Anurag1734/cuda-error-resolution-analysis
07
1[2 {3 "post_stream": {4 "posts": [5 {6 "id": 263210,7 "name": "Chai",8 "username": "Chai",9 "avatar_template": "/user_avatar/discuss.pytorch.org/chai/{size}/33705_2.png",10 "created_at": "2021-02-11T21:40:57.040Z",11 "cooked": "<p>Hey,</p>\n<p>I’m trying to do an anomaly detection on an univariate time series with a LSTM autoencoder. E.g. I have a curve like this and the LSTM autoencoder learns everything perfectly except a small part where it seems that it hasn’t learnt anything. In the graph you see the red area which is learnt very bad - maybe you guys have some hints for me that I’m able to improve it?</p>\n<p><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https://discuss.pytorch.org/uploads/default/original/3X/0/d/0d2d23aa3b14e68c214e5721f0b51feb5f5fc334.png\" data-download-href=\"https://discuss.pytorch.org/uploads/default/0d2d23aa3b14e68c214e5721f0b51feb5f5fc334\" title=\"image\"><img src=\"https://discuss.pytorch.org/uploads/default/optimized/3X/0/d/0d2d23aa3b14e68c214e5721f0b51feb5f5fc334_2_690x324.png\" alt=\"image\" data-base62-sha1=\"1SyUTsRjXyjqZPdolUbVcFvFpfC\" width=\"690\" height=\"324\" srcset=\"https://discuss.pytorch.org/uploads/default/optimized/3X/0/d/0d2d23aa3b14e68c214e5721f0b51feb5f5fc334_2_690x324.png, https://discuss.pytorch.org/uploads/default/optimized/3X/0/d/0d2d23aa3b14e68c214e5721f0b51feb5f5fc334_2_1035x486.png 1.5x, https://discuss.pytorch.org/uploads/default/optimized/3X/0/d/0d2d23aa3b14e68c214e5721f0b51feb5f5fc334_2_1380x648.png 2x\" data-dominant-color=\"FDF6F6\"><div class=\"meta\"><svg class=\"fa d-icon d-icon-far-image svg-icon\" aria-hidden=\"true\"><use href=\"#far-image\"></use></svg><span class=\"filename\">image</span><span class=\"informations\">1700×799 26.8 KB</span><svg class=\"fa d-icon d-icon-discourse-expand svg-icon\" aria-hidden=\"true\"><use href=\"#discourse-expand\"></use></svg></div></a></div></p>\n<p>This is the architecture which I’m using</p>\n<pre><code class=\"lang-auto\">class Encoder(nn.Module):\n def __init__(self):\n super(Encoder, self).__init__()\n\n self.device = get_device()\n\n self.num_features = 1 #Features\n self.num_seq_length = 180 #Sequence\n self.num_directions = 2 #BiDirectional\n self.bidirectional = True if self.num_directions > 1 else False \n self.num_hidden_states = 2 #LSTM H,C\n self.num_hidden_dim = 16\n self.num_enc_hidden_dim = 32 #Dimension\n self.num_layers = 3 #Layers\n self.dropout = 0.2 #Dropout\n self.linear_dims = (self.num_enc_hidden_dim * self.num_directions, 1) #Dimension Linear\n self.batch_size = 10\n\n self.lstm_enc1 = nn.LSTM(\n input_size=self.num_features, \n hidden_size=self.num_hidden_dim, \n dropout=self.dropout,\n num_layers=self.num_layers, \n batch_first=True,\n bidirectional=self.bidirectional \n )\n\n self.lstm_enc2 = nn.LSTM(\n input_size=self.num_hidden_dim*self.num_directions, \n hidden_size=self.num_enc_hidden_dim, \n dropout=self.dropout,\n num_layers=self.num_layers, \n batch_first=True,\n bidirectional=self.bidirectional \n )\n\n self.lstm_enc1.apply(self.init_weights)\n self.lstm_enc2.apply(self.init_weights)\n\n def init_weights(self, m):\n if type(m) == nn.LSTM:\n for name, param in m.named_parameters():\n if 'bias' in name:\n nn.init.constant(param, 0.01)\n elif 'weight' in name:\n nn.init.xavier_normal(param) #normal?\n\n def forward(self, x):\n batch_size = x.shape[0]\n\n x = x.reshape((batch_size, self.num_seq_length, self.num_features))\n\n x, (hidden, _) = self.lstm_enc1(x)\n x, (hidden, _) = self.lstm_enc2(x)\n\n x = hidden.reshape((batch_size, self.num_layers*self.num_directions, self.num_enc_hidden_dim))\n\n return x\n\nclass Decoder(nn.Module):\n def __init__(self):\n super(Decoder, self).__init__()\n\n self.device = get_device()\n\n self.num_features = 1 #Features\n self.num_seq_length = 180 #Sequence\n self.num_directions = 2 #BiDirectional\n self.bidirectional = True if self.num_directions > 1 else False \n self.num_hidden_states = 2 #LSTM H,C -> GRU H\n self.num_hidden_dim = 16\n self.num_enc_hidden_dim = 32 \n self.num_layers = 3 #Layers\n #self.dropout = 0.2 #Dropout\n self.linear_dims = (self.num_hidden_dim * self.num_directions, 1) #Dimension Linear\n self.batch_size = 10\n\n self.lstm_dec1 = nn.LSTM(\n input_size=self.num_enc_hidden_dim*self.num_directions*self.num_layers,\n hidden_size=self.num_enc_hidden_dim,\n num_layers=self.num_layers,\n batch_first=True,\n bidirectional=self.bidirectional \n )\n\n self.lstm_dec2 = nn.LSTM(\n input_size=self.num_enc_hidden_dim*self.num_directions,\n hidden_size=self.num_hidden_dim,\n num_layers=self.num_layers,\n batch_first=True,\n bidirectional=self.bidirectional \n )\n\n \n\n self.dense_layer = nn.Linear(self.num_hidden_dim*self.num_directions,self.num_features)\n\n self.lstm_dec1.apply(self.init_weights)\n self.lstm_dec2.apply(self.init_weights)\n self.dense_layer.apply(self.init_weights)\n\n def init_weights(self, m):\n if type(m) == nn.LSTM or type(m) == nn.Linear:\n for name, param in m.named_parameters():\n if 'bias' in name:\n nn.init.constant(param, 0.01)\n elif 'weight' in name:\n nn.init.xavier_normal(param) #normal?