0.004736</td>\n",
" <td>00:18</td>\n",
" <td>0.058748</td>\n",
" <td>0.009240</td>\n",
" <td>0.002706</td>\n",
" <td>00:19</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
@@ -636,7 +636,7 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 3,
"metadata": {},
"outputs": [
{
@@ -645,7 +645,7 @@
"2"
]
},
"execution_count": null,
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
@@ -663,17 +663,17 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAJcAAADACAIAAACGdmZhAADHJElEQVR4nEz9Sa9lS5Ymhn2rMdt7n+Z23r4+XnSZEZXVJMVkUYSKpEoJUhxIIn+ANNH/ECBA/0ITQQQ00qAA6R+wWEpUqqTqsovIrIh4/fPu+r33NHubrUaDfV9AZ+Dw547r551j28zW+rpFwP/088+u/4//h//9//p/8/f+2f/t//T8yfkPfnzd20lks5yHX/3tq7/99avX7ylyk0jKIKQwapXzfDgfH5gAuIjUaUwmM4sIZkJkWCfzgWBLX+4P7Xwmd3JKcyYatGSmu7fWklCnqUzDsJl2l5esFEjPsHAIS9Fknd0skklFS5KmhYoWHQgcQURCxEwCIDLczdycwiQjgohqKbUOKurm3l1YVNXNez8qN+YAEJnpgYhSpC9LApSREaWMkVXrVlW7e2bqMC3NujuxSKkiEhkxN84QQu+niKUqVDnD3XqaZyaSSdkprNtQ6lRrb+18OnNRUgnlBW4RHojORaswpxtzFhGEIQIZlClCTMy6tSD37h76/EX5b/+7//Kf/Od//P/8Z//Xpc03T57NzRHj+Yxvv33/29++efPuFLEDBwHIjPRwJ3BGsLAQA8xMSGSCQCpCoIAxKJGn4+xLS/PoZksT4rRgIkQSEBGlFMvwcPROTVprwiUZSQAxiBKcINWqoAADlAlhJWJ3z3QmKVpAnBmZBASQyBRhruzdIjzCzRqFIxlhiYQDEXBf2lyEtCqDkiiIlmVx91JUpAKoZUwamDQTGZEEyhxq1UwQgSSQ3j3cVaUokwzWI8nN3b2HO9wBYmQiSbhQKVpEBLVmpmdCGMwFAGh9lCiDMolJGMKkdUJ4mHt4pgdoXuYkZYBZ9L/97/7kJz/b/o//4v/x9de//uXPPwjH6zen5dTevpu//urw+o1pecrQyKBMgCLTez/PS1hnYVVhUlAQiBwwj3VVMxkcQb1HLEYehYWkhDkIACKCAACsCuuREdbJpFvvRlBmlmAQUUIyAFZiQWQmCa877/G5IgI4GRkIICM804FgZmYipgBlBMyMDI4IT2JyiYhwq0wEeDMQIjMzwYR8XNHMXMIFEbDM9WfJrA9TUeIMBFE3CzcQSEGSBIIQgTyyeUcmAAKBKSKseRH1jO6OTFZNdxAziFgyM4LcWgJEhUGclJmtNWQmgCRAkGjuAWIigPTlB+fT6VcP7+4+eHn5+vXbr3/7fj6eHu7n4xnLrObTyFtiQxqQzAzP7t6XpkJDHaqKrg8x2D2WSA8vzCoV5ktzqvV8mpfjwunMDFEmJoATBGKiiMwEkjM5giKRTFyUtQpJgkU0ApHklhEJEEDrmcJM60cDAgzKRCZlUqYQQIkwRhABBKLkRGQogQmUwchBqLJ26z08AVXVUrho682Rc+u9d6JlKlyL1FrrOGQiQYhwBIDM9ZENVgbRYn1Zzh5tLJyEJAIAAYOJJeHuYKIQDmIGAM4EEpQsQZLJbuRORElEKgnqPYQLERMRyXrUoKi6RTcLD11OX9y+fri+2P7dr+fT3V07zyrVjEW2ZbhML+dukgvnTIIqmuG9NSDHYRzrQJRMTJzW3ZqlOSLXDUERAkqHNVvOZ5iJqAzD+nVHJINS2Fo3JBdCws3n1nU3QgvVUbggRVjP56V3Z0itg4iYWVEpWsAZYREBJDOIyT0T60ojCchkrP+dSBCl6np/rgcBIdiXRgkFs4rUUkoNQlrvZt3TIkVlaR3BLFylMlMEkImIiAATgCIMASg9ovXWbclkFWLRyKBAEjlzJCK8e5ZEUXHPSFMp7s5gZk4PC1QRApQgxB4R7kzy+9vEA5FBBQ5YhIXrfHw7SHn3+jVFIkTrpaB4ZNDQPRymQ812SuuULCyUBOS6LzOi9YUAAqw7EkRUVYgouqV1b30+nclzqoMndfc+LypKiXBnIlGNiABUKSN663Y+lzYWrcRZmAB2p4giMl7uL54/fzZNm+V8fn//vvfZvIf3REoRLQzkeiICCfDjLYPMQEYiLXM9gwjElAEQZSShaCEREBzo1h3ZWvdIYq7TZhq3drZMuLu7qxZhZubM9AgkmEmYHJFIYpBQdFta5FBVJD0DQSDKJNFBJTykDMO0cbM2L8ycICZWlmAIN6EAQYSZGUA3MgsSZmWRwkQBrG8sRTNSOQpBOQbvwRApY2vmHkAHpdRSSnj4fF6oJWdkBiGQcOtEaEubzycRqaXCASYVjdZ7b+F9OZ7aeVnOM3UrpONQ5/REmrv3TkQaAUJz97aQktYyIzb73bivmTIvnk7z8Xx1efP06bOnT55eXFyMwzA+H+4Od99+++XrN9+15gkDgimJAsgfPjwTAZEW3cM5ACZi8Lq8Yd26mxNyKHUcBiqytG6ttXDzsEh//GrVPQm0vta7VoQzU4UACXe3nolgmDszbcbB/eyewhwRS+8ErlVAkgnhwooEnZupSBlGawYpSHhASHfjNsmW3pJYRUEpkkQCLsRKJInMBK/VR2YKlFq1BFlNcwd6ZngyQSRIO3PLAGUIJyUyPSKEWZREJMy9m3sIK0EjPT3cIxIApVNbupmFBzwISRAIiFgEVMBEIkrCHL6YGdITTFS4zoelXuzI2RZ89unPLnbX07Qd62SdT2bWeBwuPvn4J0z61Te/DRARzIwoQUGQxysEBAIQgAeCASJmhvdu1jJChVULE/fo2ax1m1vr3UiKiogosRAzHD+0MRQRbrZ2L0gIEgjPiG7BxEwU6dGFiGV9nNg9LCI8tYiqMnFEmEW3hXL9H8z1N4VYREU1c1GiTCQYBNVCpCAmZoCICBmPHUMkMtUbA5RBlMiMyC4MVbAEUcJbZCCkFiUiAtx8/Yjrv8XEVWstVYQZHBkAsQoTI5JFHh9gUSFKpshgIhAzY/0rd2dVBUWYMNcykOckA/WsVD/4+KNPPvo8UzLJDdaDhYhQatnvds9f+N3D3e3771v2UokZGQk4EYhIuDAzy1oSO+DrFu3u3UyIiImYuru1JZBEwsTMAmImJS7MAiIAJCSkTExrtRKZCGIQggFO97B0lFo9fJnPWphEKFNEpnFqzdxTWFUqLAAKJNYFIayHE8yDUDJShUilaCQygkFj0YgfjgNQEGkSETLT3COgnIK1BxLKTKSJkHIKgxERMHMigTAThTtlqBZChnUGD8NQa2VQAGuBHplEFIgAWDUSHklMIiWJ0vpacEQEZYRnN9MKpzRLcIZnLM4TqfKzpy8++ejz8JJODna3BFUuxLV3tOU8DvsPX3727t2r1vo4bQGPiIi1NoxkZyGPzKRIyiQCRB4fKzC6R4/GTM17EokIqQgJESM5LIJIpYhwrj8jqsq8fpVuCKdMymCEMNKSMzOSIgQqzIu1BKZxo1yXbkLCQdGDCWBKJhYBk7knsJbfQhSggDAIlGBWYVU1cyBz7foBEDI90ynD03W9liOM2Zkjw4XBROTESZzKyWC2tetBEhMTZa5tWaoIiMMMImWYQJnhTGx9OZ98Xtp5Waz35B/2OoiIcv2yI5mIRCJhkZHRzOJ03kyXp7uHD5998tkHn/az94467jkRSeDIRGtOKsvSh2H77Onz7fbicDJhNbe1ngFIWEQKaxEtyEAmKDLN+hKUJMqM8O7m025UIouITETQ2tt4khMxKVhZg0DCvB6RzASEhZtlerohQphIiSPBOdUSGRQJC4uOChVdm3hYSFJEpIchjJFExJSRzCSiomUtws2DM0Ufb1Na67CIBDiTidLN09kDGUocYc19AUgEzCHMlCsexZqiTM7ZvUc4gVaYRkTdo83NugmTWUCcShXWQIR7azafzud5tghWpuTebemNagEg6wVFYKLI9HAQaqkQAnFb/EcvXj67ecnQwmMZh0hlSlWBOHN6JkhKHTPJArvtZevH1jyADCRIRUodiNQy16+dKQmcmcvi5o4IYiVRJOZuzEogj2CQakmHsBYRZmViJIkyrXdIurtTpvXFeg/ryGBmEVbW7inM0zgdz0c3L1wc4d2JEYGIoJQqggCFA5SZiVwL2PWOsYhI0jrM59ncq3BGtOa1CBIRa3WZSaIiCgn2TGjkAjLmLMoqQmAl5vUQdiTgEaBSS+3WI0K1WoQ3cwuoJnELl2HkKnO0aLMy2PP88HB+OMTSNJmJQSAhJUlKAj3eKci1X+/NN7upuQlLcyRPP/6Df3R9/dI6T9ubjJLEx/PZomspYItoPYKJZutayrC5pMMt2K2FR9RaINrDaT1NLFiIGNZm72e2TuZMVEgTaZRmAhFh1ipM+ohAJPdwAaoWwGBeSFgY4eGW6ZEmnCxwZ4KqDM4M8WAkodZtpkdGoQTQe1tPg06FRYM0mRlZ1qZ2hT4sfO02ZG2zpFFaIszdzEWYhLRKIjw8MlmFlYSYWVUiIwhrCU4EziASZaZIzwhm9ZT1TskElxJmhGDliABRBpxSigix+7kv3c7L8f7eTrMkIR7BNlUtIHcTYmZwEiIyAohaNCPXMzB7fPb5H9bNk+ZlqBe1XrSWHsFUheG+ZFgZSpL3SLdcUZ6EMgu4I0GsyTBzFlEWN6NMooi22DIjnSNJGAmAhYYybvoKBxInkbszMxGlERXhoWSEn+d0B4l7M28RIUyiSlI4hKFah6AEnIkpU1gzws2J17K9OYGFAzF7EDGI16IrPcz7CigSUyZFeAYV4SRdzCORWuZIBZjY0xEJz5bJHMwiAhXKWKvdYHDy4xUvlERFkSmivS3pAUBEVHUtD3rv8zy7e0RkJi0EBTKt9/k0t2Vx91yrKA8hACBaYc8kFg44wsPBNJZ6bl2H0R2Xlzc/+fEvVLbpPO72oiNHYGl1GAvrqdnSTqrIRCQCbh2iIiLuc7gnHotKYlIBc6QEwnt4hPXwdCOWTG7NWYtIiYx0jwxSJWZlKrUykYkLMyG99whyIyAjYE6Z8ESARYS4JNSIg9YiTgVrR2xCIsIctrgnglksqIerMgsDSIdnWIQLine truncated
"text/plain": [
"<PIL.Image.Image image mode=RGB size=151x192 at 0x7FB714B9C9D0>"
"<PIL.Image.Image image mode=RGB size=151x192 at 0x7F67678F7E90>"
]
},
"execution_count": null,
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
@@ -701,13 +701,13 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 5,
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "22613e22f0c9475cb9696273da3c3d60",
"model_id": "f78619047d7544908daa7fadd3c6f0c4",
"version_major": 2,
"version_minor": 0
},
@@ -741,7 +741,7 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 6,
"metadata": {
"hide_input": true
},
@@ -754,7 +754,7 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 8,
"metadata": {},
"outputs": [
{
@@ -771,14 +771,16 @@
"name": "stdout",
"output_type": "stream",
"text": [
"Is this a cat?: True; Probability it's a cat: 0.999992\n"
"Is this a cat?: True.\n",
"Probability it's a cat: 0.999986\n"
]
}
],
"source": [
"img = PILImage.create(uploader.data[0])\n",
"is_cat,_,probs = learn.predict(img)\n",
"print(f\"Is this a cat?: {is_cat}; Probability it's a cat: {probs[1].item():.6f}\")"
"print(f\"Is this a cat?: {is_cat}.\")\n",
"print(f\"Probability it's a cat: {probs[1].item():.6f}\")"
]
},
{
@@ -2905,20 +2907,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.5"
},
"toc": {
"base_numbering": 1,
"nav_menu": {},
"number_sections": false,
"sideBar": true,
"skip_h1_title": true,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
"version": "3.7.4"
}
},
"nbformat": 4,
02_production.ipynb
+3 -15
View File
@@ -1629,7 +1629,8 @@
],
"source": [
"#hide_output\n",
"VBox([widgets.Label('Select your bear!'), btn_upload, btn_run, out_pl, lbl_pred])"
"VBox([widgets.Label('Select your bear!'), \n",
" btn_upload, btn_run, out_pl, lbl_pred])"
]
},
{
@@ -1946,20 +1947,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.5"
},
"toc": {
"base_numbering": 1,
"nav_menu": {},
"number_sections": false,
"sideBar": true,
"skip_h1_title": true,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
"version": "3.7.4"
}
},
"nbformat": 4,
03_ethics.ipynb
-28
View File
@@ -1041,34 +1041,6 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.5"
},
"toc": {
"base_numbering": 1,
"nav_menu": {
"height": "600px",
"width": "365px"
},
"number_sections": false,
"sideBar": true,
"skip_h1_title": true,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
}
},
"nbformat": 4,
05_pet_breeds.ipynb
+1 -14
View File
@@ -2543,20 +2543,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.5"
},
"toc": {
"base_numbering": 1,
"nav_menu": {},
"number_sections": false,
"sideBar": true,
"skip_h1_title": true,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
"version": "3.7.4"
}
},
"nbformat": 4,
06_multicat.ipynb
+17 -2
View File
@@ -1445,11 +1445,14 @@
"metadata": {},
"outputs": [],
"source": [
"biwi = DataBlock(blocks=(ImageBlock, PointBlock),\n",
"biwi = DataBlock(\n",
" blocks=(ImageBlock, PointBlock),\n",
" get_items=get_image_files,\n",
" get_y=get_ctr,\n",
" splitter=FuncSplitter(lambda o: o.parent.name=='13'),\n",
" batch_tfms=[*aug_transforms(size=(240,320)), Normalize.from_stats(*imagenet_stats)])"
" batch_tfms=[*aug_transforms(size=(240,320)), \n",
" Normalize.from_stats(*imagenet_stats)]\n",
")"
]
},
{
@@ -1925,6 +1928,18 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
07_sizing_and_tta.ipynb
+14 -1
View File
@@ -460,7 +460,8 @@
],
"source": [
"dls = get_dls(128, 128)\n",
"learn = Learner(dls, xresnet50(), loss_func=CrossEntropyLossFlat(), metrics=accuracy)\n",
"learn = Learner(dls, xresnet50(), loss_func=CrossEntropyLossFlat(), \n",
" metrics=accuracy)\n",
"learn.fit_one_cycle(4, 3e-3)"
]
},
@@ -1011,6 +1012,18 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
08_collab.ipynb
+12
View File
@@ -2296,6 +2296,18 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
09_tabular.ipynb
-25
View File
@@ -9740,31 +9740,6 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.5"
},
"toc": {
"base_numbering": 1,
"nav_menu": {},
"number_sections": false,
"sideBar": true,
"skip_h1_title": true,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
}
},
"nbformat": 4,
10_nlp.ipynb
+5 -16
View File
@@ -1708,7 +1708,8 @@
"TEXT = \"I liked this movie because\"\n",
"N_WORDS = 40\n",
"N_SENTENCES = 2\n",
"preds = [learn.predict(TEXT, N_WORDS, temperature=0.75) for _ in range(N_SENTENCES)]"
"preds = [learn.predict(TEXT, N_WORDS, temperature=0.75) \n",
" for _ in range(N_SENTENCES)]"
]
},
{
@@ -1891,7 +1892,8 @@
"metadata": {},
"outputs": [],
"source": [
"learn = text_classifier_learner(dls_clas, AWD_LSTM, drop_mult=0.5, metrics=accuracy).to_fp16()"
"learn = text_classifier_learner(dls_clas, AWD_LSTM, drop_mult=0.5, \n",
" metrics=accuracy).to_fp16()"
]
},
{
@@ -2260,20 +2262,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.5"
},
"toc": {
"base_numbering": 1,
"nav_menu": {},
"number_sections": false,
"sideBar": true,
"skip_h1_title": true,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
"version": "3.7.4"
}
},
"nbformat": 4,
11_midlevel_data.ipynb
+7 -17
View File
@@ -602,7 +602,8 @@
"source": [
"cut = int(len(files)*0.8)\n",
"splits = [list(range(cut)), list(range(cut,len(files)))]\n",
"tls = TfmdLists(files, [Tokenizer.from_folder(path), Numericalize], splits=splits)"
"tls = TfmdLists(files, [Tokenizer.from_folder(path), Numericalize], \n",
" splits=splits)"
]
},
{
@@ -1083,7 +1084,8 @@
"class SiameseTransform(Transform):\n",
" def __init__(self, files, label_func, splits):\n",
" self.labels = files.map(label_func).unique()\n",
" self.lbl2files = {l: L(f for f in files if label_func(f) == l) for l in self.labels}\n",
" self.lbl2files = {l: L(f for f in files if label_func(f) == l) \n",
" for l in self.labels}\n",
" self.label_func = label_func\n",
" self.valid = {f: self._draw(f) for f in files[splits[1]]}\n",
" \n",
@@ -1095,7 +1097,8 @@
" def _draw(self, f):\n",
" same = random.random() < 0.5\n",
" cls = self.label_func(f)\n",
" if not same: cls = random.choice(L(l for l in self.labels if l != cls)) \n",
" if not same: \n",
" cls = random.choice(L(l for l in self.labels if l != cls)) \n",
" return random.choice(self.lbl2files[cls]),same"
]
},
@@ -1246,20 +1249,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.5"
},
"toc": {
"base_numbering": 1,
"nav_menu": {},
"number_sections": false,
"sideBar": true,
"skip_h1_title": true,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
"version": "3.7.4"
}
},
"nbformat": 4,
12_nlp_dive.ipynb
+9 -18
View File
@@ -497,7 +497,8 @@
}
],
"source": [
"learn = Learner(dls, LMModel1(len(vocab), 64), loss_func=F.cross_entropy, metrics=accuracy)\n",
"learn = Learner(dls, LMModel1(len(vocab), 64), loss_func=F.cross_entropy, \n",
" metrics=accuracy)\n",
"learn.fit_one_cycle(4, 1e-3)"
]
},
@@ -654,7 +655,8 @@
}
],
"source": [
"learn = Learner(dls, LMModel2(len(vocab), 64), loss_func=F.cross_entropy, metrics=accuracy)\n",
"learn = Learner(dls, LMModel2(len(vocab), 64), loss_func=F.cross_entropy, \n",
" metrics=accuracy)\n",
"learn.fit_one_cycle(4, 1e-3)"
]
},
@@ -1474,7 +1476,8 @@
}
],
"source": [
"learn = Learner(dls, LMModel5(len(vocab), 64, 2), loss_func=CrossEntropyLossFlat(), \n",
"learn = Learner(dls, LMModel5(len(vocab), 64, 2), \n",
" loss_func=CrossEntropyLossFlat(), \n",
" metrics=accuracy, cbs=ModelReseter)\n",
"learn.fit_one_cycle(15, 3e-3)"
]
@@ -1867,7 +1870,8 @@
}
],
"source": [
"learn = Learner(dls, LMModel6(len(vocab), 64, 2), loss_func=CrossEntropyLossFlat(), \n",
"learn = Learner(dls, LMModel6(len(vocab), 64, 2), \n",
" loss_func=CrossEntropyLossFlat(), \n",
" metrics=accuracy, cbs=ModelReseter)\n",
"learn.fit_one_cycle(15, 1e-2)"
]
@@ -2354,20 +2358,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.5"
},
"toc": {
"base_numbering": 1,
"nav_menu": {},
"number_sections": false,
"sideBar": true,
"skip_h1_title": true,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
"version": "3.7.4"
}
},
"nbformat": 4,
13_convolutions.ipynb
+12
View File
@@ -3757,6 +3757,18 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
14_resnet.ipynb
+12
View File
@@ -1268,6 +1268,18 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
15_arch_details.ipynb
+347 -1
View File
@@ -233,7 +233,341 @@
"source": [
"This picture shows the CNN body on the left (in this case, it's a regular CNN, not a ResNet, and they're using 2x2 max pooling instead of stride 2 convolutions, since this paper was written before ResNets came along) and it shows the transposed convolutional layers on the right (they're called \"up-conv\" in this picture). Then then extra skip connections are shown as grey arrows crossing from left to right (these are sometimes called *cross connections*). You can see why it's called a \"U-net\" when you see this picture!\n",
"\n",
"With this architecture, the input to the transposed convolutions is not just the lower resolution grid in the preceding layer, but also the higher resolution grid in the resnet head. This allows the U-Net to use all of the information of the original image, as it is needed. One challenge with U-Nets is that the exact architecture depends on the image size. fastai has a unique `DynamicUnet` class which auto-generates an architecture of the right size based on the data provided."
"With this architecture, the input to the transposed convolutions is not just the lower resolution grid in the preceding layer, but also the higher resolution grid in the resnet head. This allows the U-Net to use all of the information of the original image, as it is needed. One challenge with U-Nets is that the exact architecture depends on the image size. fastai has a unique `DynamicUnet` class which auto-generates an architecture of the right size based on the data provided.\n",
"\n",
"Let's focus now on an example where we leverage the fastai library to write a custom model:"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### A Siamese network"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"#hide\n",
"from fastai2.vision.all import *\n",
"path = untar_data(URLs.PETS)\n",
"files = get_image_files(path/\"images\")\n",
"\n",
"class SiameseImage(Tuple):\n",
" def show(self, ctx=None, **kwargs): \n",
" img1,img2,same_breed = self\n",
" if not isinstance(img1, Tensor):\n",