\n\n def forward(self, x):\n batch_size = x.shape[0]\n\n x = x.repeat(1, self.num_seq_length, self.num_features)\n x = x.reshape((batch_size, self.num_seq_length, self.num_enc_hidden_dim*self.num_directions*self.num_layers))\n\n x, hidden = self.lstm_dec1(x)\n x, hidden = self.lstm_dec2(x)\n\n x = x.reshape((batch_size, self.num_seq_length, self.num_hidden_dim*self.num_directions))\n\n x = self.dense_layer(x)\n\n x = x.reshape((batch_size, self.num_features, self.num_seq_length))\n</code></pre>",12 "post_number": 1,13 "post_type": 1,14 "posts_count": 3,15 "updated_at": "2021-02-11T21:40:57.040Z",16 "reply_count": 0,17 "reply_to_post_number": null,18 "quote_count": 0,19 "incoming_link_count": 371,20 "reads": 17,21 "readers_count": 16,22 "score": 1858.4,23 "yours": false,24 "topic_id": 111676,25 "topic_slug": "lstm-autoencoder-not-able-to-learn-special-area-of-a-time-series",26 "display_username": "Chai",27 "primary_group_name": null,28 "flair_name": null,29 "flair_url": null,30 "flair_bg_color": null,31 "flair_color": null,32 "flair_group_id": null,33 "badges_granted": [],34 "version": 1,35 "can_edit": false,36 "can_delete": false,37 "can_recover": false,38 "can_see_hidden_post": false,39 "can_wiki": false,40 "link_counts": [41 {42 "url": "https://discuss.pytorch.org/uploads/default/original/3X/0/d/0d2d23aa3b14e68c214e5721f0b51feb5f5fc334.png",43 "internal": true,44 "reflection": false,45 "clicks": 046 }47 ],48 "read": true,49 "user_title": null,50 "bookmarked": false,51 "actions_summary": [],52 "moderator": false,53 "admin": false,54 "staff": false,55 "user_id": 41283,56 "hidden": false,57 "trust_level": 1,58 "deleted_at": null,59 "user_deleted": false,60 "edit_reason": null,61 "can_view_edit_history": true,62 "wiki": false,63 "post_url": "/t/lstm-autoencoder-not-able-to-learn-special-area-of-a-time-series/111676/1",64 "can_accept_answer": false,65 "can_unaccept_answer": false,66 "accepted_answer": false,67 "topic_accepted_answer": null,68 "can_vote": false69 },70 {71 "id": 263256,72 "name": "Chai",73 "username": "Chai",74 "avatar_template": "/user_avatar/discuss.pytorch.org/chai/{size}/33705_2.png",75 "created_at": "2021-02-12T08:34:13.817Z",76 "cooked": "<p>For the understanding - the curve is sectioned into 5 parts and each part defined one area for a LSTM. The red area is one part with 180 values which will be fed into an LSTM autoencoder, but it does not learn the relevant features.</p>",77 "post_number": 2,78 "post_type": 1,79 "posts_count": 3,80 "updated_at": "2021-02-12T08:34:13.817Z",81 "reply_count": 0,82 "reply_to_post_number": null,83 "quote_count": 0,84 "incoming_link_count": 10,85 "reads": 16,86 "readers_count": 15,87 "score": 53.2,88 "yours": false,89 "topic_id": 111676,90 "topic_slug": "lstm-autoencoder-not-able-to-learn-special-area-of-a-time-series",91 "display_username": "Chai",92 "primary_group_name": null,93 "flair_name": null,94 "flair_url": null,95 "flair_bg_color": null,96 "flair_color": null,97 "flair_group_id": null,98 "badges_granted": [],99 "version": 1,100 "can_edit": false,101 "can_delete": false,102 "can_recover": false,103 "can_see_hidden_post": false,104 "can_wiki": false,105 "read": true,106 "user_title": null,107 "bookmarked": false,108 "actions_summary": [],109 "moderator": false,110 "admin": false,111 "staff": false,112 "user_id": 41283,113 "hidden": false,114 "trust_level": 1,115 "deleted_at": null,116 "user_deleted": false,117 "edit_reason": null,118 "can_view_edit_history": true,119 "wiki": false,120 "post_url": "/t/lstm-autoencoder-not-able-to-learn-special-area-of-a-time-series/111676/2",121 "can_accept_answer": false,122 "can_unaccept_answer": false,123 "accepted_answer": false,124 "topic_accepted_answer": null125 },126 {127 "id": 296027,128 "name": "Existing Virtual",129 "username": "Existing_Virtual",130 "avatar_template": "/user_avatar/discuss.pytorch.org/existing_virtual/{size}/32811_2.png",131 "created_at": "2021-07-15T10:39:48.165Z",132 "cooked": "<p>I saw a few examples in the decoder part similar to your code:</p>\n<p>x = x.repeat(1, self.num_seq_length, self.num_features)<br>\nx = x.reshape((batch_size, self.num_seq_length, …</p>\n<p>i doubt it will work - essentially you feed in the a sequence of the same element, and expect the LSTM will be smart enough to output a sequence of different elements that are close to the original input.</p>\n<p>I guess you need feed the embeded vector as ‘hidden’. then the first input will be a random vector, but start from second step, you need keep feed in to LSTM with your last LSTM output .</p>",133 "post_number": 3,134 "post_type": 1,135 "posts_count": 3,136 "updated_at": "2021-07-15T10:39:48.165Z",137 "reply_count": 0,138 "reply_to_post_number": null,139 "quote_count": 0,140 "incoming_link_count": 38,141 "reads": 11,142 "readers_count": 10,143 "score": 192.2,144 "yours": false,145 "topic_id": 111676,146 "topic_slug": "lstm-autoencoder-not-able-to-learn-special-area-of-a-time-series",147 "display_username": "Existing Virtual",148 "primary_group_name": null,149 "flair_name": null,150 "flair_url": null,151 "flair_bg_color": null,152 "flair_color": null,153 "flair_group_id": null,154 "badges_granted": [],155 "version": 1,156 "can_edit": false,157 "can_delete": false,158 "can_recover": false,159 "can_see_hidden_post": false,160 "can_wiki": false,161 "read": true,162 "user_title": null,163 "bookmarked": false,164 "actions_summary": [],165 "moderator": false,166 "admin": false,167 "staff": false,168 "user_id": 47093,169 "hidden": false,170 "trust_level": 1,171 "deleted_at": null,172 "user_deleted": false,173 "edit_reason": null,174 "can_view_edit_history": true,175 "wiki": false,176 "post_url": "/t/lstm-autoencoder-not-able-to-learn-special-area-of-a-time-series/111676/3",177 "can_accept_answer": false,178 "can_unaccept_answer": false,179 "accepted_answer": false,180 "topic_accepted_answer": null181 }182 ],183 "stream": [184 263210,185 263256,186 296027187 ]188 },189 "timeline_lookup": [190 [191 1,192 1717193 ],194 [195 3,196 1563197 ]198 ],199 "suggested_topics": [200 {201 "fancy_title": "after reload model training loss keep increasing",202 "id": 217345,203 "title": "after reload model training loss keep increasing",204 "slug": "after-reload-model-training-loss-keep-increasing",205 "posts_count": 1,206 "reply_count": 0,207 "highest_post_number": 1,208 "image_url": null,209 "created_at": "2025-03-02T15:09:49.796Z",210 "last_posted_at": "2025-03-02T15:09:49.828Z",211 "bumped": true,212 "bumped_at": "2025-03-02T15:09:49.828Z",213 "archetype": "regular",214 "unseen": false,215 "pinned": false,216 "unpinned": null,217 "visible": true,218 "closed": false,219 "archived": false,220 "bookmarked": null,221 "liked": null,222 "tags_descriptions": {},223 "like_count": 0,224 "views": 37,225 "category_id": 1,226 "featured_link": null,227 "has_accepted_answer": false,228 "posters": [229 {230 "extras": "latest single",231 "description": "Original Poster, Most Recent Poster",232 "user": {233 "id": 83015,234 "username": "LOYINuts",235 "name": "LOYINuts",236 "avatar_template": "/letter_avatar_proxy/v4/letter/l/91b2a8/{size}.png",237 "trust_level": 0238 }239 }240 ]241 },242 {243 "fancy_title": "Size mismatch for model",244 "id": 212435,245 "title": "Size mismatch for model",246 "slug": "size-mismatch-for-model",247 "posts_count": 2,248 "reply_count": 0,249 "highest_post_number": 2,250 "image_url": null,251 "created_at": "2024-11-02T01:18:21.033Z",252 "last_posted_at": "2024-11-02T18:14:42.220Z",253 "bumped": true,254 "bumped_at": "2024-11-02T18:14:42.220Z",255 "archetype": "regular",256 "unseen": false,257 "pinned": false,258 "unpinned": null,259 "visible": true,260 "closed": false,261 "archived": false,262 "bookmarked": null,263 "liked": null,264 "tags_descriptions": {},265 "like_count": 0,266 "views": 817,267 "category_id": 1,268 "featured_link": null,269 "has_accepted_answer": false,270 "posters": [271 {272 "extras": null,273 "description": "Original Poster",274 "user": {275 "id": 80643,276 "username": "Andile_Zungu",277 "name": "Andile Zungu",278 "avatar_template": "/user_avatar/discuss.pytorch.org/andile_zungu/{size}/73200_2.png",279 "trust_level": 0280 }281 },282 {283 "extras": "latest",284 "description": "Most Recent Poster",285 "user": {286 "id": 3534,287 "username": "ptrblck",288 "name": "",289 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",290 "admin": true,291 "moderator": true,292 "trust_level": 2293 }294 }295 ]296 },297 {298 "fancy_title": "ComfyUI - Flux dev GGUF Generation Issue, getting Black output everytime",299 "id": 212758,300 "title": "ComfyUI - Flux dev GGUF Generation Issue, getting Black output everytime",301 "slug": "comfyui-flux-dev-gguf-generation-issue-getting-black-output-everytime",302 "posts_count": 1,303 "reply_count": 0,304 "highest_post_number": 1,305 "image_url": null,306 "created_at": "2024-11-09T22:33:12.593Z",307 "last_posted_at": "2024-11-09T22:33:12.663Z",308 "bumped": true,309 "bumped_at": "2024-11-09T22:33:12.663Z",310 "archetype": "regular",311 "unseen": false,312 "pinned": false,313 "unpinned": null,314 "visible": true,315 "closed": false,316 "archived": false,317 "bookmarked": null,318 "liked": null,319 "tags_descriptions": {},320 "like_count": 0,321 "views": 106,322 "category_id": 1,323 "featured_link": null,324 "has_accepted_answer": false,325 "posters": [326 {327 "extras": "latest single",328 "description": "Original Poster, Most Recent Poster",329 "user": {330 "id": 80796,331 "username": "Adildo_Goat",332 "name": "Adildo Goat",333 "avatar_template": "/user_avatar/discuss.pytorch.org/adildo_goat/{size}/73891_2.png",334 "trust_level": 0335 }336 }337 ]338 },339 {340 "fancy_title": "Troubleshooting LSTM Forecasting Function: What am I doing wrong?",341 "id": 215536,342 "title": "Troubleshooting LSTM Forecasting Function: What am I doing wrong?",343 "slug": "troubleshooting-lstm-forecasting-function-what-am-i-doing-wrong",344 "posts_count": 1,345 "reply_count": 0,346 "highest_post_number": 1,347 "image_url": null,348 "created_at": "2025-01-18T04:26:09.899Z",349 "last_posted_at": "2025-01-18T04:26:09.938Z",350 "bumped": true,351 "bumped_at": "2025-01-18T07:21:34.553Z",352 "archetype": "regular",353 "unseen": false,354 "pinned": false,355 "unpinned": null,356 "visible": true,357 "closed": false,358 "archived": false,359 "bookmarked": null,360 "liked": null,361 "tags_descriptions": {},362 "like_count": 0,363 "views": 26,364 "category_id": 1,365 "featured_link": null,366 "has_accepted_answer": false,367 "posters": [368 {369 "extras": "latest single",370 "description": "Original Poster, Most Recent Poster",371 "user": {372 "id": 82151,373 "username": "NGA",374 "name": "",375 "avatar_template": "/user_avatar/discuss.pytorch.org/nga/{size}/75120_2.png",376 "trust_level": 0377 }378 }379 ]380 },381 {382 "fancy_title": "APL support in pytorch",383 "id": 218060,384 "title": "APL support in pytorch",385 "slug": "apl-support-in-pytorch",386 "posts_count": 4,387 "reply_count": 1,388 "highest_post_number": 4,389 "image_url": null,390 "created_at": "2025-03-20T08:36:58.282Z",391 "last_posted_at": "2025-04-30T19:39:28.503Z",392 "bumped": true,393 "bumped_at": "2025-04-30T19:39:28.503Z",394 "archetype": "regular",395 "unseen": false,396 "pinned": false,397 "unpinned": null,398 "visible": true,399 "closed": false,400 "archived": false,401 "bookmarked": null,402 "liked": null,403 "tags_descriptions": {},404 "like_count": 0,405 "views": 107,406 "category_id": 1,407 "featured_link": null,408 "has_accepted_answer": false,409 "posters": [410 {411 "extras": null,412 "description": "Original Poster",413 "user": {414 "id": 83375,415 "username": "Meghana_R_Prakash",416 "name": "Meghana R Prakash",417 "avatar_template": "/user_avatar/discuss.pytorch.org/meghana_r_prakash/{size}/74876_2.png",418 "trust_level": 0419 }420 },421 {422 "extras": "latest",423 "description": "Most Recent Poster",424 "user": {425 "id": 77605,426 "username": "mhall119",427 "name": "Michael Hall",428 "avatar_template": "/user_avatar/discuss.pytorch.org/mhall119/{size}/71599_2.png",429 "trust_level": 1430 }431 }432 ]433 }434 ],435 "tags_descriptions": {},436 "fancy_title": "LSTM Autoencoder not able to learn special area of a time series",437 "id": 111676,438 "title": "LSTM Autoencoder not able to learn special area of a time series",439 "posts_count": 3,440 "created_at": "2021-02-11T21:40:56.926Z",441 "views": 1180,442 "reply_count": 0,443 "like_count": 0,444 "last_posted_at": "2021-07-15T10:39:48.165Z",445 "visible": true,446 "closed": false,447 "archived": false,448 "has_summary": false,449 "archetype": "regular",450 "slug": "lstm-autoencoder-not-able-to-learn-special-area-of-a-time-series",451 "category_id": 1,452 "word_count": 665,453 "deleted_at": null,454 "user_id": 41283,455 "featured_link": null,456 "pinned_globally": false,457 "pinned_at": null,458 "pinned_until": null,459 "image_url": "https://discuss.pytorch.org/uploads/default/optimized/3X/0/d/0d2d23aa3b14e68c214e5721f0b51feb5f5fc334_2_1023x481.png",460 "slow_mode_seconds": 0,461 "draft": null,462 "draft_key": "topic_111676",463 "draft_sequence": null,464 "unpinned": null,465 "pinned": false,466 "current_post_number": 1,467 "highest_post_number": 3,468 "deleted_by": null,469 "actions_summary": [470 {471 "id": 4,472 "count": 0,473 "hidden": false,474 "can_act": false475 },476 {477 "id": 8,478 "count": 0,479 "hidden": false,480 "can_act": false481 },482 {483 "id": 10,484 "count": 0,485 "hidden": false,486 "can_act": false487 },488 {489 "id": 7,490 "count": 0,491 "hidden": false,492 "can_act": false493 }494 ],495 "chunk_size": 20,496 "bookmarked": false,497 "topic_timer": null,498 "message_bus_last_id": 0,499 "participant_count": 2,500 "show_read_indicator": false,501 "thumbnails": [502 {503 "max_width": null,504 "max_height": null,505 "width": 1700,506 "height": 799,507 "url": "https://discuss.pytorch.org/uploads/default/original/3X/0/d/0d2d23aa3b14e68c214e5721f0b51feb5f5fc334.png"508 },509 {510 "max_width": 1024,511 "max_height": 1024,512 "width": 1023,513 "height": 481,514 "url": "https://discuss.pytorch.org/uploads/default/optimized/3X/0/d/0d2d23aa3b14e68c214e5721f0b51feb5f5fc334_2_1023x481.png"515 }516 ],517 "slow_mode_enabled_until": null,518 "can_vote": false,519 "vote_count": 0,520 "user_voted": false,521 "discourse_zendesk_plugin_zendesk_id": null,522 "discourse_zendesk_plugin_zendesk_url": "https://your-url.zendesk.com/agent/tickets/",523 "details": {524 "can_edit": false,525 "notification_level": 1,526 "participants": [527 {528 "id": 41283,529 "username": "Chai",530 "name": "Chai",531 "avatar_template": "/user_avatar/discuss.pytorch.org/chai/{size}/33705_2.png",532 "post_count": 2,533 "primary_group_name": null,534 "flair_name": null,535 "flair_url": null,536 "flair_color": null,537 "flair_bg_color": null,538 "flair_group_id": null,539 "trust_level": 1540 },541 {542 "id": 47093,543 "username": "Existing_Virtual",544 "name": "Existing Virtual",545 "avatar_template": "/user_avatar/discuss.pytorch.org/existing_virtual/{size}/32811_2.png",546 "post_count": 1,547 "primary_group_name": null,548 "flair_name": null,549 "flair_url": null,550 "flair_color": null,551 "flair_bg_color": null,552 "flair_group_id": null,553 "trust_level": 1554 }555 ],556 "created_by": {557 "id": 41283,558 "username": "Chai",559 "name": "Chai",560 "avatar_template": "/user_avatar/discuss.pytorch.org/chai/{size}/33705_2.png"561 },562 "last_poster": {563 "id": 47093,564 "username": "Existing_Virtual",565 "name": "Existing Virtual",566 "avatar_template": "/user_avatar/discuss.pytorch.org/existing_virtual/{size}/32811_2.png"567 }568 },569 "bookmarks": []570 },571 {572 "post_stream": {573 "posts": [574 {575 "id": 295848,576 "name": "Azzeddine Yacine",577 "username": "DarQ",578 "avatar_template": "/user_avatar/discuss.pytorch.org/darq/{size}/39606_2.png",579 "created_at": "2021-07-14T18:29:26.296Z",580 "cooked": "<p>I have an iterable dataset and let’s say I have 2 workers: worker 0 generates an audio clip and combines it with other pre-generated audio clips to create a data point, and workers 1 only combines pre-generated audio clips to create a data point. Naturally, worker 0 takes more time then worker 1 (audio clip generation and saving to the disk takes ≈ 0.8s while loading 12 audio clips takes 0.18s).