" if img2.size != img1.size: img2 = img2.resize(img1.size)\n",
" t1,t2 = tensor(img1),tensor(img2)\n",
" t1,t2 = t1.permute(2,0,1),t2.permute(2,0,1)\n",
" else: t1,t2 = img1,img2\n",
" line = t1.new_zeros(t1.shape[0], t1.shape[1], 10)\n",
" return show_image(torch.cat([t1,line,t2], dim=2), \n",
" title=same_breed, ctx=ctx)\n",
" \n",
"def label_func(fname):\n",
" return re.match(r'^(.*)_\\d+.jpg$', fname.name).groups()[0]\n",
"\n",
"class SiameseTransform(Transform):\n",
" def __init__(self, files, label_func, splits):\n",
" self.labels = files.map(label_func).unique()\n",
" self.lbl2files = {l: L(f for f in files if label_func(f) == l) for l in self.labels}\n",
" self.label_func = label_func\n",
" self.valid = {f: self._draw(f) for f in files[splits[1]]}\n",
" \n",
" def encodes(self, f):\n",
" f2,t = self.valid.get(f, self._draw(f))\n",
" img1,img2 = PILImage.create(f),PILImage.create(f2)\n",
" return SiameseImage(img1, img2, t)\n",
" \n",
" def _draw(self, f):\n",
" same = random.random() < 0.5\n",
" cls = self.label_func(f)\n",
" if not same: cls = random.choice(L(l for l in self.labels if l != cls)) \n",
" return random.choice(self.lbl2files[cls]),same\n",
" \n",
"splits = RandomSplitter()(files)\n",
"tfm = SiameseTransform(files, label_func, splits)\n",
"tls = TfmdLists(files, tfm, splits=splits)\n",
"dls = tls.dataloaders(after_item=[Resize(224), ToTensor], \n",
" after_batch=[IntToFloatTensor, Normalize.from_stats(*imagenet_stats)])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's go back to the input pipeline we set up in <<chapter_midlevel_data>> for a Siamese network. If your remember, it consisted of pair of images with the label being `True` or `False`, depending on if they were in the same class or not.\n",
"\n",
"Using what we just saw, let's build a custom model for this task and train it. How? We will use a pretrained architecture and pass our two images throught it. Then we can concatenate the results and send them to a custom head that will return two predictions. In terms of modules, this looks like this:"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class SiameseModel(Module):\n",
" def __init__(self, encoder, head):\n",
" self.encoder,self.head = encoder,head\n",
" \n",
" def forward(self, x1, x2):\n",
" ftrs = torch.cat([self.encoder(x1), self.encoder(x2)], dim=1)\n",
" return self.head(ftrs)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"To create our encoder, we just need to take a pretrained model and cut it, as we explained before. The function `create_body` does that for us, we just have to pass it the place we want to cut. If we remember our look in the dictionary of metadata for pretrained models, the cut value for a resnet is -2:"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"encoder = create_body(resnet34, cut=-2)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Then we can create our head. A look at the encoder tells us the last layer has 512 features, so this head will need to receive `512*4`. Why 4? First we have to multiply by 2 because we have two images. Then we need a second multiplication by 2 because of our concat-pool trick."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"head = create_head(512*4, 2, ps=0.5)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"With our encoder and head, we can now build our model."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"model = SiameseModel(encoder, head)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Before using `Learner`, we have two more things to define. First, we must define the loss function we want to use. It's regular cross entropy, but since our targets are booleans, we need to convert them to integers or PyTorch will throw an error."
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {},
"outputs": [],
"source": [
"def loss_func(out, targ):\n",
" return nn.CrossEntropyLoss()(out, targ.long())"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"More importantly, to take full advantage of transfer learning, we have to define a custom *splitter*. A splitter is a function that tells the fastai library how to split the model in several parameter groups. This is what is used behind the scenes not only train the head of a model when we do transfert learning. \n",
"\n",
"Here we want two parameter groups: one for the encoder and one for the head. We can thus define the following splitter (`params` is jsut a function that returns all parameters of a given module):"
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [],
"source": [
"def siamese_splitter(model):\n",
" return [params(model.encoder), params(model.head)]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Then we can define our `Learner` by passing the data, model, loss function, spliiter and any metric we want. Since we are not using a convenience function from fastai for transfer learning (like `cnn_learner`), we have to call `learn.freeze` manually. This will make sure only the last parameter groups (in this case, the head) is trained. "
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [],
"source": [
"learn = Learner(dls, model, loss_func=loss_func, \n",
" splitter=siamese_splitter, metrics=accuracy)\n",
"learn.freeze()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Then we can directly train our model with the usual methods:"
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {
"scrolled": true
},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>0.367015</td>\n",
" <td>0.281242</td>\n",
" <td>0.885656</td>\n",
" <td>00:26</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>0.307688</td>\n",
" <td>0.214721</td>\n",
" <td>0.915426</td>\n",
" <td>00:26</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>0.275221</td>\n",
" <td>0.170615</td>\n",
" <td>0.936401</td>\n",
" <td>00:26</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>0.223771</td>\n",
" <td>0.159633</td>\n",
" <td>0.943843</td>\n",
" <td>00:26</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn.fit_one_cycle(4, 3e-3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Before unfreezing and training a bit more with discriminative learning rates..."
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>0.212744</td>\n",
" <td>0.159033</td>\n",
" <td>0.944520</td>\n",
" <td>00:35</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>0.201893</td>\n",
" <td>0.159615</td>\n",
" <td>0.942490</td>\n",
" <td>00:35</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>0.204606</td>\n",
" <td>0.152338</td>\n",
" <td>0.945196</td>\n",
" <td>00:36</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>0.213203</td>\n",
" <td>0.148346</td>\n",
" <td>0.947903</td>\n",
" <td>00:36</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn.unfreeze()\n",
"learn.fit_one_cycle(4, slice(1e-6,1e-4))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"94.8% is very good when we remember a classifier trained the same way (with no data augmentation) had an arror rate of 7%."
]
},
{
@@ -476,6 +810,18 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
16_accel_sgd.ipynb
+12
View File
@@ -1307,6 +1307,18 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
17_foundations.ipynb
+16 -2
View File
@@ -2037,7 +2037,8 @@
" return self.out\n",
" \n",
" def backward(self):\n",
" self.inp.g = 2. * (self.inp.squeeze() - self.targ).unsqueeze(-1) / self.targ.shape[0]"
" x = (self.inp.squeeze()-self.targ).unsqueeze(-1)\n",
" self.inp.g = 2.*x/self.targ.shape[0]"
]
},
{
@@ -2189,7 +2190,8 @@
"source": [
"class Mse(LayerFunction):\n",
" def forward (self, inp, targ): return (inp.squeeze() - targ).pow(2).mean()\n",
" def bwd(self, out, inp, targ): inp.g = 2*(inp.squeeze()-targ).unsqueeze(-1) / targ.shape[0]"
" def bwd(self, out, inp, targ): \n",
" inp.g = 2*(inp.squeeze()-targ).unsqueeze(-1) / targ.shape[0]"
]
},
{
@@ -2446,6 +2448,18 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
18_CAM.ipynb
+12
View File
@@ -653,6 +653,18 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
19_learner.ipynb
+135 -105
View File
@@ -2,7 +2,7 @@
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -26,6 +26,13 @@
"The end of chapter questionnaire is particularly important for this chapter. This is where we will be getting you started on the many interesting directions that you could take, using this chapter as your starting out point. What we are really saying is: follow through with this chapter on your computer, not on paper, and do lots of experiments, web searches, and whatever else you need to understand what's going on. You've built up the skills and expertise to do this in the rest of this book, so we think you are going to go great!"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"First let's start with gathering (manually) the data."
]
},
{
"cell_type": "markdown",
"metadata": {},
@@ -42,7 +49,7 @@
},
{
"cell_type": "code",
"execution_count": 2,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -58,7 +65,7 @@
},
{
"cell_type": "code",
"execution_count": 3,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -67,7 +74,7 @@
"Path('/home/jhoward/.fastai/data/imagenette2-160/val/n03417042/n03417042_3752.JPEG')"
]
},
"execution_count": 3,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -86,7 +93,7 @@
},
{
"cell_type": "code",
"execution_count": 4,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -95,7 +102,7 @@
"Path('/home/jhoward/.fastai/data/imagenette2-160/val/n03417042/n03417042_3752.JPEG')"
]
},
"execution_count": 4,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -117,7 +124,7 @@
},
{
"cell_type": "code",
"execution_count": 5,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -127,7 +134,7 @@
"<PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=213x160 at 0x7FA45AC27D50>"
]
},
"execution_count": 5,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -139,7 +146,7 @@
},
{
"cell_type": "code",
"execution_count": 6,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -148,7 +155,7 @@
"torch.Size([160, 213, 3])"
]
},
"execution_count": 6,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -167,7 +174,7 @@
},
{
"cell_type": "code",
"execution_count": 7,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -176,7 +183,7 @@
"(#10) ['n03417042','n03445777','n03888257','n03394916','n02979186','n03000684','n03425413','n01440764','n03028079','n02102040']"
]
},
"execution_count": 7,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -194,7 +201,7 @@
},
{
"cell_type": "code",
"execution_count": 8,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -212,7 +219,7 @@
" 'n02102040': 9}"
]
},
"execution_count": 8,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -225,14 +232,14 @@
"cell_type": "markdown",
"metadata": {},
"source": [
"That's all the pieces we need to put together our `Dataset`:"
"That's all the pieces we need to put together our `Dataset`."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Dataset"
"### Dataset"
]
},
{
@@ -244,7 +251,7 @@
},
{
"cell_type": "code",
"execution_count": 9,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -266,7 +273,7 @@
},
{
"cell_type": "code",
"execution_count": 10,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -275,7 +282,7 @@
"(9469, 3925)"
]
},
"execution_count": 10,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -295,7 +302,7 @@
},
{
"cell_type": "code",
"execution_count": 11,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -304,7 +311,7 @@
"(torch.Size([64, 64, 3]), tensor(0))"
]
},
"execution_count": 11,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -317,7 +324,7 @@
},
{
"cell_type": "code",
"execution_count": 12,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -346,7 +353,7 @@
},
{
"cell_type": "code",
"execution_count": 13,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -364,7 +371,7 @@
},
{
"cell_type": "code",
"execution_count": 14,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -373,7 +380,7 @@
"(torch.Size([2, 64, 64, 3]), tensor([0, 0]))"
]
},
"execution_count": 14,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -392,7 +399,7 @@
},
{
"cell_type": "code",
"execution_count": 15,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -419,7 +426,7 @@
},
{
"cell_type": "code",
"execution_count": 16,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -428,7 +435,7 @@
"(torch.Size([128, 64, 64, 3]), torch.Size([128]), 74)"
]
},
"execution_count": 16,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -452,7 +459,7 @@
},
{
"cell_type": "code",
"execution_count": 17,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -461,7 +468,7 @@
"[tensor([0.4544, 0.4453, 0.4141]), tensor([0.2812, 0.2766, 0.2981])]"
]
},
"execution_count": 17,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -480,7 +487,7 @@
},
{
"cell_type": "code",
"execution_count": 18,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -501,7 +508,7 @@
},
{
"cell_type": "code",
"execution_count": 19,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -511,7 +518,7 @@
},
{
"cell_type": "code",
"execution_count": 20,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -520,7 +527,7 @@
"(tensor([0.3732, 0.4907, 0.5633]), tensor([1.0212, 1.0311, 1.0131]))"
]
},
"execution_count": 20,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -537,6 +544,13 @@
"Here `tfm_x` isn't just applying `Normalize`, but is also permuting the axis order from `NHWC` to `NCHW` (see <<chapter_convolutions>> if you need a reminder what these acronyms refer to). PIL uses `HWC` axis order, which we can't use with PyTorch, hence the need for this `permute`."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"That's all we need for the data for our model. So now we need the model itself!"
]
},
{
"cell_type": "markdown",
"metadata": {},
@@ -548,12 +562,12 @@
"cell_type": "markdown",
"metadata": {},
"source": [
"That's all we need for the data for our model. So now we need the model itself! To create a model, we'll need `Module`. To create `Module`, we'll need `Parameter`, so let's start there. Recall that in <<chapter_collab>> we said that `Parameter` \"this class doesn't actually add any functionality (other than automatically calling `requires_grad_()` for us). It's only used as a 'marker' to show what to include in `parameters()`\". Here's a definition which does exactly that:"
"To create a model, we'll need `Module`. To create `Module`, we'll need `Parameter`, so let's start there. Recall that in <<chapter_collab>> we said that `Parameter` \"this class doesn't actually add any functionality (other than automatically calling `requires_grad_()` for us). It's only used as a 'marker' to show what to include in `parameters()`\". Here's a definition which does exactly that:"
]
},
{
"cell_type": "code",
"execution_count": 21,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -573,7 +587,7 @@
},
{
"cell_type": "code",
"execution_count": 22,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -582,7 +596,7 @@
"tensor(3., requires_grad=True)"
]
},
"execution_count": 22,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -600,7 +614,7 @@
},
{
"cell_type": "code",
"execution_count": 23,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -658,7 +672,7 @@
},
{
"cell_type": "code",
"execution_count": 24,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -686,7 +700,7 @@
},
{
"cell_type": "code",
"execution_count": 25,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -695,7 +709,7 @@
"2"
]
},
"execution_count": 25,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -714,7 +728,7 @@
},
{
"cell_type": "code",
"execution_count": 26,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -723,7 +737,7 @@
"torch.Size([128, 4, 64, 64])"
]
},
"execution_count": 26,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -743,7 +757,7 @@
},
{
"cell_type": "code",
"execution_count": 27,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -766,7 +780,7 @@
},
{
"cell_type": "code",
"execution_count": 28,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -775,7 +789,7 @@
"torch.Size([3, 2])"
]
},
"execution_count": 28,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -795,7 +809,7 @@
},
{
"cell_type": "code",
"execution_count": 29,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -814,7 +828,7 @@
},
{
"cell_type": "code",
"execution_count": 30,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -823,7 +837,7 @@
"4"
]
},
"execution_count": 30,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -842,7 +856,7 @@
},
{
"cell_type": "code",
"execution_count": 31,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -851,7 +865,7 @@
"device(type='cuda', index=5)"
]
},
"execution_count": 31,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -865,7 +879,14 @@
"cell_type": "markdown",
"metadata": {},
"source": [
"## Simple CNN"
"We can now use those pieces to create a CNN."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Simple CNN"
]
},
{
@@ -877,7 +898,7 @@
},
{
"cell_type": "code",
"execution_count": 32,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -915,7 +936,7 @@
},
{
"cell_type": "code",
"execution_count": 33,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -932,7 +953,7 @@
},
{
"cell_type": "code",
"execution_count": 34,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -956,7 +977,7 @@
},
{
"cell_type": "code",
"execution_count": 35,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -965,7 +986,7 @@
"10"
]
},
"execution_count": 35,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -984,7 +1005,7 @@
},
{
"cell_type": "code",
"execution_count": 36,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -1003,7 +1024,7 @@
"torch.Size([128, 10])"
]
},
"execution_count": 36,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -1016,6 +1037,13 @@
"r.shape"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We have data and model. Now we need a loss function."
]
},
{
"cell_type": "markdown",
"metadata": {},
@@ -1027,12 +1055,12 @@
"cell_type": "markdown",
"metadata": {},
"source": [
"We have data and model. Now we need a loss function. We've already seen how to define \"negative log likelihood\":"
"We've already seen how to define \"negative log likelihood\":"
]
},
{
"cell_type": "code",
"execution_count": 37,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -1048,7 +1076,7 @@
},
{
"cell_type": "code",
"execution_count": 38,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -1057,7 +1085,7 @@
"tensor(-1.2790, grad_fn=<SelectBackward>)"
]
},
"execution_count": 38,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -1077,7 +1105,7 @@
},
{
"cell_type": "code",
"execution_count": 39,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -1086,7 +1114,7 @@
"tensor(2.5666, grad_fn=<NegBackward>)"
]
},
"execution_count": 39,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -1109,7 +1137,7 @@
},
{
"cell_type": "code",
"execution_count": 40,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -1118,7 +1146,7 @@
"tensor(-1.2790, grad_fn=<SelectBackward>)"
]
},
"execution_count": 40,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -1144,7 +1172,7 @@
},
{
"cell_type": "code",
"execution_count": 41,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -1153,7 +1181,7 @@
"tensor(True)"
]
},
"execution_count": 41,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -1173,7 +1201,7 @@
},
{
"cell_type": "code",
"execution_count": 42,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -1182,7 +1210,7 @@
"tensor(3.9784, grad_fn=<SelectBackward>)"
]
},
"execution_count": 42,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -1204,7 +1232,7 @@
},
{
"cell_type": "code",
"execution_count": 43,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -1220,7 +1248,7 @@
},
{
"cell_type": "code",
"execution_count": 44,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -1229,7 +1257,7 @@
"tensor(-1.2790, grad_fn=<SelectBackward>)"
]
},
"execution_count": 44,
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
@@ -1247,13 +1275,20 @@
},
{
"cell_type": "code",
"execution_count": 45,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def cross_entropy(preds, yb): return nll(log_softmax(preds), yb).mean()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's now combine all those pieces together to create a `Learner`."
]
},
{
"cell_type": "markdown",
"metadata": {},
@@ -1270,7 +1305,7 @@
},
{
"cell_type": "code",
"execution_count": 46,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -1291,7 +1326,7 @@
},
{
"cell_type": "code",
"execution_count": 47,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -1310,12 +1345,10 @@
},
{
"cell_type": "code",
"execution_count": 57,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"#id class_learner\n",
"#caption The Learner class\n",
"class Learner:\n",
" def __init__(self, model, dls, loss_func, lr, cbs, opt_func=SGD):\n",
" store_attr(self, 'model,dls,loss_func,lr,cbs,opt_func')\n",
@@ -1375,7 +1408,7 @@
"cell_type": "markdown",
"metadata": {},
"source": [
"## Callbacks"
"### Callbacks"
]
},
{
@@ -1391,7 +1424,7 @@
},
{
"cell_type": "code",
"execution_count": 58,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -1407,7 +1440,7 @@
},
{
"cell_type": "code",
"execution_count": 59,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -1430,7 +1463,7 @@
},
{
"cell_type": "code",
"execution_count": 60,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -1460,7 +1493,7 @@
},
{
"cell_type": "code",
"execution_count": 61,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -1508,14 +1541,14 @@
"cell_type": "markdown",
"metadata": {},
"source": [
"It's quite amazing to realize that we can implement all the key ideas from fastai's `Learner` in so little code!"
"It's quite amazing to realize that we can implement all the key ideas from fastai's `Learner` in so little code! Let's now add some learning rate scheduling."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Annealing"
"### Scheduling the learning rate"
]
},
{
@@ -1527,7 +1560,7 @@
},
{
"cell_type": "code",
"execution_count": 62,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -1556,7 +1589,7 @@
},
{
"cell_type": "code",
"execution_count": 63,
"execution_count": null,
"metadata": {},
"outputs": [
{
@@ -1719,7 +1752,7 @@
},
{
"cell_type": "code",
"execution_count": 64,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
@@ -1756,6 +1789,13 @@
"plt.plot(onecyc.lrs);"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"TK conclusion"
]
},
{
"cell_type": "markdown",
"metadata": {},
@@ -1841,6 +1881,9 @@
}
],
"metadata": {
"jupytext": {
"split_at_heading": true
},
"kernelspec": {
"display_name": "Python 3",
"language": "python",
@@ -1856,20 +1899,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.5"
},
"toc": {
"base_numbering": 1,
"nav_menu": {},
"number_sections": false,
"sideBar": true,
"skip_h1_title": true,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
"version": "3.7.4"
}
},
"nbformat": 4,
20_conclusion.ipynb
+12
View File
@@ -69,6 +69,18 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
app_blog.ipynb
+12
View File
@@ -292,6 +292,18 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
app_jupyter.ipynb
-25
View File
@@ -533,31 +533,6 @@
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.5"
},
"toc": {
"base_numbering": 1,
"nav_menu": {},
"number_sections": false,
"sideBar": true,
"skip_h1_title": true,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
}
},
"nbformat": 4,
clean/01_intro.ipynb
+1504
View File
File diff suppressed because it is too large. Load diff
clean/02_production.ipynb
+1029
View File
File diff suppressed because it is too large. Load diff
clean/03_ethics.ipynb
+261
View File
@@ -0,0 +1,261 @@
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Data Ethics"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Sidebar: Acknowledgement: Dr Rachel Thomas"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### End sidebar"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Key examples for data ethics"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Bugs and recourse: Buggy algorithm used for healthcare benefits"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Feedback loops: YouTube's recommendation system"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Bias: Professor Lantanya Sweeney \"arrested\""
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Why does this matter?"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Integrating machine learning with product design"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Topics in Data Ethics"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Recourse and accountability"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Feedback loops"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Bias"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Historical bias"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Measurement bias"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Aggregation Bias"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Representation Bias"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Addressing different types of bias"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Humans are biased, so does algorithmic bias matter?"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Disinformation"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Identifying and addressing ethical issues"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Analyze a project you are working on"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Processes to implement"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Ethical Lenses"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### The power of diversity"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Fairness, accountability, and transparency"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Role of Policy"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### The effectiveness of regulation"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Rights and policy"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Cars: a historical precedent"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Conclusion"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Questionnaire"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Further research:"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Section 1: that's a wrap!"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"jupytext": {
"split_at_heading": true
},
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
}
},
"nbformat": 4,
"nbformat_minor": 4
}
clean/04_mnist_basics.ipynb
+3966
View File
File diff suppressed because it is too large. Load diff
clean/05_pet_breeds.ipynb
+1732
View File
File diff suppressed because it is too large. Load diff
clean/06_multicat.ipynb
+1332
View File
File diff suppressed because it is too large. Load diff
clean/07_sizing_and_tta.ipynb
+654
View File
@@ -0,0 +1,654 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"#hide\n",
"from utils import *"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Training a state-of-the-art model"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Imagenette"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from fastai2.vision.all import *\n",
"path = untar_data(URLs.IMAGENETTE)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"dblock = DataBlock(blocks=(ImageBlock(), CategoryBlock()),\n",
" get_items=get_image_files,\n",
" get_y=parent_label,\n",
" item_tfms=Resize(460),\n",
" batch_tfms=aug_transforms(size=224, min_scale=0.75))\n",
"dls = dblock.dataloaders(path, bs=64)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>1.583403</td>\n",
" <td>2.064317</td>\n",
" <td>0.401792</td>\n",
" <td>01:03</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>1.208877</td>\n",
" <td>1.260106</td>\n",
" <td>0.601568</td>\n",
" <td>01:02</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>0.925265</td>\n",
" <td>1.036154</td>\n",
" <td>0.664302</td>\n",
" <td>01:03</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>0.730190</td>\n",
" <td>0.700906</td>\n",
" <td>0.777819</td>\n",
" <td>01:03</td>\n",
" </tr>\n",
" <tr>\n",
" <td>4</td>\n",
" <td>0.585707</td>\n",
" <td>0.541810</td>\n",
" <td>0.825243</td>\n",
" <td>01:03</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"model = xresnet50()\n",
"learn = Learner(dls, model, loss_func=CrossEntropyLossFlat(), metrics=accuracy)\n",
"learn.fit_one_cycle(5, 3e-3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Normalization"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(TensorImage([0.4842, 0.4711, 0.4511], device='cuda:5'),\n",
" TensorImage([0.2873, 0.2893, 0.3110], device='cuda:5'))"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"x,y = dls.one_batch()\n",
"x.mean(dim=[0,2,3]),x.std(dim=[0,2,3])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def get_dls(bs, size):\n",
" dblock = DataBlock(blocks=(ImageBlock, CategoryBlock),\n",
" get_items=get_image_files,\n",
" get_y=parent_label,\n",
" item_tfms=Resize(460),\n",
" batch_tfms=[*aug_transforms(size=size, min_scale=0.75),\n",
" Normalize.from_stats(*imagenet_stats)])\n",
" return dblock.dataloaders(path, bs=bs)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"dls = get_dls(64, 224)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(TensorImage([-0.0787, 0.0525, 0.2136], device='cuda:5'),\n",
" TensorImage([1.2330, 1.2112, 1.3031], device='cuda:5'))"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"x,y = dls.one_batch()\n",
"x.mean(dim=[0,2,3]),x.std(dim=[0,2,3])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>1.632865</td>\n",
" <td>2.250024</td>\n",
" <td>0.391337</td>\n",
" <td>01:02</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>1.294041</td>\n",
" <td>1.579932</td>\n",
" <td>0.517177</td>\n",
" <td>01:02</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>0.960535</td>\n",
" <td>1.069164</td>\n",
" <td>0.657207</td>\n",
" <td>01:04</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>0.730220</td>\n",
" <td>0.767433</td>\n",
" <td>0.771845</td>\n",
" <td>01:05</td>\n",
" </tr>\n",
" <tr>\n",
" <td>4</td>\n",
" <td>0.577889</td>\n",
" <td>0.550673</td>\n",
" <td>0.824496</td>\n",
" <td>01:06</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"model = xresnet50()\n",
"learn = Learner(dls, model, loss_func=CrossEntropyLossFlat(), metrics=accuracy)\n",
"learn.fit_one_cycle(5, 3e-3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Progressive resizing"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>1.902943</td>\n",
" <td>2.447006</td>\n",
" <td>0.401419</td>\n",
" <td>00:30</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>1.315203</td>\n",
" <td>1.572992</td>\n",
" <td>0.525765</td>\n",
" <td>00:30</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>1.001199</td>\n",
" <td>0.767886</td>\n",
" <td>0.759149</td>\n",
" <td>00:30</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>0.765864</td>\n",
" <td>0.665562</td>\n",
" <td>0.797984</td>\n",
" <td>00:30</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"dls = get_dls(128, 128)\n",
"learn = Learner(dls, xresnet50(), loss_func=CrossEntropyLossFlat(), \n",
" metrics=accuracy)\n",
"learn.fit_one_cycle(4, 3e-3)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>0.985213</td>\n",
" <td>1.654063</td>\n",