</p>\n<p>The workers obviously work in parallel so I expect the average loading time to be equal to 0.18 seconds at most (while worker 0 is generating the audio clip, worker 1 will take on the task of generating data points) that is not the case however, since I’m getting avg loading time of about 0.52s (almost thrice as slow). How can I solve this issue? is that even possible?<br>\nhere is some pseudo code:</p>\n<pre><code class=\"lang-auto\">class GeneratedDataset(IterableDataset):\n def __init__(self, note_num):\n super(GeneratedDataset).__init__()\n self.note_num = note_num\n \n def __iter__(self):\n worker = torch.utils.data.get_worker_info()\n if worker:\n id = worker.id\n np.random.seed(int(2**31*torch.rand(1)))\n else:\n id = 0\n while True:\n if id==0: # this is the code that only runs on worker 0 \n \n # GENERATE AUDIO CLIP\n # GENERATE RESPECTIVE LABELS\n # SAVE BOTH\n \n ------------------------\n \n # LOADING AUDIO CLIP\n # GENERATING SPECTROGRAM\n # TENSOR MANIPULATION (padding, normalizing, etc.)\n # SAVING THE TENSOR\n else:\n # LOAD RANDOM (pre-saved) TENSOR\n # LOAD RESPECTIVE LABELS\n for _ in range(12):\n # MORE TENSOR MANIPULATION\n \n # PROCESSING THE LABELS\n \n # CHOOSE ANOTHER (TENSOR, LABELS) PAIR FOR NEXT LOOP\n\n # CONCATENATE THE LOADED TENSORS. SAME FOR THE LABELS\n yield # TENSOR, LABELS\n</code></pre>\n<p>Any help would be greatly appreciated.</p>",581 "post_number": 1,582 "post_type": 1,583 "posts_count": 1,584 "updated_at": "2021-07-15T10:28:30.938Z",585 "reply_count": 0,586 "reply_to_post_number": null,587 "quote_count": 0,588 "incoming_link_count": 11,589 "reads": 2,590 "readers_count": 1,591 "score": 55.4,592 "yours": false,593 "topic_id": 126761,594 "topic_slug": "a-worker-instance-holds-down-other-workers-when-processing-load-is-unevenly-distributed",595 "display_username": "Azzeddine Yacine",596 "primary_group_name": null,597 "flair_name": null,598 "flair_url": null,599 "flair_bg_color": null,600 "flair_color": null,601 "flair_group_id": null,602 "badges_granted": [],603 "version": 2,604 "can_edit": false,605 "can_delete": false,606 "can_recover": false,607 "can_see_hidden_post": false,608 "can_wiki": false,609 "read": true,610 "user_title": null,611 "bookmarked": false,612 "actions_summary": [],613 "moderator": false,614 "admin": false,615 "staff": false,616 "user_id": 46557,617 "hidden": false,618 "trust_level": 1,619 "deleted_at": null,620 "user_deleted": false,621 "edit_reason": null,622 "can_view_edit_history": true,623 "wiki": false,624 "post_url": "/t/a-worker-instance-holds-down-other-workers-when-processing-load-is-unevenly-distributed/126761/1",625 "can_accept_answer": false,626 "can_unaccept_answer": false,627 "accepted_answer": false,628 "topic_accepted_answer": null,629 "can_vote": false630 }631 ],632 "stream": [633 295848634 ]635 },636 "timeline_lookup": [637 [638 1,639 1564640 ]641 ],642 "suggested_topics": [643 {644 "fancy_title": "Wrong nvcc version when compiling plugins",645 "id": 216293,646 "title": "Wrong nvcc version when compiling plugins",647 "slug": "wrong-nvcc-version-when-compiling-plugins",648 "posts_count": 1,649 "reply_count": 0,650 "highest_post_number": 1,651 "image_url": null,652 "created_at": "2025-02-06T01:00:59.715Z",653 "last_posted_at": "2025-02-06T01:00:59.757Z",654 "bumped": true,655 "bumped_at": "2025-02-06T01:02:58.568Z",656 "archetype": "regular",657 "unseen": false,658 "pinned": false,659 "unpinned": null,660 "visible": true,661 "closed": false,662 "archived": false,663 "bookmarked": null,664 "liked": null,665 "tags_descriptions": {},666 "like_count": 0,667 "views": 82,668 "category_id": 1,669 "featured_link": null,670 "has_accepted_answer": false,671 "posters": [672 {673 "extras": "latest single",674 "description": "Original Poster, Most Recent Poster",675 "user": {676 "id": 29433,677 "username": "cltexe",678 "name": "Omer Faruk Soylemez",679 "avatar_template": "/user_avatar/discuss.pytorch.org/cltexe/{size}/41817_2.png",680 "trust_level": 1681 }682 }683 ]684 },685 {686 "fancy_title": "RuntimeError: The size of tensor a (80) must match the size of tensor b (95) at non-singleton dimension 2",687 "id": 214768,688 "title": "RuntimeError: The size of tensor a (80) must match the size of tensor b (95) at non-singleton dimension 2",689 "slug": "runtimeerror-the-size-of-tensor-a-80-must-match-the-size-of-tensor-b-95-at-non-singleton-dimension-2",690 "posts_count": 12,691 "reply_count": 11,692 "highest_post_number": 13,693 "image_url": null,694 "created_at": "2024-12-30T02:09:47.988Z",695 "last_posted_at": "2025-01-01T07:23:52.070Z",696 "bumped": true,697 "bumped_at": "2025-01-01T07:23:52.070Z",698 "archetype": "regular",699 "unseen": false,700 "pinned": false,701 "unpinned": null,702 "visible": true,703 "closed": false,704 "archived": false,705 "bookmarked": null,706 "liked": null,707 "tags_descriptions": {},708 "like_count": 0,709 "views": 208,710 "category_id": 1,711 "featured_link": null,712 "has_accepted_answer": true,713 "posters": [714 {715 "extras": "latest",716 "description": "Original Poster, Most Recent Poster",717 "user": {718 "id": 81774,719 "username": "Thierry_Roger_Bayala",720 "name": "Thierry Roger Bayala",721 "avatar_template": "/user_avatar/discuss.pytorch.org/thierry_roger_bayala/{size}/74787_2.png",722 "trust_level": 1723 }724 },725 {726 "extras": null,727 "description": "Frequent Poster, Accepted Answer",728 "user": {729 "id": 81709,730 "username": "QLYYLQ",731 "name": "qly",732 "avatar_template": "/user_avatar/discuss.pytorch.org/qlyylq/{size}/74718_2.png",733 "trust_level": 1734 }735 }736 ]737 },738 {739 "fancy_title": "GNN for electrical distribution systems, HELP",740 "id": 215046,741 "title": "GNN for electrical distribution systems, HELP",742 "slug": "gnn-for-electrical-distribution-systems-help",743 "posts_count": 1,744 "reply_count": 0,745 "highest_post_number": 1,746 "image_url": null,747 "created_at": "2025-01-06T21:52:12.794Z",748 "last_posted_at": "2025-01-06T21:52:12.832Z",749 "bumped": true,750 "bumped_at": "2025-01-06T21:52:12.832Z",751 "archetype": "regular",752 "unseen": false,753 "pinned": false,754 "unpinned": null,755 "visible": true,756 "closed": false,757 "archived": false,758 "bookmarked": null,759 "liked": null,760 "tags_descriptions": {},761 "like_count": 0,762 "views": 30,763 "category_id": 1,764 "featured_link": null,765 "has_accepted_answer": false,766 "posters": [767 {768 "extras": "latest single",769 "description": "Original Poster, Most Recent Poster",770 "user": {771 "id": 81908,772 "username": "Diego_Paul_Guachichu",773 "name": "Diego Paul Guachichullca Bermeo",774 "avatar_template": "/user_avatar/discuss.pytorch.org/diego_paul_guachichu/{size}/74933_2.png",775 "trust_level": 0776 }777 }778 ]779 },780 {781 "fancy_title": "RuntimeError: mat1 and mat2 shapes cannot be multiplied (1x3 and 20x10)",782 "id": 213493,783 "title": "RuntimeError: mat1 and mat2 shapes cannot be multiplied (1x3 and 20x10)",784 "slug": "runtimeerror-mat1-and-mat2-shapes-cannot-be-multiplied-1x3-and-20x10",785 "posts_count": 1,786 "reply_count": 0,787 "highest_post_number": 1,788 "image_url": null,789 "created_at": "2024-11-26T21:12:05.959Z",790 "last_posted_at": "2024-11-26T21:12:06.050Z",791 "bumped": true,792 "bumped_at": "2024-11-26T21:12:06.050Z",793 "archetype": "regular",794 "unseen": false,795 "pinned": false,796 "unpinned": null,797 "visible": true,798 "closed": false,799 "archived": false,800 "bookmarked": null,801 "liked": null,802 "tags_descriptions": {},803 "like_count": 0,804 "views": 23,805 "category_id": 1,806 "featured_link": null,807 "has_accepted_answer": false,808 "posters": [809 {810 "extras": "latest single",811 "description": "Original Poster, Most Recent Poster",812 "user": {813 "id": 81152,814 "username": "dpalate",815 "name": "Devashish Palate",816 "avatar_template": "/user_avatar/discuss.pytorch.org/dpalate/{size}/74220_2.png",817 "trust_level": 0818 }819 }820 ]821 },822 {823 "fancy_title": "“turing_fp16_s1688gemm_fp16_128x128_ldg8_relu_f2f_tn”",824 "id": 219212,825 "title": "\"turing_fp16_s1688gemm_fp16_128x128_ldg8_relu_f2f_tn\"",826 "slug": "turing-fp16-s1688gemm-fp16-128x128-ldg8-relu-f2f-tn",827 "posts_count": 6,828 "reply_count": 4,829 "highest_post_number": 6,830 "image_url": null,831 "created_at": "2025-04-17T16:50:15.925Z",832 "last_posted_at": "2025-04-21T12:41:41.181Z",833 "bumped": true,834 "bumped_at": "2025-04-21T12:41:41.181Z",835 "archetype": "regular",836 "unseen": false,837 "pinned": false,838 "unpinned": null,839 "visible": true,840 "closed": false,841 "archived": false,842 "bookmarked": null,843 "liked": null,844 "tags_descriptions": {},845 "like_count": 0,846 "views": 132,847 "category_id": 1,848 "featured_link": null,849 "has_accepted_answer": false,850 "posters": [851 {852 "extras": null,853 "description": "Original Poster",854 "user": {855 "id": 83873,856 "username": "JB13",857 "name": null,858 "avatar_template": "/letter_avatar_proxy/v4/letter/j/71c47a/{size}.png",859 "trust_level": 0860 }861 },862 {863 "extras": "latest",864 "description": "Most Recent Poster",865 "user": {866 "id": 3534,867 "username": "ptrblck",868 "name": "",869 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",870 "admin": true,871 "moderator": true,872 "trust_level": 2873 }874 }875 ]876 }877 ],878 "tags_descriptions": {},879 "fancy_title": "A worker instance holds down other workers when processing load is unevenly distributed",880 "id": 126761,881 "title": "A worker instance holds down other workers when processing load is unevenly distributed",882 "posts_count": 1,883 "created_at": "2021-07-14T18:29:26.221Z",884 "views": 244,885 "reply_count": 0,886 "like_count": 0,887 "last_posted_at": "2021-07-14T18:29:26.296Z",888 "visible": true,889 "closed": false,890 "archived": false,891 "has_summary": false,892 "archetype": "regular",893 "slug": "a-worker-instance-holds-down-other-workers-when-processing-load-is-unevenly-distributed",894 "category_id": 1,895 "word_count": 271,896 "deleted_at": null,897 "user_id": 46557,898 "featured_link": null,899 "pinned_globally": false,900 "pinned_at": null,901 "pinned_until": null,902 "image_url": null,903 "slow_mode_seconds": 0,904 "draft": null,905 "draft_key": "topic_126761",906 "draft_sequence": null,907 "unpinned": null,908 "pinned": false,909 "current_post_number": 1,910 "highest_post_number": 1,911 "deleted_by": null,912 "actions_summary": [913 {914 "id": 4,915 "count": 0,916 "hidden": false,917 "can_act": false918 },919 {920 "id": 8,921 "count": 0,922 "hidden": false,923 "can_act": false924 },925 {926 "id": 10,927 "count": 0,928 "hidden": false,929 "can_act": false930 },931 {932 "id": 7,933 "count": 0,934 "hidden": false,935 "can_act": false936 }937 ],938 "chunk_size": 20,939 "bookmarked": false,940 "topic_timer": null,941 "message_bus_last_id": 0,942 "participant_count": 1,943 "show_read_indicator": false,944 "thumbnails": null,945 "slow_mode_enabled_until": null,946 "can_vote": false,947 "vote_count": 0,948 "user_voted": false,949 "discourse_zendesk_plugin_zendesk_id": null,950 "discourse_zendesk_plugin_zendesk_url": "https://your-url.zendesk.com/agent/tickets/",951 "details": {952 "can_edit": false,953 "notification_level": 1,954 "participants": [955 {956 "id": 46557,957 "username": "DarQ",958 "name": "Azzeddine Yacine",959 "avatar_template": "/user_avatar/discuss.pytorch.org/darq/{size}/39606_2.png",960 "post_count": 1,961 "primary_group_name": null,962 "flair_name": null,963 "flair_url": null,964 "flair_color": null,965 "flair_bg_color": null,966 "flair_group_id": null,967 "trust_level": 1968 }969 ],970 "created_by": {971 "id": 46557,972 "username": "DarQ",973 "name": "Azzeddine Yacine",974 "avatar_template": "/user_avatar/discuss.pytorch.org/darq/{size}/39606_2.png"975 },976 "last_poster": {977 "id": 46557,978 "username": "DarQ",979 "name": "Azzeddine Yacine",980 "avatar_template": "/user_avatar/discuss.pytorch.org/darq/{size}/39606_2.png"981 }982 },983 "bookmarks": []984 },985 {986 "post_stream": {987 "posts": [988 {989 "id": 295689,990 "name": "Prateek Varshney",991 "username": "Prateek_Varshney",992 "avatar_template": "/user_avatar/discuss.pytorch.org/prateek_varshney/{size}/16778_2.png",993 "created_at": "2021-07-14T07:57:39.460Z",994 "cooked": "<p>I am trying to subset particular class (in particular, samples from labels 0, 4, 8) samples from the MNIST-M Dataset <a href=\"https://drive.google.com/file/d/0B9Z4d7lAwbnTNDdNeFlERWRGNVk/view?resourcekey=0-QEalUvOz5FDK-aVK7ZOALg\" rel=\"noopener nofollow ugc\">(source)</a>. Since torchvision does not have a predefined library function to load the MNIST-M Dataset, I am using the following custom dataset class function:</p>\n<pre><code class=\"lang-auto\">class MNIST_M(torch.utils.data.Dataset):\n def __init__(self, root, train, transform=None):\n self.train = train\n self.transform = transform\n if train:\n self.image_dir = os.path.join(root, 'mnist_m_train')\n self.labels_file = os.path.join(root, \"mnist_m_train_labels.txt\")\n else:\n self.image_dir = os.path.join(root, 'mnist_m_test')\n self.labels_file = os.path.join(root, \"mnist_m_test_labels.txt\")\n\n with open(self.labels_file, \"r\") as fp:\n \tcontent = fp.readlines()\n self.mapping = list(map(lambda x: (x[0], int(x[1])), [c.strip().split() for c in content]))\n\n def __len__(self):\n return len(self.mapping)\n\n def __getitem__(self, idx):\n image, labels = self.mapping[idx]\n image = os.path.join(self.image_dir, image)\n image = self.transform(Image.open(image).convert('RGB'))\n return image, labels\n\n def _load_data(self):\n data = read_image_file(self.image_dir)\n targets = read_label_file(self.labels_file)\n\n return data, targets\n</code></pre>\n<p>To create a subset of the dataset, I am using the following code which I have ported from the similar case of subsetting MNIST Dataset (<a href=\"https://discuss.pytorch.org/t/how-to-use-one-class-of-number-in-mnist/26276\">subsetting MNIST reference</a>):</p>\n<pre><code class=\"lang-auto\">mnist_train_ds_modded = datasets.MNIST(root_dir, download=True, train=True, transform=source_transform)\n\nmnistm_train_ds_modded = MNIST_M(root=root_dir, train=True,\n transform=transforms.Compose([\n transforms.Scale(imageSize),\n transforms.ToTensor(),\n transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)),\n ]))\n\nmnistm_train_ds_modded.labels = torch.tensor(mnistm_train_ds_modded.labels)\n\nmnistm_train_indexes_0 = 1*(mnistm_train_ds_modded.labels == 0).nonzero().flatten().tolist()\nmnistm_train_indexes_4 = 1*(mnistm_train_ds_modded.labels == 4).nonzero().flatten().tolist()\nmnistm_train_indexes_8 = 1*(mnistm_train_ds_modded.labels == 8).nonzero().flatten().tolist()\n\n\nmnistm_train_modded_idx = mnistm_train_indexes_0 + mnistm_train_indexes_4 + mnistm_train_indexes_8\nmnistm_train_ds_modded.labels = mnistm_train_ds_modded.labels[mnistm_train_modded_idx]\nmnistm_train_ds_modded.data = mnistm_train_ds_modded.data[mnistm_train_modded_idx]\n</code></pre>\n<p>Clearly, the MNIST-M dataset class written above does not have any attributes called data and labels. So the above code will not work as it does in the case of MNIST.</p>\n<p>I went through the source code of MNIST to define the class attributes data and labels but I am unable to do the same for png files (MNIST-M has png files). Kindly help me define the class attributes so that I can subset the dataset.</p>",995 "post_number": 1,996 "post_type": 1,997 "posts_count": 3,998 "updated_at": "2021-07-14T07:57:39.460Z",999 "reply_count": 0,1000 "reply_to_post_number": null,1001 "quote_count": 0,1002 "incoming_link_count": 1050,1003 "reads": 18,1004 "readers_count": 17,1005 "score": 5253.6,1006 "yours": false,1007 "topic_id": 126694,1008 "topic_slug": "class-wise-subset-of-mnist-m-dataset",1009 "display_username": "Prateek Varshney",1010 "primary_group_name": null,1011 "flair_name": null,1012 "flair_url": null,1013 "flair_bg_color": null,1014 "flair_color": null,1015 "flair_group_id": null,1016 "badges_granted": [],1017 "version": 1,1018 "can_edit": false,1019 "can_delete": false,1020 "can_recover": false,1021 "can_see_hidden_post": false,1022 "can_wiki": false,1023 "link_counts": [1024 {1025 "url": "https://drive.google.com/file/d/0B9Z4d7lAwbnTNDdNeFlERWRGNVk/view?resourcekey=0-QEalUvOz5FDK-aVK7ZOALg",1026 "internal": false,1027 "reflection": false,1028 "clicks": 581029 },1030 {1031 "url": "https://discuss.pytorch.org/t/how-to-use-one-class-of-number-in-mnist/26276",1032 "internal": true,1033 "reflection": false,1034 "title": "How to use one class of number in MNIST",1035 "clicks": 181036 }1037 ],1038 "read": true,1039 "user_title": null,1040 "bookmarked": false,1041 "actions_summary": [],1042 "moderator": false,1043 "admin": false,1044 "staff": false,1045 "user_id": 35992,1046 "hidden": false,1047 "trust_level": 1,1048 "deleted_at": null,1049 "user_deleted": false,1050 "edit_reason": null,1051 "can_view_edit_history": true,1052 "wiki": false,1053 "post_url": "/t/class-wise-subset-of-mnist-m-dataset/126694/1",1054 "can_accept_answer": false,1055 "can_unaccept_answer": false,1056 "accepted_answer": false,1057 "topic_accepted_answer": null,1058 "can_vote": false1059 },1060 {1061 "id": 295946,1062 "name": "",1063 "username": "ptrblck",1064 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",1065 "created_at": "2021-07-15T06:40:34.336Z",1066 "cooked": "<p>I don’t know how <code>MNIST-M</code> is stored, but in case the images are stored in folders for each class, you could use <code>ImageFolder</code> to lazily load (and process) each image and target.</p>",1067 "post_number": 2,1068 "post_type": 1,1069 "posts_count": 3,1070 "updated_at": "2021-07-15T06:40:34.336Z",1071 "reply_count": 1,1072 "reply_to_post_number": null,1073 "quote_count": 0,1074 "incoming_link_count": 23,1075 "reads": 14,1076 "readers_count": 13,1077 "score": 122.8,1078 "yours": false,1079 "topic_id": 126694,1080 "topic_slug": "class-wise-subset-of-mnist-m-dataset",1081 "display_username": "",1082 "primary_group_name": null,1083 "flair_name": null,1084 "flair_url": null,1085 "flair_bg_color": null,1086 "flair_color": null,1087 "flair_group_id": null,1088 "badges_granted": [],1089 "version": 1,1090 "can_edit": false,1091 "can_delete": false,1092 "can_recover": false,1093 "can_see_hidden_post": false,1094 "can_wiki": false,1095 "read": true,1096 "user_title": "",1097 "bookmarked": false,1098 "actions_summary": [],1099 "moderator": true,1100 "admin": true,1101 "staff": true,1102 "user_id": 3534,1103 "hidden": false,1104 "trust_level": 2,1105 "deleted_at": null,1106 "user_deleted": false,1107 "edit_reason": null,1108 "can_view_edit_history": true,1109 "wiki": false,1110 "post_url": "/t/class-wise-subset-of-mnist-m-dataset/126694/2",1111 "can_accept_answer": false,1112 "can_unaccept_answer": false,1113 "accepted_answer": false,1114 "topic_accepted_answer": null1115 },1116 {1117 "id": 296015,1118 "name": "Prateek Varshney",1119 "username": "Prateek_Varshney",1120 "avatar_template": "/user_avatar/discuss.pytorch.org/prateek_varshney/{size}/16778_2.png",1121 "created_at": "2021-07-15T09:54:59.586Z",1122 "cooked": "<p>Here is the drive location of the MNIST-M Dataset: <a href=\"https://drive.google.com/drive/folders/1FW8xkzrFI5qJCev-rRwasycyiXFBUS6S?usp=sharing\" rel=\"noopener nofollow ugc\">MNIST-M folder</a>. So the images are not stored in folders for each class, rather all the images have been stored in the same folder.</p>\n<p>Kindly note that the MNIST-M class defined above also loads the samples lazily like ImageFolder but since I’m only interested in a subset of the data, this will amount to reading the entire dataset again and again and only filtering the classes I’m interested in, which is computationally expensive as compared to having a preprocessed dataset containing only the classes of interest.</p>",1123 "post_number": 3,1124 "post_type": 1,1125 "posts_count": 3,1126 "updated_at": "2021-07-15T10:23:39.321Z",1127 "reply_count": 0,1128 "reply_to_post_number": 2,1129 "quote_count": 0,1130 "incoming_link_count": 3,1131 "reads": 14,1132 "readers_count": 13,1133 "score": 17.8,1134 "yours": false,1135 "topic_id": 126694,1136 "topic_slug": "class-wise-subset-of-mnist-m-dataset",1137 "display_username": "Prateek Varshney",1138 "primary_group_name": null,1139 "flair_name": null,1140 "flair_url": null,1141 "flair_bg_color": null,1142 "flair_color": null,1143 "flair_group_id": null,1144 "badges_granted": [],1145 "version": 3,1146 "can_edit": false,1147 "can_delete": false,1148 "can_recover": false,1149 "can_see_hidden_post": false,1150 "can_wiki": false,1151 "link_counts": [1152 {1153 "url": "https://drive.google.com/drive/folders/1FW8xkzrFI5qJCev-rRwasycyiXFBUS6S?usp=sharing",1154 "internal": false,1155 "reflection": false,1156 "title": "mnist_m - Google Drive",1157 "clicks": 911158 }1159 ],1160 "read": true,1161 "user_title": null,1162 "reply_to_user": {1163 "id": 3534,1164 "username": "ptrblck",1165 "name": "",1166 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png"1167 },1168 "bookmarked": false,1169 "actions_summary": [],1170 "moderator": false,1171 "admin": false,1172 "staff": false,1173 "user_id": 35992,1174 "hidden": false,1175 "trust_level": 1,1176 "deleted_at": null,1177 "user_deleted": false,1178 "edit_reason": null,1179 "can_view_edit_history": true,1180 "wiki": false,1181 "post_url": "/t/class-wise-subset-of-mnist-m-dataset/126694/3",1182 "can_accept_answer": false,1183 "can_unaccept_answer": false,1184 "accepted_answer": false,1185 "topic_accepted_answer": null1186 }1187 ],1188 "stream": [1189 295689,1190 295946,1191 2960151192 ]1193 },1194 "timeline_lookup": [1195 [1196 1,1197 15651198 ],1199 [1200 2,