" <td>0.565721</td>\n",
" <td>01:06</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>0.706869</td>\n",
" <td>0.689622</td>\n",
" <td>0.784541</td>\n",
" <td>01:07</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>0.739217</td>\n",
" <td>0.928541</td>\n",
" <td>0.712472</td>\n",
" <td>01:07</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>0.629462</td>\n",
" <td>0.788906</td>\n",
" <td>0.764003</td>\n",
" <td>01:07</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>0.491912</td>\n",
" <td>0.502622</td>\n",
" <td>0.836445</td>\n",
" <td>01:06</td>\n",
" </tr>\n",
" <tr>\n",
" <td>4</td>\n",
" <td>0.414880</td>\n",
" <td>0.431332</td>\n",
" <td>0.863331</td>\n",
" <td>01:06</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn.dls = get_dls(64, 224)\n",
"learn.fine_tune(5, 1e-3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Test time augmentation"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/plain": [
"0.8737863898277283"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"preds,targs = learn.tta()\n",
"accuracy(preds, targs).item()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Mixup"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Sidebar: Papers and math"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### End sidebar"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"hide_input": true
},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 864x288 with 3 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"church = PILImage.create(get_image_files_sorted(path/'train'/'n03028079')[0])\n",
"gas = PILImage.create(get_image_files_sorted(path/'train'/'n03425413')[0])\n",
"church = church.resize((256,256))\n",
"gas = gas.resize((256,256))\n",
"tchurch = tensor(church).float() / 255.\n",
"tgas = tensor(gas).float() / 255.\n",
"\n",
"_,axs = plt.subplots(1, 3, figsize=(12,4))\n",
"show_image(tchurch, ax=axs[0]);\n",
"show_image(tgas, ax=axs[1]);\n",
"show_image((0.3*tchurch + 0.7*tgas), ax=axs[2]);"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Label smoothing"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Sidebar: Label smoothing, the paper"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### End sidebar"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Conclusion"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Questionnaire"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Further research\n",
"\n",
"1. Use the fastai documentation to build a function that crops an image to a square in the four corners, then implement a TTA method that averages the predictions on a center crop and those four crops. Did it help? Is it better than the TTA method of fastai?\n",
"1. Find the Mixup paper on arxiv and read it. Pick one or two more recent articles introducing variants of Mixup and read them, then try to implement them on your problem.\n",
"1. Find the script training Imagenette using Mixup and use it as an example to build a script for a long training on your own project. Execute it and see if it helped.\n",
"1. Read the sidebar on the math of label smoothing, and look at the relevant section of the original paper, and see if you can follow it. Don't be afraid to ask for help!"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"jupytext": {
"split_at_heading": true
},
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
clean/08_collab.ipynb
+1685
View File
File diff suppressed because it is too large. Load diff
clean/09_tabular.ipynb
+8301
View File
File diff suppressed because it is too large. Load diff
clean/10_nlp.ipynb
+1531
View File
File diff suppressed because it is too large. Load diff
clean/11_midlevel_data.ipynb
+840
View File
@@ -0,0 +1,840 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"#hide\n",
"from utils import *\n",
"from IPython.display import display,HTML"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Data munging with fastai's mid-level API"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Going deeper into fastai's layered API"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from fastai2.text.all import *\n",
"\n",
"dls = TextDataLoaders.from_folder(untar_data(URLs.IMDB), valid='test')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"path = untar_data(URLs.IMDB)\n",
"dls = DataBlock(\n",
" blocks=(TextBlock.from_folder(path),CategoryBlock),\n",
" get_y = parent_label,\n",
" get_items=partial(get_text_files, folders=['train', 'test']),\n",
" splitter=GrandparentSplitter(valid_name='test')\n",
").dataloaders(path)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Transforms"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"files = get_text_files(path, folders = ['train', 'test'])\n",
"txts = L(o.open().read() for o in files[:2000])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(#228) ['xxbos','xxmaj','this','movie',',','which','i','just','discovered','at'...]"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tok = Tokenizer.from_folder(path)\n",
"tok.setup(txts)\n",
"toks = txts.map(tok)\n",
"toks[0]"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"tensor([ 2, 8, 20, 27, 11, 88, 18, 53, 3286, 45])"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"num = Numericalize()\n",
"num.setup(toks)\n",
"nums = toks.map(num)\n",
"nums[0][:10]"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"tensor([ 2, 8, 20, 27, 11, 88, 18, 53, 3286, 45])"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"num = Numericalize()\n",
"num.setup(toks)\n",
"nums = toks.map(num)\n",
"nums[0][:10]"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(#10) ['xxbos','xxmaj','this','movie',',','which','i','just','discovered','at']"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"nums_dec = num.decode(nums[0][:10]); nums_dec"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"'xxbos xxmaj this movie , which i just discovered at'"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tok.decode(nums_dec)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"((#374) ['xxbos','xxmaj','well',',','\"','cube','\"','(','1997',')'...],\n",
" (#207) ['xxbos','xxmaj','conrad','xxmaj','hall','went','out','with','a','bang'...])"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tok((txts[0], txts[1]))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Writing your own Transform"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"3"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"def f(x): return x+1\n",
"tfm = Transform(f)\n",
"tfm(2)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class NormalizeMean(Transform):\n",
" def setups(self, items): self.mean = sum(items)/len(items)\n",
" def encodes(self, x): return x-self.mean\n",
" def decodes(self, x): return x+self.mean"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(3.0, 5.0, 2.0)"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tfm = NormalizeMean()\n",
"tfm.setup([1,2,3,4,5])\n",
"start = 2\n",
"y = tfm(start)\n",
"z = tfm.decode(y)\n",
"tfm.mean,y,z"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Pipeline"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"tensor([ 2, 8, 76, 10, 23, 3112, 23, 34, 3113, 33, 10, 8, 4477, 22, 88, 32, 10, 27, 42, 14])"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tfms = Pipeline([tok, num])\n",
"t = tfms(txts[0]); t[:20]"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"'xxbos xxmaj well , \" cube \" ( 1997 ) , xxmaj vincenzo \\'s first movie , was one of the most interesti'"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tfms.decode(t)[:100]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## TfmdLists and Datasets: Transformed collections"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### TfmdLists"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"tls = TfmdLists(files, [Tokenizer.from_folder(path), Numericalize])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"tensor([ 2, 8, 91, 11, 22, 5793, 22, 37, 4910, 34, 11, 8, 13042, 23, 107, 30, 11, 25, 44, 14])"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"t = tls[0]; t[:20]"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"'xxbos xxmaj well , \" cube \" ( 1997 ) , xxmaj vincenzo \\'s first movie , was one of the most interesti'"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tls.decode(t)[:100]"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"xxbos xxmaj well , \" cube \" ( 1997 ) , xxmaj vincenzo 's first movie , was one of the most interesting and tricky ideas that xxmaj i 've ever seen when talking about movies . xxmaj they had just one scenery , a bunch of actors and a plot . xxmaj so , what made it so special were all the effective direction , great dialogs and a bizarre condition that characters had to deal like rats in a labyrinth . xxmaj his second movie , \" cypher \" ( 2002 ) , was all about its story , but it was n't so good as \" cube \" but here are the characters being tested like rats again . \n",
"\n",
" \" nothing \" is something very interesting and gets xxmaj vincenzo coming back to his ' cube days ' , locking the characters once again in a very different space with no time once more playing with the characters like playing with rats in an experience room . xxmaj but instead of a thriller sci - fi ( even some of the promotional teasers and trailers erroneous seemed like that ) , \" nothing \" is a loose and light comedy that for sure can be called a modern satire about our society and also about the intolerant world we 're living . xxmaj once again xxmaj xxunk amaze us with a great idea into a so small kind of thing . 2 actors and a blinding white scenario , that 's all you got most part of time and you do n't need more than that . xxmaj while \" cube \" is a claustrophobic experience and \" cypher \" confusing , \" nothing \" is completely the opposite but at the same time also desperate . \n",
"\n",
" xxmaj this movie proves once again that a smart idea means much more than just a millionaire budget . xxmaj of course that the movie fails sometimes , but its prime idea means a lot and offsets any flaws . xxmaj there 's nothing more to be said about this movie because everything is a brilliant surprise and a totally different experience that i had in movies since \" cube \" .\n"
]
}
],
"source": [
"tls.show(t)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"cut = int(len(files)*0.8)\n",
"splits = [list(range(cut)), list(range(cut,len(files)))]\n",
"tls = TfmdLists(files, [Tokenizer.from_folder(path), Numericalize], \n",
" splits=splits)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"tensor([ 2, 8, 20, 30, 87, 510, 1570, 12, 408, 379, 4196, 10, 8, 20, 30, 16, 13, 12216, 202, 509])"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tls.valid[0][:20]"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(#50000) ['pos','pos','pos','pos','pos','pos','pos','pos','pos','pos'...]"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lbls = files.map(parent_label)\n",
"lbls"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"((#2) ['neg','pos'], TensorCategory(1))"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"cat = Categorize()\n",
"cat.setup(lbls)\n",
"cat.vocab, cat(lbls[0])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"TensorCategory(1)"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tls_y = TfmdLists(files, [parent_label, Categorize()])\n",
"tls_y[0]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Datasets"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"x_tfms = [Tokenizer.from_folder(path), Numericalize]\n",
"y_tfms = [parent_label, Categorize()]\n",
"dsets = Datasets(files, [x_tfms, y_tfms])\n",
"x,y = dsets[0]\n",
"x[:20],y"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(tensor([ 2, 8, 20, 30, 87, 510, 1570, 12, 408, 379, 4196, 10, 8, 20, 30, 16, 13, 12216, 202, 509]),\n",
" TensorCategory(0))"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"x_tfms = [Tokenizer.from_folder(path), Numericalize]\n",
"y_tfms = [parent_label, Categorize()]\n",
"dsets = Datasets(files, [x_tfms, y_tfms], splits=splits)\n",
"x,y = dsets.valid[0]\n",
"x[:20],y"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"('xxbos xxmaj this movie had horrible lighting and terrible camera movements . xxmaj this movie is a jumpy horror flick with no meaning at all . xxmaj the slashes are totally fake looking . xxmaj it looks like some 17 year - old idiot wrote this movie and a 10 year old kid shot it . xxmaj with the worst acting you can ever find . xxmaj people are tired of knives . xxmaj at least move on to guns or fire . xxmaj it has almost exact lines from \" when a xxmaj stranger xxmaj calls \" . xxmaj with gruesome killings , only crazy people would enjoy this movie . xxmaj it is obvious the writer does n\\'t have kids or even care for them . i mean at show some mercy . xxmaj just to sum it up , this movie is a \" b \" movie and it sucked . xxmaj just for your own sake , do n\\'t even think about wasting your time watching this crappy movie .',\n",
" 'neg')"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"t = dsets.valid[0]\n",
"dsets.decode(t)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"dls = dsets.dataloaders(bs=64, before_batch=pad_input)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"tfms = [[Tokenizer.from_folder(path), Numericalize], [parent_label, Categorize]]\n",
"files = get_text_files(path, folders = ['train', 'test'])\n",
"splits = GrandparentSplitter(valid_name='test')(files)\n",
"dsets = Datasets(files, tfms, splits=splits)\n",
"dls = dsets.dataloaders(dl_type=SortedDL, before_batch=pad_input)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"path = untar_data(URLs.IMDB)\n",
"dls = DataBlock(\n",
" blocks=(TextBlock.from_folder(path),CategoryBlock),\n",
" get_y = parent_label,\n",
" get_items=partial(get_text_files, folders=['train', 'test']),\n",
" splitter=GrandparentSplitter(valid_name='test')\n",
").dataloaders(path)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Applying the mid-tier data API: SiamesePair"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from fastai2.vision.all import *\n",
"path = untar_data(URLs.PETS)\n",
"files = get_image_files(path/\"images\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class SiameseImage(Tuple):\n",
" def show(self, ctx=None, **kwargs): \n",
" img1,img2,same_breed = self\n",
" if not isinstance(img1, Tensor):\n",
" if img2.size != img1.size: img2 = img2.resize(img1.size)\n",
" t1,t2 = tensor(img1),tensor(img2)\n",
" t1,t2 = t1.permute(2,0,1),t2.permute(2,0,1)\n",
" else: t1,t2 = img1,img2\n",
" line = t1.new_zeros(t1.shape[0], t1.shape[1], 10)\n",
" return show_image(torch.cat([t1,line,t2], dim=2), \n",
" title=same_breed, ctx=ctx)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAASUAAAB6CAYAAAD5yEXhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOy9abBl2VXn91tr73Pu8Kacs8asQSWVpKqShIRoCZWwEGISjRp344Y23e5AbjrCPRmD3SbcER3tiG5/dgcQgcM4gmgMNHQwBRjaCBkxWQEITUgqqaSqylJlZdaQw8s33OGcvffyh7XveS+zBqAMTUG8HSFlZb773jv3nN9ew3+tta+YGUfraB2to/VqWfoXfQFH62gdraN1eB0ZpaN1tI7Wq2odGaWjdbSO1qtqHRmlo3W0jtarah0ZpaN1tI7Wq2odGaWjdbSO1qtqHRmlo3W0jtarah0ZpaP1giUie4f+V0Rkfujv3/UXfX1H66/2kqPmyaP1cktEzgP/wMx+/WVeE80s/ae7qqP1V3kdRUpH60+9RORfi8hPi8hPicgu8HdF5P8UkX916DXvqwZt9fc7ROTnReR5EXlCRP7xX8ClH62/BOvIKB2tV7r+c+AngS3gp1/uhSISgF8G/gC4Hfh64H8Qka/7877Io/WXbx0ZpaP1StfvmNkvmVkxs/kf89p3AJtm9r+YWWdmXwL+D+A7//wv82j9ZVvxL/oCjtZf2vXUn+K1dwHnRGT70L8F4CN/pld0tP5KrCOjdLRe6bq5QrIPTA/9/ZZD//0U8EUze8Of+1Udrb/06yh9O1p/VuuTwLeIyHERuRX4Z4e+9lGgE5HvF5GxiAQReUhE3vYXc6lH69W8jozS0fqzWj8GPAI8CfxH4N+vvlDbBd4PfBVwHrgM/G/A5n/qizxar/511Kd0tI7W0XpVraNI6WgdraP1qlpHRuloHa2j9apaR0bpaB2to/WqWkdG6WgdraP1qlpHRuloHa2j9apaL9s8uSYTy3HBODSIFkIQQJEWWEIusLQAkggEoiQUIYiSVBEx2pJIGjlc5RMroALWkCyjBgEBE7ImtCgaAqYFKLTWkrWQLREskMWgGJiSrSAiAAQF/5IgmikIFENVSX0hqH9NA0BBaDAyIsKorfZZwnCdIQTMhCSZaEKfE+NRJPf+Pe24RVByzpj47/YfocTRmGYyZTJdR0Q4feasX4fB1rFjmApShCw9ljI5GX23YDydIKmQrLDYnxEk0o5HaFPvoQhmRp+WUARVkCLM9vaZbm5h6sP6uS8s+wV50WEmTKdTigqNKMvUI8UgKKoKuTDreiLCrJsTi6JRSF2mxEg322d35yopC2E0RtsRsye/hKU5jQa+eHlvuGdTUXIolZlECAGRAI3BElKGDmem1YDYjcwEycRSsNCSsoGUF2UmYojpC5ghGGaZpjSUYGRLUATU7xtZKBhmzkVQ/N9NEc2YyfC1gRn87wUjEDEyIQhNvJEZEUFV/WdIQQwsF0KjlJxpmgaNiqCUUvzn4r9PgtJMpsTRmMl0ndFoxNax48SgdOVGZoomulmHiFC6Jc107MykxGKxIISGEALNeERKiRCdndl8jya0RAGzysyxTYwMgHWFeVogqdD3mel0Shy1WMosc0JyeQEzIsJyOaexgARIXSZboe97dneugkWsadDxmNkTX8TSHFXlsef35KXszstGSjZONLGhK8YyGbMlLPrCfNeY9ZllSaS8IGVjWTqWSehN2E9GnzJqSg4NFEMM1AIqkWIBy0AphOwPXERIISNFsWCklLAiiASSZEIBK26QxMAEtLhxEHOwCoZpQACKYMXfd7ZCCYKhmBZyMgricAmgghTD8J+btBokUbIWvz6DJrS+gZpAjG7PzQwlIAYRwdSwkun7JaVPqBhtOwJzQ6Mls+g75vu7lBjRLCwzSAxMp1OCRDIGKAEowei6BfvzPSxEJuMxmgOp62lMiBYIonRpyXI+p9ExsYzQ7F+L2rB/7RJXLl8iGfRWyKVAdmMuqqRUmO9cpaBYX0AS1kYsQlsM9X+imy3ou8xaM6L0M6QY+Sa0bKSHmBFmS5h3+RAz/cDMvE8sk9AVBmasBHJofKNSUAtgOjAjZoRsJGRgRi28LDMhhAOHQTlkPCozKIJvVPP9SbY0MCMi5OSGbGAGPWDGhKT4z6zMgDMjAYJEQnPg7MwMM+dWYGAmdfkFzBgyMLO32KPECJ24UY6BtbW1gRlTIQBtG1kmZyaMxgMz0QKNCcRmYEYyAzMUf03JeWCm7zK9FUrOg+EVsYEZy2C939MVM2JlYGZv9xp9l1m3ODDT/TFtSC/bpzRqo+lqwwv+wANoEcqh12mAnLN7nhAwM3LORFFijCRLSDHWRiMCRhaj9OYPWyGaQP0dYu6Z/M1Xg6OFIB4VOT5QMu5FzT2YA6BYLmgMqJh7DzH8UUFPIZpgWjckgbZxTxUUkAYRIys0REyglIL1GZ0osYDogR1XEXfC9R7l4h6w6xaoKsfPnGJr8zjt5gbrm8ex1DPb36G/8Dhta+zHLTbP3sF0c4sYAhoC43bCsl/QzfbZ3t4mXb/GeOcSUFhKg2yeYnrydnIQxpMJMbaU3LP92OdJV5+l2zrN5m3nWF87Ri5LlrM59tQX6Hf32E+ZbtRw8v63kFOim+0znWyys7MNly+xvbvN8XsfJCFo6VnkJfPnn6fPRhxPafeu0VNYaoPMF4yCkIDz12bDPWmbwM3MaPBo7mZmVuxpvac5Z9oQUVVy7hEJTNsIlkHlBmaUghal+h03YoeYEa2vvYmZ1etelBkrQxQlImB6IzMEj5Abqc/fbmAmijuqUoo74hZCOXh/Ujd1ETxCKoWUnb+uW9CuTdjY2GBr8zijzS3WNrduYGakxt54ixN3vZa2bYkh0LQtTWhZ9gtm16+xtz+je/4Sa/MrmMBMx8SN40xP3k4hM1rfYKwtnTkz/c4V+rXjbN52js2NY/Rpye61qzTPfZl+d48rqUNHY8688e0s9nZJ/ZJ2NGG2tw+XL3Ft/zon7n6APncEURZ5yfWnvow0Y+J4Sphfp+REDhPybI9REBbFeGp79pKR0ssbpaa1QnZrXp++qGEl+E6kpizVOIg6iCNxb7V6uNkEsUzRwLEmkOiw5K+nOEiZg59Xgkcu7sn9T7VQf0+gWI8JkP3BGoL5haBSPKWRwDIbjTkIVqGXCJINQiQoqIJqpIme2oXQUEpi0fVYLqS+oDGysd4gEhAxVAQ1ZXe5RESYxsjurKPvOlCHPYqydeY4m6dPs7V5nM1jx+hTYeexT3NKM51lmh52Uw/3PsTGsdM8e+lpUr9gb/say6uXWV+fstYnehGWOdW0DfYFjt1yG5effJolmVvvuItQOq5evEgBln1i/dhxJiePs//ssywWCyx5xCgGqNBlT0kLRgnQ1JQirzaOCUGNKOrBvQLFPXik0ObCaCR0+Uaj1DQNZitmAlD+WGYQoRFuYCYVEMtYiGxGpUiPJXVjlgFRSjVzZp6KHmZGcaezSv9W6RL1KLoigo/vCdHDeISGLvVEUaiSQM4ZixBuYiZqQ4hGMSFqQ7bEbLFEDFJfGE3GjEeeLospJRhSjL0uISJEaVgul3TLJRL8njRxzObJNTZPn+b4idOsr6+zs7tPf/GLnNLsBrIzrqXC9C3vZqTKhSe+hARhb/sa3fPPsLa1yTQlOqDPmZwzuSj7AutnzrL91DPkNnD6zK2E0nHt0iWyGcs+cfzsWeL6GteffIpEwZI78lAgK5RihAJJDNSZSVYwu5EZl1MEFBQjxhuZ2Vtmnt5ZvDKj1IRoVEgBVgZqeNimK1oH4yHFIIYa8ibIhRhd7JnQEKaRtEw1BXMwpOoklj3NiiZudCrAEt2zYAqSKNlD6mVJ1eMJvRXaEMEKo6BgyrLvKny+H2IMQ+pnZCQ0jIK/vxCERTIsmW/OGq5WuYHxeMy4hUYVCUrXF7Z3ZjUUr9cqQikHGtd0fcod585x+eoVlsuOsuxpLdGo0ZshRWjEWIYAsSHt7VfDJ0SF3iCI0hTog+8mK4KFgHQZicEjs9TTxBFajCR58OjJeiJKLgfp7SqyK8EjilIKEpRUClp8kw7/L4pRUHSYvm2bhkYDljPHJoFry57ndxaHmTmINFbMmD8vpKCmlBUzKwNiYEFfkhnTDBJekpkchaY4MyKuI97ITBki60Wp8ZoJqTITzIhBKBkWOQ3XvmJmiK4sobFlFKhambFIhiRIJdOXglYeJCpt2zIdtygZDcLePDGbLW5ixje7p5TCeDrhjnPnuPjMJcyg358xCUKjRlf8/jVi9KN1Col+PkOL37sgntYqQlMgxTzoZRYCpUvEGMk50+f0AmYAMqlGcX5/U817a/xQ77tv+1QKweoeeQlmrGRGozEBQcw4Ngk8M1tyfW/5yoxSGxsTtbrRrDq6iFnGVBwmCRjZxeUKWrTgoqB4mE4utEEwDYzFWKIIBQtu1Mz8Z5fimyYUKBRCCAeGSQ1VKFndGJWycoJYUHqDeGhw3UoZvDtVfwDIAsFcpNeQmTYT2iYwaSJX9hfs90vIocKdKKiH5dJh0nBiLbCXCsuF60MAOQhYFUnFaC0MRpVgNKJE8dQ2kNHq4YsGKAGxHlVlaUawqmMMHqdgJiuFiULGsutuqzdk6l7NBIIEsERRkAqrid+HgnlEqoLlg2TKBLK5V8MCSHYBGSNUIdbUr2tC5ti4oc/KUlxcvTY7OAm3jY0bgT8BM4aCeZovNQ1XjBAFNXNBVgMjETpcD8rika3kBBIo5VCEfYgZpRrBQ8zknIn196Tgnj1iiIIVZ8Y3343MJIyIkEtBgzFtJoxHgVFwZva6BVJivY5MQQkiiGSiNkw3WubLnn6RKo9AKGTz6DBbprXgBlxcrB9X56SqRCkoxe+bQC5K0GoQJHg0hzts1UgpCTH1/6b3Z5yVgJECZDMUsASooChLine truncated
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"img = PILImage.create(files[0])\n",
"s = SiameseImage(img, img, True)\n",
"s.show();"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"img1 = PILImage.create(files[1])\n",
"s1 = SiameseImage(img, img1, False)\n",
"s1.show();"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"s2 = Resize(224)(s1)\n",
"s2.show();"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def label_func(fname):\n",
" return re.match(r'^(.*)_\\d+.jpg$', fname.name).groups()[0]"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class SiameseTransform(Transform):\n",
" def __init__(self, files, label_func, splits):\n",
" self.labels = files.map(label_func).unique()\n",
" self.lbl2files = {l: L(f for f in files if label_func(f) == l) \n",
" for l in self.labels}\n",
" self.label_func = label_func\n",
" self.valid = {f: self._draw(f) for f in files[splits[1]]}\n",
" \n",
" def encodes(self, f):\n",
" f2,t = self.valid.get(f, self._draw(f))\n",
" img1,img2 = PILImage.create(f),PILImage.create(f2)\n",
" return SiameseImage(img1, img2, t)\n",
" \n",
" def _draw(self, f):\n",
" same = random.random() < 0.5\n",
" cls = self.label_func(f)\n",
" if not same: \n",
" cls = random.choice(L(l for l in self.labels if l != cls)) \n",
" return random.choice(self.lbl2files[cls]),same"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAASUAAAB6CAYAAAD5yEXhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOy8d3QUV7r1/TvnVHVQlshC5GgyGBsMmGBwwAkb5zg2xgmHcRjHcQCPs8c5x3FO2OCAE8YkAyaLJJKEUEAoZ6lj1TnfH9Vgz/3W+F3vrHvf63uX9lq9pO7qrj5dXWfX3s+zTwtjDG1oQxva8EeB/O8eQBva0IY2/BZtpNSGNrThD4U2UmpDG9rwh0IbKbWhDW34Q6GNlNrQhjb8odBGSm1oQxv+UGgjpTa0oQ1/KLSRUhv+fxBCtPzmpoUQ4d/cv+i/e3xt+N8N0RaebMPvQQhRBMw2xiz5nedYxhjn/92o2vC/GW1KqQ3/1xBCPCiE+EQI8ZEQohm4WAjxvhBi7m+eMy1BaIfu5wghFgohqoUQ+4UQ1/03DL0N/wPQRkpt+HdxJvAhkA588ntPFEIoYBGwAegKHA/cJoSY+l89yDb8z0MbKbXh38UqY8zXxhhtjAn/H547FkgzxjxsjIkZYwqAN4Hz/+uH2Yb/abD+uwfQhv+xKP2/eG4PoLsQouE3jylg+X/qiNrwvwJtpNSGfxf/sUPSCiT95n7n3/xfCuQbY474Lx9VG/7Ho82+teE/C1uAU4QQmUKILsCNv9n2CxATQtwqhAgIIZQQYqgQ4sj/nqG24Y+MNlJqw38W3gZ2AcXA98DHhzYk4gInA0cDRUAN8CqQ9v96kG3446Mtp9SGNrThD4U2pdSGNrThD4U2UmpDG9rwh0IbKbWhDW34Q6GNlNrQhjb8odBGSm1oQxv+UPjd8GSyCBrXihBQNkJqlBKARPiAKLgaokaBcFAoLOEgESghcaRECINPOzjS4rddPmE0SAHGxjEu0oBCgBG40kFqiVQKIzWg8RkfrtS4xkEZhSsMaANG4hqNEAIAJcHbJBDSRSNAG6SUOHGNkt42qQA0AhuDixACvy/Bz0IdHqdSCmMEjnCxjCDuOgT8Fm7ce40v4EMgcV0XI7z39nYhsfwB7GASwaQUhBB06NjJG4eB9IwMjBQILXBFHOO4uI4hHosQSAoiHI1jNJHWEEpY+AJ+pJ04hkJgjCHuREELpAShBaGWVpLS0jHSW6zvxjXReAQ3EsMYQVJSEloKbCGJOnGENqAkUkpwNaFYHAtBKBbG0hJpCZyYi7YsYqFWmpvqcFyB8geQPj+h4gKME8aWivyalsPHrGvHdjSnhBjcL5X6aID4wUpSOim6Dp7GtJMs4sGjyF/5Lm88sYtBY2Fkx7HsoJHP/z6LR+95ntaI5sLrLya5fTvOuOBOAllw7slDWLdhPz8+/zZXzbuR8qYaZl92E189+RE/lBez+YeX+XLVJ4zvPJkmU80bi7/gyI6jGN9jGFc98waR+iaceJzWRoc7/nYidn2Yq/98I4W1RWzLXcXGNTuoj1dz6vjL+WT5Ak4cPZmXPvieTskpxBtbKNdNJKdaNNSG+O7rL5n/48uM7ZPNCw+s56xZU/nsm6+Y//zPDJzUHbcmnb5dHQb0asdZl5/HL4s3cuWcq8jdsJg6y2Zz3iZ27Gxk/YZ8jj9rIge3bcUfaeXe2x7l09w3+GVpHvc8+BgPzXmcqtZWJk0WXHvVDVx++UtYqpXtqxQzrmlHs9XMnhUh3v54Dnef9iV74mVsXvYtN9x3J7vL8ulswmR0aE+VaeCZx+bx98feormmiZkXn8bZk8/jxy++Y79/LW+8sZbChcvoO+V4enVzEIEAvY/qzZb5Bzj1nAlMvnQyexbXsalgPq/O/ZmbLx9HQ2YTRx1zPO+/8QnnnnMha7bsRYhGkivS6HPEYHqc3p5Xb1+AlWVTE91HhupOwcYidB/FETk5zJx1Kjce/4L4V7yj5s6d+y9J6aHH5s21pMLRBlcb4q7ANQYnCnGtcdE42kEbgYuD0RKEIOqCMQZbKozHBHh0phBSoTVgDMIIhDZIpZAIHOkitQRl0I5GCG/SGAyWEThGYKRBGDAClKsxUnr3MSBBSAthDBwiJwwagxYCicRIjXEB6W1TAoyU2IARAqTElXhjFxJXaKSUSA1SWlgShBIoqRBSgAFhJGCwELjSgPHeU0mbQNCPzxcg4PfhCpBaYyyLeDSMCiYjHZeIq1GWwm/bSGHhaAeQmHgMbQm06xB1Y1j+JJL8ftyoIRoL4RcWSiikELSEmhFIAoFkpKswjusddyNorisjEotiJ6UCBsd1ka5BWAqpFE7MoaWxFtufQjwSRSkNAT9au/i0wInFiYXCRMJRUDbpySm0VBYjDRglqQ/FDp8znbJitMQVqX39nHbqOJZ/vYfW+jgPP3YvfTIcSg4epDhvLwfKWgnWCujYyPSpk9GpDYzoO5EXvv2O8ceNJBxvZsa0QQyeOZUkGhjVrRuTRvTjlRVL0AegQ6dUfli1haPH9CCnVw/eeOFTjj5lKB8v+BpfAMaPGkd5vImcdh1YtmIbxvVR73c46eSBqJBk+b61yFiAx2//gCD1VBiXvC1ljBubzdNPLSUcaaW0qJZpJ0xl+LDB7N+/F59M4ofvvqeuvohRg/rz4gvL+XrRKtykCL17t7Dw63W4JkBpRQPp7VIJOBnUVa6lZ88hVFXs5/s1qxhw5Hh6djIU761k6nGj2Zm7kX59uzN4yEiS0qL0OaIPS5Yupuagn/ZdXEIyRP6mZkrLG1AByd33XkRSph8n1Ex5scWE8dN487NF2BGbHZuXc/NfL+Suv/yVlMxWOnT3s/rnYk6aOZ57bvmC9B6K5gPNdO5tyMrOJBoq5JwLZiObyklKTabVNHLG+ePYs62AxnKX777aQ6yhkszkKMdOH0Fu7jqyKsLsDx/gqDEzuPbC61i47DNyl+2kpbWO2vIyquJlDBo9CuW6zJpzGqeNGctDj33HzFlTuf+iq6ipqGNw764M6XfivH+LlB544IG5RhuMMbh4ykgLMBpcI9BagJFICdrVGAFGSDSGuOPiuBptJBHtEnc0tmWhEuRhXO89XCW8k9sYQCDwRJCQCSJ1QePtW6MPqxHjgk6wk8YjKoREOy4oiZAeaXnbPNKLo1HGIz1PXSksWyK8l4KwEXiEJ72RYrTBRF3wC+8RIQDhrbEw4GLAaE+9uAI3bgiHIzgxh0BSgIAdwAoG8AWTwXVpbW2iefcWnLpSqiqqcJHYto1MjN9nB9DG4ETCNDY3E6qqwBzYjVtVQmNZMY0NDWhp4RiDsC1QHomFy4oJFe2itqkR7bPx+ZMwQhOPx/HVlkN1JY2l+6mvPog/syPReIyWhnpwBPV1tejKg1Ts340vtR3hmCbW0kJzqJma0iIaGurQQuIP1eOG6mloqIWYiy0lrjY0ROKHz5mPF99NUiBI7qZtLF+cx4kn9GffbsM7iz5gyLCuZPmaGTtuGjt2bqJDlwh7cmPU11SzYP5Kjp7SlytuvIS6ulK+/mYl6R0kyQfKaAlFWfH5ZuYt/JT6/Bay+6ZyzrCxfLlxE9h+Bo/qjb86Qs/+A1m8ehG9Wy2sQDtWLlnFxXfN4YJzjmZ7eT5zLj+bV17+mCxl8/HirfQdIDl2/BAGZw9hT1MDQ3sFqC5LZW/ZQXw+w9gB/dl6oIQDu3ZwYE8Lnfp1QdGLaFUzJ50xnB8X5/Pd968jwha9B49mzaKfGJnTnp+Xf8PVtz9L3+5JbNyzGxoruPepDfTr3JmrL7uQfSUHaWpq4b13lvPlJwvJ6Ozw9+dfIWTVsGlFMb1H9KAkP4/vv1jM3vx8LrvgWr788lOCSnHNbRcTlDkM6zqUc2dN4ot3l1FaUUJTjaEoXk9uwXIGZqfz2DOfMvWE0az/vpSJU0ey9OtNWKkhunWCwSMm8P2yhWzatZE/nX4t+5qb2VK0k8n9RnLxCTdw1TUvMn36cRx/ZgfWbm2lPtTM2eNOIXfDdwS6dmZXbYzVyxYy8sQhDEkaQ5G7g4FHTeGoCX0YP3Eay1b8xIi+R7Hio7dZ/NkvdB4aYv3iUp74230syPuYVUvWcsnZt/1LUvr9tW9G4uAiBBidIAljcI3Em0UJu6Q1AgUY4trFLwSWZWOMR2gYC3BpjmsybIUxMTxZo1GuAHFocnsT3SjPdgEgEwRiJEoARqFN3FvO6UqM1oDwCMrVSKE9kkARAWzt+SoDCFejLRCOwSgLKTWu6ykgKUEbjZQ2aIdQLIxxNU5cIy2LVJ+NEQqMQQqBNJLmSBQhBEmWRXMoRjwW88ZrwBKSWDhCNMUhoMHCEEfiVJTQwSeIGU1aqIHm/Gpaeg8lNaMDB0uLceIRWhrqidbVkJKSRHLcoUEIoq6LMQ7xulJai0vJ6JxNTXEZUVy65PRA6Rh19WF0fQnR/EJSMjIJtsuktbKSSCSCcfCsrglTWvkzMdezpBqDVmAj0VpTvnUzQgi0EShpsITEBWhuAm3w+fxYxPAZPBvs/vMps6W5kR83rme41ZN9OSWs3FCMTJEIN0BaVpgFn6/j6EkxKqItBLpnMP2co1n68WJGXzqYWL0gXhVhz/oC8otLeeLK27n+/jk0hdJoCafjy27hz9P60m3cLDYvXsSyb+7kxW920yM5jbpJx3DV44/y6u3n0xTqz/ljcug+pgczh1/Mw69fx1XnTieITbw1mbe+2M2QMR0YOnIim5YsZ9N3G6nJFDR068RHH23nulOPZk1lPZffdSazrn+YWK0mK7M9GfEUqlt3M2hoOkEV5NQJw5GBIDl9/KTqVCaOGszQlB6cdMIkOk/VXH3RBE64YC2vlO4gf8n3nHDDuZRXNFK3p4ULZx/N0iW7yN22gp/XbaasqYXZx05kzhUzOG/ybBaseorLrzmbyqZqGsvDXDt7Dvfc/w9ef/dtbr7Line truncated
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"splits = RandomSplitter()(files)\n",
"tfm = SiameseTransform(files, label_func, splits)\n",
"tfm(files[0]).show();"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"tls = TfmdLists(files, tfm, splits=splits)\n",
"show_at(tls.valid, 0);"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"dls = tls.dataloaders(after_item=[Resize(224), ToTensor], \n",
" after_batch=[IntToFloatTensor, Normalize.from_stats(*imagenet_stats)])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Becoming a deep learning practitioner"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"jupytext": {
"split_at_heading": true
},
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
clean/12_nlp_dive.ipynb
+1590
View File
File diff suppressed because it is too large. Load diff
clean/13_convolutions.ipynb
+2626
View File
File diff suppressed because it is too large. Load diff
clean/14_resnet.ipynb
+862
View File
@@ -0,0 +1,862 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"hide_input": false
},
"outputs": [],
"source": [
"#hide\n",
"from utils import *"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Resnets"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Going back to Imagenette"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def get_data(url, presize, resize):\n",
" path = untar_data(url)\n",
" return DataBlock(\n",
" blocks=(ImageBlock, CategoryBlock), get_items=get_image_files, \n",
" splitter=GrandparentSplitter(valid_name='val'),\n",
" get_y=parent_label, item_tfms=Resize(presize),\n",
" batch_tfms=[*aug_transforms(min_scale=0.5, size=resize),\n",
" Normalize.from_stats(*imagenet_stats)],\n",
" ).dataloaders(path, bs=128)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"dls = get_data(URLs.IMAGENETTE_160, 160, 128)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 432x432 with 4 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"dls.show_batch(max_n=4)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def avg_pool(x): return x.mean((2,3))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def block(ni, nf): return ConvLayer(ni, nf, stride=2)\n",
"def get_model():\n",
" return nn.Sequential(\n",
" block(3, 16),\n",
" block(16, 32),\n",
" block(32, 64),\n",
" block(64, 128),\n",
" block(128, 256),\n",
" nn.AdaptiveAvgPool2d(1),\n",
" Flatten(),\n",
" nn.Linear(256, dls.c))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def get_learner(m):\n",
" return Learner(dls, m, loss_func=nn.CrossEntropyLoss(), metrics=accuracy\n",
" ).to_fp16()\n",
"\n",
"learn = get_learner(get_model())"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/plain": [
"(0.47863011360168456, 3.981071710586548)"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"learn.lr_find()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>1.901582</td>\n",
" <td>2.155090</td>\n",
" <td>0.325350</td>\n",
" <td>00:07</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>1.559855</td>\n",
" <td>1.586795</td>\n",
" <td>0.507771</td>\n",
" <td>00:07</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>1.296350</td>\n",
" <td>1.295499</td>\n",
" <td>0.571720</td>\n",
" <td>00:07</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>1.144139</td>\n",
" <td>1.139257</td>\n",
" <td>0.639236</td>\n",
" <td>00:07</td>\n",
" </tr>\n",
" <tr>\n",
" <td>4</td>\n",
" <td>1.049770</td>\n",
" <td>1.092619</td>\n",
" <td>0.659108</td>\n",
" <td>00:07</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn.fit_one_cycle(5, 3e-3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Building a modern CNN: ResNet"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Skip-connections"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class ResBlock(Module):\n",
" def __init__(self, ni, nf):\n",
" self.convs = nn.Sequential(\n",
" ConvLayer(ni,nf),\n",
" ConvLayer(nf,nf, norm_type=NormType.BatchZero))\n",
" \n",
" def forward(self, x): return x + self.convs(x)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def _conv_block(ni,nf,stride):\n",
" return nn.Sequential(\n",
" ConvLayer(ni, nf, stride=stride),\n",
" ConvLayer(nf, nf, act_cls=None, norm_type=NormType.BatchZero))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class ResBlock(Module):\n",
" def __init__(self, ni, nf, stride=1):\n",
" self.convs = _conv_block(ni,nf,stride)\n",
" self.idconv = noop if ni==nf else ConvLayer(ni, nf, 1, act_cls=None)\n",
" self.pool = noop if stride==1 else nn.AvgPool2d(2, ceil_mode=True)\n",
"\n",
" def forward(self, x):\n",
" return F.relu(self.convs(x) + self.idconv(self.pool(x)))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def block(ni,nf): return ResBlock(ni, nf, stride=2)\n",
"learn = get_learner(get_model())"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>1.973174</td>\n",
" <td>1.845491</td>\n",
" <td>0.373248</td>\n",
" <td>00:08</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>1.678627</td>\n",
" <td>1.778713</td>\n",
" <td>0.439236</td>\n",
" <td>00:08</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>1.386163</td>\n",
" <td>1.596503</td>\n",
" <td>0.507261</td>\n",
" <td>00:08</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>1.177839</td>\n",
" <td>1.102993</td>\n",
" <td>0.644841</td>\n",
" <td>00:09</td>\n",
" </tr>\n",
" <tr>\n",
" <td>4</td>\n",
" <td>1.052435</td>\n",
" <td>1.038013</td>\n",
" <td>0.667771</td>\n",
" <td>00:09</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn.fit_one_cycle(5, 3e-3)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def block(ni, nf):\n",
" return nn.Sequential(ResBlock(ni, nf, stride=2), ResBlock(nf, nf))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>1.964076</td>\n",
" <td>1.864578</td>\n",
" <td>0.355159</td>\n",
" <td>00:12</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>1.636880</td>\n",
" <td>1.596789</td>\n",
" <td>0.502675</td>\n",
" <td>00:12</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>1.335378</td>\n",
" <td>1.304472</td>\n",
" <td>0.588535</td>\n",
" <td>00:12</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>1.089160</td>\n",
" <td>1.065063</td>\n",
" <td>0.663185</td>\n",
" <td>00:12</td>\n",
" </tr>\n",
" <tr>\n",
" <td>4</td>\n",
" <td>0.942904</td>\n",
" <td>0.963589</td>\n",
" <td>0.692739</td>\n",
" <td>00:12</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn = get_learner(get_model())\n",
"learn.fit_one_cycle(5, 3e-3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### A state-of-the-art ResNet"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def _resnet_stem(*sizes):\n",
" return [\n",
" ConvLayer(sizes[i], sizes[i+1], 3, stride = 2 if i==0 else 1)\n",
" for i in range(len(sizes)-1)\n",
" ] + [nn.MaxPool2d(kernel_size=3, stride=2, padding=1)]"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[ConvLayer(\n",
" (0): Conv2d(3, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
" (1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
" (2): ReLU()\n",
" ), ConvLayer(\n",
" (0): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
" (1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
" (2): ReLU()\n",
" ), ConvLayer(\n",
" (0): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
" (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
" (2): ReLU()\n",
" ), MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)]"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"_resnet_stem(3,32,32,64)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class ResNet(nn.Sequential):\n",
" def __init__(self, n_out, layers, expansion=1):\n",
" stem = _resnet_stem(3,32,32,64)\n",
" self.block_szs = [64, 64, 128, 256, 512]\n",
" for i in range(1,5): self.block_szs[i] *= expansion\n",
" blocks = [self._make_layer(*o) for o in enumerate(layers)]\n",
" super().__init__(*stem, *blocks,\n",
" nn.AdaptiveAvgPool2d(1), Flatten(),\n",
" nn.Linear(self.block_szs[-1], n_out))\n",
" \n",
" def _make_layer(self, idx, n_layers):\n",
" stride = 1 if idx==0 else 2\n",
" ch_in,ch_out = self.block_szs[idx:idx+2]\n",
" return nn.Sequential(*[\n",
" ResBlock(ch_in if i==0 else ch_out, ch_out, stride if i==0 else 1)\n",
" for i in range(n_layers)\n",
" ])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"rn = ResNet(dls.c, [2,2,2,2])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>1.673882</td>\n",
" <td>1.828394</td>\n",
" <td>0.413758</td>\n",
" <td>00:13</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>1.331675</td>\n",
" <td>1.572685</td>\n",
" <td>0.518217</td>\n",
" <td>00:13</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>1.087224</td>\n",
" <td>1.086102</td>\n",
" <td>0.650701</td>\n",
" <td>00:13</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>0.900428</td>\n",
" <td>0.968219</td>\n",
" <td>0.684331</td>\n",
" <td>00:12</td>\n",
" </tr>\n",
" <tr>\n",
" <td>4</td>\n",
" <td>0.760280</td>\n",
" <td>0.782558</td>\n",
" <td>0.757197</td>\n",
" <td>00:12</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn = get_learner(rn)\n",
"learn.fit_one_cycle(5, 3e-3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Bottleneck layers"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def _conv_block(ni,nf,stride):\n",
" return nn.Sequential(\n",
" ConvLayer(ni, nf//4, 1),\n",
" ConvLayer(nf//4, nf//4, stride=stride), \n",
" ConvLayer(nf//4, nf, 1, act_cls=None, norm_type=NormType.BatchZero))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"dls = get_data(URLs.IMAGENETTE_320, presize=320, resize=224)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"rn = ResNet(dls.c, [3,4,6,3], 4)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>1.613448</td>\n",
" <td>1.473355</td>\n",
" <td>0.514140</td>\n",
" <td>00:31</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>1.359604</td>\n",
" <td>2.050794</td>\n",
" <td>0.397452</td>\n",
" <td>00:31</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>1.253112</td>\n",
" <td>4.511735</td>\n",
" <td>0.387006</td>\n",
" <td>00:31</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>1.133450</td>\n",
" <td>2.575221</td>\n",
" <td>0.396178</td>\n",
" <td>00:31</td>\n",
" </tr>\n",
" <tr>\n",
" <td>4</td>\n",
" <td>1.054752</td>\n",
" <td>1.264525</td>\n",
" <td>0.613758</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>5</td>\n",
" <td>0.927930</td>\n",
" <td>2.670484</td>\n",
" <td>0.422675</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>6</td>\n",
" <td>0.838268</td>\n",
" <td>1.724588</td>\n",
" <td>0.528662</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>7</td>\n",
" <td>0.748289</td>\n",
" <td>1.180668</td>\n",
" <td>0.666497</td>\n",
" <td>00:31</td>\n",
" </tr>\n",
" <tr>\n",
" <td>8</td>\n",
" <td>0.688637</td>\n",
" <td>1.245039</td>\n",
" <td>0.650446</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>9</td>\n",
" <td>0.645530</td>\n",
" <td>1.053691</td>\n",
" <td>0.674904</td>\n",
" <td>00:31</td>\n",
" </tr>\n",
" <tr>\n",
" <td>10</td>\n",
" <td>0.593401</td>\n",
" <td>1.180786</td>\n",
" <td>0.676433</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>11</td>\n",
" <td>0.536634</td>\n",
" <td>0.879937</td>\n",
" <td>0.713885</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>12</td>\n",
" <td>0.479208</td>\n",
" <td>0.798356</td>\n",
" <td>0.741656</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>13</td>\n",
" <td>0.440071</td>\n",
" <td>0.600644</td>\n",
" <td>0.806879</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>14</td>\n",
" <td>0.402952</td>\n",
" <td>0.450296</td>\n",
" <td>0.858599</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>15</td>\n",
" <td>0.359117</td>\n",
" <td>0.486126</td>\n",
" <td>0.846369</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>16</td>\n",
" <td>0.313642</td>\n",
" <td>0.442215</td>\n",
" <td>0.861911</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>17</td>\n",
" <td>0.294050</td>\n",
" <td>0.485967</td>\n",
" <td>0.853503</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>18</td>\n",
" <td>0.270583</td>\n",
" <td>0.408566</td>\n",
" <td>0.875924</td>\n",
" <td>00:32</td>\n",
" </tr>\n",
" <tr>\n",
" <td>19</td>\n",
" <td>0.266003</td>\n",
" <td>0.411752</td>\n",
" <td>0.872611</td>\n",
" <td>00:33</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn = get_learner(rn)\n",
"learn.fit_one_cycle(20, 3e-3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Questionnaire"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Further research"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"s"
]
}
],
"metadata": {
"jupytext": {
"split_at_heading": true
},
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
clean/15_arch_details.ipynb
+421
View File
@@ -0,0 +1,421 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"#hide\n",
"from utils import *"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Application architectures deep dive"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Computer vision"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### cnn_learner"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"{'cut': -2,\n",
" 'split': <function fastai2.vision.learner._resnet_split(m)>,\n",
" 'stats': ([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])}"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"model_meta[resnet50]"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Sequential(\n",
" (0): AdaptiveConcatPool2d(\n",
" (ap): AdaptiveAvgPool2d(output_size=1)\n",
" (mp): AdaptiveMaxPool2d(output_size=1)\n",
" )\n",
" (1): full: False\n",
" (2): BatchNorm1d(20, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
" (3): Dropout(p=0.25, inplace=False)\n",
" (4): Linear(in_features=20, out_features=512, bias=False)\n",
" (5): ReLU(inplace=True)\n",
" (6): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
" (7): Dropout(p=0.5, inplace=False)\n",
" (8): Linear(in_features=512, out_features=2, bias=False)\n",
")"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"create_head(20,2)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### unet_learner"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### A Siamese network"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"#hide\n",
"from fastai2.vision.all import *\n",
"path = untar_data(URLs.PETS)\n",
"files = get_image_files(path/\"images\")\n",
"\n",
"class SiameseImage(Tuple):\n",
" def show(self, ctx=None, **kwargs): \n",
" img1,img2,same_breed = self\n",
" if not isinstance(img1, Tensor):\n",
" if img2.size != img1.size: img2 = img2.resize(img1.size)\n",
" t1,t2 = tensor(img1),tensor(img2)\n",
" t1,t2 = t1.permute(2,0,1),t2.permute(2,0,1)\n",
" else: t1,t2 = img1,img2\n",
" line = t1.new_zeros(t1.shape[0], t1.shape[1], 10)\n",
" return show_image(torch.cat([t1,line,t2], dim=2), \n",
" title=same_breed, ctx=ctx)\n",
" \n",
"def label_func(fname):\n",
" return re.match(r'^(.*)_\\d+.jpg$', fname.name).groups()[0]\n",
"\n",
"class SiameseTransform(Transform):\n",
" def __init__(self, files, label_func, splits):\n",
" self.labels = files.map(label_func).unique()\n",
" self.lbl2files = {l: L(f for f in files if label_func(f) == l) for l in self.labels}\n",
" self.label_func = label_func\n",
" self.valid = {f: self._draw(f) for f in files[splits[1]]}\n",
" \n",
" def encodes(self, f):\n",
" f2,t = self.valid.get(f, self._draw(f))\n",
" img1,img2 = PILImage.create(f),PILImage.create(f2)\n",
" return SiameseImage(img1, img2, t)\n",
" \n",
" def _draw(self, f):\n",
" same = random.random() < 0.5\n",
" cls = self.label_func(f)\n",
" if not same: cls = random.choice(L(l for l in self.labels if l != cls)) \n",
" return random.choice(self.lbl2files[cls]),same\n",
" \n",
"splits = RandomSplitter()(files)\n",
"tfm = SiameseTransform(files, label_func, splits)\n",
"tls = TfmdLists(files, tfm, splits=splits)\n",
"dls = tls.dataloaders(after_item=[Resize(224), ToTensor], \n",
" after_batch=[IntToFloatTensor, Normalize.from_stats(*imagenet_stats)])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class SiameseModel(Module):\n",
" def __init__(self, encoder, head):\n",
" self.encoder,self.head = encoder,head\n",
" \n",
" def forward(self, x1, x2):\n",
" ftrs = torch.cat([self.encoder(x1), self.encoder(x2)], dim=1)\n",
" return self.head(ftrs)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"encoder = create_body(resnet34, cut=-2)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"head = create_head(512*4, 2, ps=0.5)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"model = SiameseModel(encoder, head)"
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {},
"outputs": [],
"source": [
"def loss_func(out, targ):\n",
" return nn.CrossEntropyLoss()(out, targ.long())"
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [],
"source": [
"def siamese_splitter(model):\n",
" return [params(model.encoder), params(model.head)]"
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [],
"source": [
"learn = Learner(dls, model, loss_func=loss_func, \n",
" splitter=siamese_splitter, metrics=accuracy)\n",
"learn.freeze()"
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {
"scrolled": true
},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>0.367015</td>\n",
" <td>0.281242</td>\n",
" <td>0.885656</td>\n",
" <td>00:26</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>0.307688</td>\n",
" <td>0.214721</td>\n",
" <td>0.915426</td>\n",
" <td>00:26</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>0.275221</td>\n",
" <td>0.170615</td>\n",
" <td>0.936401</td>\n",
" <td>00:26</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>0.223771</td>\n",
" <td>0.159633</td>\n",
" <td>0.943843</td>\n",
" <td>00:26</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn.fit_one_cycle(4, 3e-3)"
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>0.212744</td>\n",
" <td>0.159033</td>\n",
" <td>0.944520</td>\n",
" <td>00:35</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>0.201893</td>\n",
" <td>0.159615</td>\n",
" <td>0.942490</td>\n",
" <td>00:35</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>0.204606</td>\n",
" <td>0.152338</td>\n",
" <td>0.945196</td>\n",
" <td>00:36</td>\n",
" </tr>\n",
" <tr>\n",
" <td>3</td>\n",
" <td>0.213203</td>\n",
" <td>0.148346</td>\n",
" <td>0.947903</td>\n",
" <td>00:36</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn.unfreeze()\n",
"learn.fit_one_cycle(4, slice(1e-6,1e-4))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Natural language processing"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Tabular"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Wrapping up architectures"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Questionnaire"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Further research"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"jupytext": {
"split_at_heading": true
},
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
clean/16_accel_sgd.ipynb
+719
View File
@@ -0,0 +1,719 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"hide_input": false
},
"outputs": [],
"source": [
"#hide\n",
"from utils import *"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# The training process"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Let's start with SGD"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def get_data(url, presize, resize):\n",
" path = untar_data(url)\n",
" return DataBlock(\n",
" blocks=(ImageBlock, CategoryBlock), get_items=get_image_files, \n",
" splitter=GrandparentSplitter(valid_name='val'),\n",
" get_y=parent_label, item_tfms=Resize(presize),\n",
" batch_tfms=[*aug_transforms(min_scale=0.5, size=resize),\n",
" Normalize.from_stats(*imagenet_stats)],\n",
" ).dataloaders(path, bs=128)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"dls = get_data(URLs.IMAGENETTE_160, 160, 128)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def get_learner(**kwargs):\n",
" return cnn_learner(dls, resnet34, pretrained=False,\n",
" metrics=accuracy, **kwargs).to_fp16()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>2.571932</td>\n",
" <td>2.685040</td>\n",
" <td>0.322548</td>\n",
" <td>00:11</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>1.904674</td>\n",
" <td>1.852589</td>\n",
" <td>0.437452</td>\n",
" <td>00:11</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>1.586909</td>\n",
" <td>1.374908</td>\n",
" <td>0.594904</td>\n",
" <td>00:11</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn = get_learner()\n",
"learn.fit_one_cycle(3, 0.003)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"learn = get_learner(opt_func=SGD)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/plain": [
"(0.017378008365631102, 3.019951861915615e-07)"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"learn.lr_find()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>2.969412</td>\n",
" <td>2.214596</td>\n",
" <td>0.242038</td>\n",
" <td>00:09</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>2.442730</td>\n",
" <td>1.845950</td>\n",
" <td>0.362548</td>\n",
" <td>00:09</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>2.157159</td>\n",
" <td>1.741143</td>\n",
" <td>0.408917</td>\n",
" <td>00:09</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn.fit_one_cycle(3, 0.03, moms=(0,0,0))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## A generic optimizer"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def sgd_cb(p, lr, **kwargs): p.data.add_(-lr, p.grad.data)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"opt_func = partial(Optimizer, cbs=[sgd_step])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>2.730918</td>\n",
" <td>2.009971</td>\n",
" <td>0.332739</td>\n",
" <td>00:09</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>2.204893</td>\n",
" <td>1.747202</td>\n",
" <td>0.441529</td>\n",
" <td>00:09</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>1.875621</td>\n",
" <td>1.684515</td>\n",
" <td>0.445350</td>\n",
" <td>00:09</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn = get_learner(opt_func=opt_func)\n",
"learn.fit(3, 0.03)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Momentum"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"hide_input": true
},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"x = np.linspace(-4, 4, 100)\n",
"y = 1 - (x/3) ** 2\n",
"x1 = x + np.random.randn(100) * 0.1\n",
"y1 = y + np.random.randn(100) * 0.1\n",
"plt.scatter(x1,y1)\n",
"idx = x1.argsort()\n",
"beta,avg,res = 0.7,0,[]\n",
"for i in idx:\n",
" avg = beta * avg + (1-beta) * y1[i]\n",
" res.append(avg/(1-beta**(i+1)))\n",
"plt.plot(x1[idx],np.array(res), color='red');"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"hide_input": true
},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 864x576 with 4 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"x = np.linspace(-4, 4, 100)\n",
"y = 1 - (x/3) ** 2\n",
"x1 = x + np.random.randn(100) * 0.1\n",
"y1 = y + np.random.randn(100) * 0.1\n",
"_,axs = plt.subplots(2,2, figsize=(12,8))\n",
"betas = [0.5,0.7,0.9,0.99]\n",
"idx = x1.argsort()\n",
"for beta,ax in zip(betas, axs.flatten()):\n",
" ax.scatter(x1,y1)\n",
" avg,res = 0,[]\n",
" for i in idx:\n",
" avg = beta * avg + (1-beta) * y1[i]\n",
" res.append(avg)#/(1-beta**(i+1)))\n",
" ax.plot(x1[idx],np.array(res), color='red');\n",
" ax.set_title(f'beta={beta}')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def average_grad(p, mom, grad_avg=None, **kwargs):\n",
" if grad_avg is None: grad_avg = torch.zeros_like(p.grad.data)\n",
" return {'grad_avg': grad_avg*mom + p.grad.data}"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def momentum_step(p, lr, grad_avg, **kwargs): p.data.add_(-lr, grad_avg)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"opt_func = partial(Optimizer, cbs=[average_grad,momentum_step], mom=0.9)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>2.856000</td>\n",
" <td>2.493429</td>\n",
" <td>0.246115</td>\n",
" <td>00:10</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>2.504205</td>\n",
" <td>2.463813</td>\n",
" <td>0.348280</td>\n",
" <td>00:10</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>2.187387</td>\n",
" <td>1.755670</td>\n",
" <td>0.418853</td>\n",
" <td>00:10</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn = get_learner(opt_func=opt_func)\n",
"learn.fit_one_cycle(3, 0.03)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 864x288 with 2 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"learn.recorder.plot_sched()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## RMSProp"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def average_sqr_grad(p, sqr_mom, sqr_avg=None, **kwargs):\n",
" if sqr_avg is None: sqr_avg = torch.zeros_like(p.grad.data)\n",
" return {'sqr_avg': sqr_avg*sqr_mom + p.grad.data**2}"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def rms_prop_step(p, lr, sqr_avg, eps, grad_avg=None, **kwargs):\n",
" denom = sqr_avg.sqrt().add_(eps)\n",
" p.data.addcdiv_(-lr, p.grad, denom)\n",
"\n",
"opt_func = partial(Optimizer, cbs=[average_sqr_grad,rms_prop_step],\n",
" sqr_mom=0.99, eps=1e-7)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>accuracy</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>2.766912</td>\n",
" <td>1.845900</td>\n",
" <td>0.402548</td>\n",
" <td>00:11</td>\n",
" </tr>\n",
" <tr>\n",
" <td>1</td>\n",
" <td>2.194586</td>\n",
" <td>1.510269</td>\n",
" <td>0.504459</td>\n",
" <td>00:11</td>\n",
" </tr>\n",
" <tr>\n",
" <td>2</td>\n",
" <td>1.869099</td>\n",
" <td>1.447939</td>\n",
" <td>0.544968</td>\n",
" <td>00:11</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"learn = get_learner(opt_func=opt_func)\n",
"learn.fit_one_cycle(3, 0.003)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Adam"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Decoupled weight_decay"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Callbacks"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Creating a callback"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class ModelReseter(Callback):\n",
" def begin_train(self): self.model.reset()\n",
" def begin_validate(self): self.model.reset()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class RNNRegularizer(Callback):\n",
" def __init__(self, alpha=0., beta=0.): self.alpha,self.beta = alpha,beta\n",
"\n",
" def after_pred(self):\n",
" self.raw_out,self.out = self.pred[1],self.pred[2]\n",
" self.learn.pred = self.pred[0]\n",
"\n",
" def after_loss(self):\n",
" if not self.training: return\n",
" if self.alpha != 0.:\n",
" self.learn.loss += self.alpha * self.out[-1].float().pow(2).mean()\n",
" if self.beta != 0.:\n",
" h = self.raw_out[-1]\n",
" if len(h)>1:\n",
" self.learn.loss += self.beta * (h[:,1:] - h[:,:-1]\n",
" ).float().pow(2).mean()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Callback ordering and exceptions"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class TerminateOnNaNCallback(Callback):\n",
" run_before=Recorder\n",
" def after_batch(self):\n",
" if torch.isinf(self.loss) or torch.isnan(self.loss):\n",
" raise CancelFitException"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Questionnaire"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Further research"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Foundations of Deep Learning: Wrap up"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"jupytext": {
"split_at_heading": true
},
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
clean/17_foundations.ipynb
+1565
View File
File diff suppressed because it is too large. Load diff
clean/18_CAM.ipynb
+462
View File
@@ -0,0 +1,462 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"hide_input": false
},
"outputs": [],
"source": [
"#hide\n",
"from utils import *"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# CNN interpretation with CAM"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## CAM and hooks"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>error_rate</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>0.181760</td>\n",
" <td>0.032238</td>\n",
" <td>0.009472</td>\n",
" <td>00:14</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: left;\">\n",
" <th>epoch</th>\n",
" <th>train_loss</th>\n",
" <th>valid_loss</th>\n",
" <th>error_rate</th>\n",
" <th>time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <td>0</td>\n",
" <td>0.059119</td>\n",
" <td>0.014090</td>\n",
" <td>0.002706</td>\n",
" <td>00:18</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"path = untar_data(URLs.PETS)/'images'\n",
"def is_cat(x): return x[0].isupper()\n",
"dls = ImageDataLoaders.from_name_func(\n",
" path, get_image_files(path), valid_pct=0.2, seed=42,\n",
" label_func=is_cat, item_tfms=Resize(224))\n",
"learn = cnn_learner(dls, resnet34, metrics=error_rate)\n",
"learn.fine_tune(1)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"img = PILImage.create('images/chapter1_cat_example.jpg')\n",
"x, = first(dls.test_dl([img]))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class Hook():\n",
" def hook_func(self, m, i, o): self.stored = o.detach().clone()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"hook_output = Hook()\n",
"hook = learn.model[0].register_forward_hook(hook_output.hook_func)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"with torch.no_grad(): output = learn.model.eval()(x)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"act = hook_output.stored[0]"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"tensor([[2.7374e-09, 1.0000e+00]], device='cuda:5')"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"F.softmax(output, dim=-1)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(#2) [False,True]"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"dls.vocab"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"torch.Size([1, 3, 224, 224])"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"x.shape"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"torch.Size([2, 7, 7])"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"cam_map = torch.einsum('ck,kij->cij', learn.model[1][-1].weight, act)\n",
"cam_map.shape"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"x_dec = TensorImage(dls.train.decode((x,))[0][0])\n",
"_,ax = plt.subplots()\n",
"x_dec.show(ctx=ax)\n",
"ax.imshow(cam_map[0].detach().cpu(), alpha=0.6, extent=(0,224,224,0),\n",
" interpolation='bilinear', cmap='magma');"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"hook.remove()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class Hook():\n",
" def __init__(self, m):\n",
" self.hook = m.register_forward_hook(self.hook_func) \n",
" def hook_func(self, m, i, o): self.stored = o.detach().clone()\n",
" def __enter__(self, *args): return self\n",
" def __exit__(self, *args): self.hook.remove()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"with Hook(learn.model[0]) as hook:\n",
" with torch.no_grad(): output = learn.model.eval()(x.cuda())\n",
" act = hook.stored"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Gradient CAM"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"class HookBwd():\n",
" def __init__(self, m):\n",
" self.hook = m.register_backward_hook(self.hook_func) \n",
" def hook_func(self, m, gi, go): self.stored = go[0].detach().clone()\n",
" def __enter__(self, *args): return self\n",
" def __exit__(self, *args): self.hook.remove()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"cls = 1\n",
"with HookBwd(learn.model[0]) as hookg:\n",
" with Hook(learn.model[0]) as hook:\n",
" output = learn.model.eval()(x.cuda())\n",
" act = hook.stored\n",
" output[0,cls].backward()\n",
" grad = hookg.stored"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"w = grad[0].mean(dim=[1,2], keepdim=True)\n",
"cam_map = (w * act[0]).sum(0)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"_,ax = plt.subplots()\n",
"x_dec.show(ctx=ax)\n",
"ax.imshow(cam_map.detach().cpu(), alpha=0.6, extent=(0,224,224,0),\n",
" interpolation='bilinear', cmap='magma');"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"with HookBwd(learn.model[0][-2]) as hookg:\n",
" with Hook(learn.model[0][-2]) as hook:\n",
" output = learn.model.eval()(x.cuda())\n",
" act = hook.stored\n",
" output[0,cls].backward()\n",
" grad = hookg.stored"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"w = grad[0].mean(dim=[1,2], keepdim=True)\n",
"cam_map = (w * act[0]).sum(0)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"_,ax = plt.subplots()\n",
"x_dec.show(ctx=ax)\n",
"ax.imshow(cam_map.detach().cpu(), alpha=0.6, extent=(0,224,224,0),\n",
" interpolation='bilinear', cmap='magma');"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Questionnaire"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Further research"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"jupytext": {
"split_at_heading": true
},
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
clean/19_learner.ipynb
+1319
View File
File diff suppressed because it is too large. Load diff
clean/20_conclusion.ipynb
+42
View File
@@ -0,0 +1,42 @@
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Concluding thoughts"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"jupytext": {
"split_at_heading": true
},
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
clean/app_blog.ipynb
+95
View File
@@ -0,0 +1,95 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"#hide\n",
"from utils import *\n",
"from fastai2.vision.widgets import *"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Creating a blog"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Blogging with GitHub Pages"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Creating the repository"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Setting up your homepage"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Creating posts"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Synchronizing GitHub and your computer"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Jupyter for blogging"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"jupytext": {
"split_at_heading": true
},
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
clean/app_jupyter.ipynb
+273
View File
@@ -0,0 +1,273 @@
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Appendix: Jupyter notebook 101"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Introduction"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"2"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"1+1"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Writing"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"1.5"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"3/2"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Modes"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Other Important Considerations"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Markdown formatting\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Italics, Bold, Strikethrough, Inline, Blockquotes and Links"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Headings"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Lists"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Code Capabilities"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# Import necessary libraries\n",
"from fastai2.vision.all import * \n",
"import matplotlib.pyplot as plt"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from PIL import Image"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(1, 2, 4, 8)"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"a = 1\n",
"b = a + 1\n",
"c = b + a + 1\n",
"d = c + b + a + 1\n",
"a, b, c ,d"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plt.plot([a,b,c,d])\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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 truncated
"text/plain": [
"<PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=394x500 at 0x7F689F7B16D0>"
]
},
"execution_count": null,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"Image.open('images/chapter1_cat_example.jpg')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Running the app locally"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Creating a notebook"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Shortcuts and tricks"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Command Mode Shortcuts"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Cell Tricks"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Line Magics"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"56.1 µs ± 592 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n"
]
}
],
"source": [
"%timeit [i+1 for i in range(1000)]"
]
}
],
"metadata": {
"jupytext": {
"split_at_heading": true
},
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
clean/images
Symlink
+1
View File
@@ -0,0 +1 @@
../images
images/0_jupyter.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 41 KiB

After

Width:  |  Height:  |  Size: 63 KiB

images/Dropout.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 102 KiB

After

Width:  |  Height:  |  Size: 143 KiB

images/Dropout1.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 99 KiB

After

Width:  |  Height:  |  Size: 138 KiB

images/LSTM.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 71 KiB

After

Width:  |  Height:  |  Size: 112 KiB

images/att_00000.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 68 KiB

After

Width:  |  Height:  |  Size: 108 KiB

images/att_00001.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 48 KiB

After

Width:  |  Height:  |  Size: 98 KiB

images/att_00002.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 180 KiB

After

Width:  |  Height:  |  Size: 244 KiB

images/att_00003.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 808 B

After

Width:  |  Height:  |  Size: 1.0 KiB

images/att_00004.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 28 KiB

After

Width:  |  Height:  |  Size: 37 KiB

images/att_00005.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 664 B

After

Width:  |  Height:  |  Size: 884 B

images/att_00006.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 33 KiB

After

Width:  |  Height:  |  Size: 45 KiB

images/att_00007.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 180 KiB

After

Width:  |  Height:  |  Size: 244 KiB

images/att_00008.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 808 B

After

Width:  |  Height:  |  Size: 1.0 KiB

images/att_00009.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 28 KiB

After

Width:  |  Height:  |  Size: 37 KiB

images/att_00010.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 664 B

After

Width:  |  Height:  |  Size: 884 B

images/att_00011.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 33 KiB

After

Width:  |  Height:  |  Size: 45 KiB

images/att_00012.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 444 KiB

After

Width:  |  Height:  |  Size: 685 KiB

images/att_00013.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 155 KiB

After

Width:  |  Height:  |  Size: 371 KiB

images/att_00014.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 585 KiB

After

Width:  |  Height:  |  Size: 819 KiB

images/att_00015.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 623 KiB

After

Width:  |  Height:  |  Size: 911 KiB

images/att_00016.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 350 KiB

After

Width:  |  Height:  |  Size: 471 KiB

images/att_00017.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 180 KiB

After

Width:  |  Height:  |  Size: 295 KiB

images/att_00018.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 37 KiB

After

Width:  |  Height:  |  Size: 70 KiB

images/att_00019.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 60 KiB

After

Width:  |  Height:  |  Size: 111 KiB

images/att_00020.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 13 KiB

After

Width:  |  Height:  |  Size: 57 KiB

images/att_00021.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 5.7 KiB

After

Width:  |  Height:  |  Size: 12 KiB

images/att_00022.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 19 KiB

After

Width:  |  Height:  |  Size: 82 KiB

images/att_00023.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 8.2 KiB

After

Width:  |  Height:  |  Size: 42 KiB

images/att_00024.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 7.9 KiB

After

Width:  |  Height:  |  Size: 38 KiB

images/att_00025.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 30 KiB

After

Width:  |  Height:  |  Size: 73 KiB

images/att_00026.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 15 KiB

After

Width:  |  Height:  |  Size: 62 KiB

images/att_00027.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 67 KiB

After

Width:  |  Height:  |  Size: 102 KiB

images/att_00028.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 69 KiB

After

Width:  |  Height:  |  Size: 106 KiB

images/att_00029.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 111 KiB

After

Width:  |  Height:  |  Size: 161 KiB

images/att_00030.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 86 KiB

After

Width:  |  Height:  |  Size: 124 KiB

images/att_00031.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 55 KiB

After

Width:  |  Height:  |  Size: 194 KiB

images/att_00032.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 5.4 KiB

After

Width:  |  Height:  |  Size: 7.8 KiB

images/att_00033.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 3.5 KiB

After

Width:  |  Height:  |  Size: 4.8 KiB

images/att_00034.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 3.2 KiB

After

Width:  |  Height:  |  Size: 4.4 KiB

images/att_00035.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 51 KiB

After

Width:  |  Height:  |  Size: 83 KiB

images/att_00036.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 33 KiB

After

Width:  |  Height:  |  Size: 56 KiB

images/att_00037.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 50 KiB

After

Width:  |  Height:  |  Size: 69 KiB

images/att_00038.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 121 KiB

After

Width:  |  Height:  |  Size: 177 KiB

images/att_00039.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 41 KiB

After

Width:  |  Height:  |  Size: 59 KiB

images/att_00040.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 22 KiB

After

Width:  |  Height:  |  Size: 43 KiB

images/att_00041.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 89 KiB

After

Width:  |  Height:  |  Size: 133 KiB

images/att_00042.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 141 KiB

After

Width:  |  Height:  |  Size: 170 KiB

images/att_00043.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 16 KiB

After

Width:  |  Height:  |  Size: 22 KiB

images/att_00044.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 653 KiB

After

Width:  |  Height:  |  Size: 1.1 MiB

images/att_00045.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 84 KiB

After

Width:  |  Height:  |  Size: 110 KiB

images/att_00046.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 29 KiB

After

Width:  |  Height:  |  Size: 43 KiB

images/att_00047.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 15 KiB

After

Width:  |  Height:  |  Size: 35 KiB

images/att_00048.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 10 KiB

After

Width:  |  Height:  |  Size: 28 KiB

images/att_00049.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 27 KiB

After

Width:  |  Height:  |  Size: 38 KiB

images/att_00050.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 21 KiB

After

Width:  |  Height:  |  Size: 55 KiB

images/att_00051.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 61 KiB

After

Width:  |  Height:  |  Size: 83 KiB

Loaded 100 of 196 files, more files were not shown because too many files have changed in this diff. Show more