Anurag1734/cuda-error-resolution-analysis
07
1[2 {3 "post_stream": {4 "posts": [5 {6 "id": 292603,7 "name": "Giuseppe Torrisi",8 "username": "Nexkira",9 "avatar_template": "/letter_avatar_proxy/v4/letter/n/9fc29f/{size}.png",10 "created_at": "2021-06-27T21:32:23.460Z",11 "cooked": "<p>Hi, when I am using torch.use_deterministic_algorithms(True) and I run this code, I got this error.<br>\nsum_img are images coverted into a torch tensor of dimension: batch_size x channel x img_size (128x128)</p>\n<p>If I move sum_img to cpu instead of cuda, I have got no error</p>\n<p>The code gives me error even if I put a number instead a list of number<br>\nIf I run this code with deterministic algorithms false run without a problem, how can I fix it? thank you in advice<br>\n<br>\n</p>\n<pre><code class=\"lang-auto\">list_n = [0.85,0.5,0.3] \nfor i in range(3):\n\t\tchannel = sum_img[:,i,:,:]\n\t\tchannel[channel>=5] = list_n[i] \n\n\n\n\nRuntimeError: linearIndex.numel()*sliceSize*nElemBefore == value.numel()INTERNAL ASSERT FAILED at \"/pytorch/aten/src/ATen/native/cuda/Indexing.cu\":253, please report a bug to PyTorch. number of flattened indices did not match number of elements in the value tensor2684801</code></pre>",12 "post_number": 1,13 "post_type": 1,14 "posts_count": 9,15 "updated_at": "2021-06-28T08:34:15.297Z",16 "reply_count": 0,17 "reply_to_post_number": null,18 "quote_count": 0,19 "incoming_link_count": 1723,20 "reads": 33,21 "readers_count": 32,22 "score": 8581.6,23 "yours": false,24 "topic_id": 125200,25 "topic_slug": "torch-deterministic-algorithms-error",26 "display_username": "Giuseppe Torrisi",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": 4,35 "can_edit": false,36 "can_delete": false,37 "can_recover": false,38 "can_see_hidden_post": false,39 "can_wiki": false,40 "read": true,41 "user_title": null,42 "bookmarked": false,43 "actions_summary": [],44 "moderator": false,45 "admin": false,46 "staff": false,47 "user_id": 46670,48 "hidden": false,49 "trust_level": 1,50 "deleted_at": null,51 "user_deleted": false,52 "edit_reason": null,53 "can_view_edit_history": true,54 "wiki": false,55 "post_url": "/t/torch-deterministic-algorithms-error/125200/1",56 "can_accept_answer": false,57 "can_unaccept_answer": false,58 "accepted_answer": false,59 "topic_accepted_answer": null,60 "can_vote": false61 },62 {63 "id": 292727,64 "name": "",65 "username": "ptrblck",66 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",67 "created_at": "2021-06-28T08:42:26.708Z",68 "cooked": "<p>Could you post the output of <code>python -m torch.utils.collect_env</code> as well as the shapes for all tensors, please?<br>\nBased on the error message it could be a valid error, which we would have to debug and fix (if it’s still observed in the current nightly release).</p>",69 "post_number": 2,70 "post_type": 1,71 "posts_count": 9,72 "updated_at": "2021-06-28T08:42:26.708Z",73 "reply_count": 2,74 "reply_to_post_number": null,75 "quote_count": 0,76 "incoming_link_count": 8,77 "reads": 32,78 "readers_count": 31,79 "score": 56.4,80 "yours": false,81 "topic_id": 125200,82 "topic_slug": "torch-deterministic-algorithms-error",83 "display_username": "",84 "primary_group_name": null,85 "flair_name": null,86 "flair_url": null,87 "flair_bg_color": null,88 "flair_color": null,89 "flair_group_id": null,90 "badges_granted": [],91 "version": 1,92 "can_edit": false,93 "can_delete": false,94 "can_recover": false,95 "can_see_hidden_post": false,96 "can_wiki": false,97 "read": true,98 "user_title": "",99 "bookmarked": false,100 "actions_summary": [],101 "moderator": true,102 "admin": true,103 "staff": true,104 "user_id": 3534,105 "hidden": false,106 "trust_level": 2,107 "deleted_at": null,108 "user_deleted": false,109 "edit_reason": null,110 "can_view_edit_history": true,111 "wiki": false,112 "post_url": "/t/torch-deterministic-algorithms-error/125200/2",113 "can_accept_answer": false,114 "can_unaccept_answer": false,115 "accepted_answer": false,116 "topic_accepted_answer": null117 },118 {119 "id": 292736,120 "name": "Giuseppe Torrisi",121 "username": "Nexkira",122 "avatar_template": "/letter_avatar_proxy/v4/letter/n/9fc29f/{size}.png",123 "created_at": "2021-06-28T08:57:02.813Z",124 "cooked": "<aside class=\"quote no-group\" data-username=\"ptrblck\" data-post=\"2\" data-topic=\"125200\">\n<div class=\"title\">\n<div class=\"quote-controls\"></div>\n<img loading=\"lazy\" alt=\"\" width=\"24\" height=\"24\" src=\"https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/ptrblck/48/1823_2.png\" class=\"avatar\"> ptrblck:</div>\n<blockquote>\n<p><code>python -m torch.utils.collect_env</code></p>\n</blockquote>\n</aside>\n<p>thank you for answering the output is:<br>\n<br>\n</p>\n<pre><code class=\"lang-auto\">Collecting environment information...\nPyTorch version: 1.9.0+cu102\nIs debug build: False\nCUDA used to build PyTorch: 10.2\nROCM used to build PyTorch: N/A\n\nOS: Ubuntu 18.04.5 LTS (x86_64)\nGCC version: (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0\nClang version: 6.0.0-1ubuntu2 (tags/RELEASE_600/final)\nCMake version: version 3.12.0\nLibc version: glibc-2.26\n\nPython version: 3.7 (64-bit runtime)\nPython platform: Linux-5.4.104+-x86_64-with-Ubuntu-18.04-bionic\nIs CUDA available: True\nCUDA runtime version: 11.0.221\nGPU models and configuration: GPU 0: Tesla P100-PCIE-16GB\nNvidia driver version: 460.32.03\ncuDNN version: Probably one of the following:\n/usr/lib/x86_64-linux-gnu/libcudnn.so.7.6.5\n/usr/lib/x86_64-linux-gnu/libcudnn.so.8.0.4\n/usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.0.4\n/usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.0.4\n/usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.0.4\n/usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.0.4\n/usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.0.4\n/usr/lib/x86_64-linux-gnu/libcudnn_ops_train.so.8.0.4\nHIP runtime version: N/A\nMIOpen runtime version: N/A\n\nVersions of relevant libraries:\n[pip3] numpy==1.19.5\n[pip3] torch==1.9.0+cu102\n[pip3] torchsummary==1.5.1\n[pip3] torchtext==0.10.0\n[pip3] torchvision==0.10.0+cu102\n[conda] Could not collect\n\n</code></pre>\n<p>the shape are:<br>\n<br>\n</p>\n<pre><code class=\"lang-auto\">sum_img = torch.Size([64, 3, 128, 128])\nchannel = torch.Size([64, 128, 128])</code></pre>",125 "post_number": 3,126 "post_type": 1,127 "posts_count": 9,128 "updated_at": "2021-06-28T08:58:25.897Z",129 "reply_count": 0,130 "reply_to_post_number": 2,131 "quote_count": 1,132 "incoming_link_count": 49,133 "reads": 29,134 "readers_count": 28,135 "score": 250.8,136 "yours": false,137 "topic_id": 125200,138 "topic_slug": "torch-deterministic-algorithms-error",139 "display_username": "Giuseppe Torrisi",140 "primary_group_name": null,141 "flair_name": null,142 "flair_url": null,143 "flair_bg_color": null,144 "flair_color": null,145 "flair_group_id": null,146 "badges_granted": [],147 "version": 1,148 "can_edit": false,149 "can_delete": false,150 "can_recover": false,151 "can_see_hidden_post": false,152 "can_wiki": false,153 "read": true,154 "user_title": null,155 "bookmarked": false,156 "actions_summary": [],157 "moderator": false,158 "admin": false,159 "staff": false,160 "user_id": 46670,161 "hidden": false,162 "trust_level": 1,163 "deleted_at": null,164 "user_deleted": false,165 "edit_reason": null,166 "can_view_edit_history": true,167 "wiki": false,168 "post_url": "/t/torch-deterministic-algorithms-error/125200/3",169 "can_accept_answer": false,170 "can_unaccept_answer": false,171 "accepted_answer": false,172 "topic_accepted_answer": null173 },174 {175 "id": 292963,176 "name": "Giuseppe Torrisi",177 "username": "Nexkira",178 "avatar_template": "/letter_avatar_proxy/v4/letter/n/9fc29f/{size}.png",179 "created_at": "2021-06-29T08:04:02.995Z",180 "cooked": "<p>Do you need more information?</p>",181 "post_number": 4,182 "post_type": 1,183 "posts_count": 9,184 "updated_at": "2021-06-29T08:05:05.584Z",185 "reply_count": 0,186 "reply_to_post_number": 2,187 "quote_count": 0,188 "incoming_link_count": 69,189 "reads": 27,190 "readers_count": 26,191 "score": 350.4,192 "yours": false,193 "topic_id": 125200,194 "topic_slug": "torch-deterministic-algorithms-error",195 "display_username": "Giuseppe Torrisi",196 "primary_group_name": null,197 "flair_name": null,198 "flair_url": null,199 "flair_bg_color": null,200 "flair_color": null,201 "flair_group_id": null,202 "badges_granted": [],203 "version": 1,204 "can_edit": false,205 "can_delete": false,206 "can_recover": false,207 "can_see_hidden_post": false,208 "can_wiki": false,209 "read": true,210 "user_title": null,211 "reply_to_user": {212 "id": 3534,213 "username": "ptrblck",214 "name": "",215 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png"216 },217 "bookmarked": false,218 "actions_summary": [],219 "moderator": false,220 "admin": false,221 "staff": false,222 "user_id": 46670,223 "hidden": false,224 "trust_level": 1,225 "deleted_at": null,226 "user_deleted": false,227 "edit_reason": null,228 "can_view_edit_history": true,229 "wiki": false,230 "post_url": "/t/torch-deterministic-algorithms-error/125200/4",231 "can_accept_answer": false,232 "can_unaccept_answer": false,233 "accepted_answer": false,234 "topic_accepted_answer": null235 },236 {237 "id": 293221,238 "name": "Stavros Sykiotis",239 "username": "ssykiotis",240 "avatar_template": "/user_avatar/discuss.pytorch.org/ssykiotis/{size}/39814_2.png",241 "created_at": "2021-06-30T08:15:47.573Z",242 "cooked": "<p>Hi, I am assuming that you want to make the output of your model deterministic. I had exactly the similar issue in my project and solved it using the following solution:</p>\n<p>torch.backends.cudnn.deterministic = True<br>\ntorch.backends.cudnn.benchmark = False<br>\ntorch.manual_seed(seed)<br>\ntorch.cuda.manual_seed(seed)<br>\ntorch.cuda.manual_seed_all(seed)</p>\n<p>Avoid use_deterministic_algorithms(True). The above code solved the determinism issue for me. If you run your code on colab, a couple of additional lines are needed, let me know if that is the case so that I can provide it to you.</p>",243 "post_number": 5,244 "post_type": 1,245 "posts_count": 9,246 "updated_at": "2021-06-30T09:31:37.912Z",247 "reply_count": 2,248 "reply_to_post_number": null,249 "quote_count": 0,250 "incoming_link_count": 85,251 "reads": 28,252 "readers_count": 27,253 "score": 465.6,254 "yours": false,255 "topic_id": 125200,256 "topic_slug": "torch-deterministic-algorithms-error",257 "display_username": "Stavros Sykiotis",258 "primary_group_name": null,259 "flair_name": null,260 "flair_url": null,261 "flair_bg_color": null,262 "flair_color": null,263 "flair_group_id": null,264 "badges_granted": [],265 "version": 1,266 "can_edit": false,267 "can_delete": false,268 "can_recover": false,269 "can_see_hidden_post": false,270 "can_wiki": false,271 "read": true,272 "user_title": null,273 "bookmarked": false,274 "actions_summary": [275 {276 "id": 2,277 "count": 2278 }279 ],280 "moderator": false,281 "admin": false,282 "staff": false,283 "user_id": 46756,284 "hidden": false,285 "trust_level": 0,286 "deleted_at": null,287 "user_deleted": false,288 "edit_reason": null,289 "can_view_edit_history": true,290 "wiki": false,291 "post_url": "/t/torch-deterministic-algorithms-error/125200/5",292 "can_accept_answer": false,293 "can_unaccept_answer": false,294 "accepted_answer": false,295 "topic_accepted_answer": null296 },297 {298 "id": 293222,299 "name": "Giuseppe Torrisi",300 "username": "Nexkira",301 "avatar_template": "/letter_avatar_proxy/v4/letter/n/9fc29f/{size}.png",302 "created_at": "2021-06-30T08:36:54.873Z",303 "cooked": "<p>Yes, thank you, this fix my problem. I think torch.cuda.manual_seed and seed_all no need it cause torch.manual_seed should enable both cpu and gpu</p>",304 "post_number": 6,305 "post_type": 1,306 "posts_count": 9,307 "updated_at": "2021-06-30T08:36:54.873Z",308 "reply_count": 1,309 "reply_to_post_number": 5,310 "quote_count": 0,311 "incoming_link_count": 46,312 "reads": 28,313 "readers_count": 27,314 "score": 240.6,315 "yours": false,316 "topic_id": 125200,317 "topic_slug": "torch-deterministic-algorithms-error",318 "display_username": "Giuseppe Torrisi",319 "primary_group_name": null,320 "flair_name": null,321 "flair_url": null,322 "flair_bg_color": null,323 "flair_color": null,324 "flair_group_id": null,325 "badges_granted": [],326 "version": 1,327 "can_edit": false,328 "can_delete": false,329 "can_recover": false,330 "can_see_hidden_post": false,331 "can_wiki": false,332 "read": true,333 "user_title": null,334 "reply_to_user": {335 "id": 46756,336 "username": "ssykiotis",337 "name": "Stavros Sykiotis",338 "avatar_template": "/user_avatar/discuss.pytorch.org/ssykiotis/{size}/39814_2.png"339 },340 "bookmarked": false,341 "actions_summary": [],342 "moderator": false,343 "admin": false,344 "staff": false,345 "user_id": 46670,346 "hidden": false,347 "trust_level": 1,348 "deleted_at": null,349 "user_deleted": false,350 "edit_reason": null,351 "can_view_edit_history": true,352 "wiki": false,353 "post_url": "/t/torch-deterministic-algorithms-error/125200/6",354 "can_accept_answer": false,355 "can_unaccept_answer": false,356 "accepted_answer": false,357 "topic_accepted_answer": null358 },359 {360 "id": 293227,361 "name": "",362 "username": "ptrblck",363 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",364 "created_at": "2021-06-30T09:12:25.441Z",365 "cooked": "<p>It’s not sufficient to only seed the code and disable non-deterministic algorithms in cudnn as described in the <a href=\"https://pytorch.org/docs/stable/notes/randomness.html\">reproducibility docs</a>, so the suggestion is not a solution.<br>\nDo not avoid using <code>torch.use_deterministic_algorithms(True)</code>, as it’s the proper way to get a full deterministic code.</p>\n<p>In any case, thanks for the additional information. We are tracking the issue <a href=\"https://github.com/pytorch/pytorch/issues/61032\">here</a> and will fix it.</p>",366 "post_number": 7,367 "post_type": 1,368 "posts_count": 9,369 "updated_at": "2021-06-30T09:13:17.397Z",370 "reply_count": 0,371 "reply_to_post_number": 6,372 "quote_count": 0,373 "incoming_link_count": 25,374 "reads": 27,375 "readers_count": 26,376 "score": 130.4,377 "yours": false,378 "topic_id": 125200,379 "topic_slug": "torch-deterministic-algorithms-error",380 "display_username": "",381 "primary_group_name": null,382 "flair_name": null,383 "flair_url": null,384 "flair_bg_color": null,385 "flair_color": null,386 "flair_group_id": null,387 "badges_granted": [],388 "version": 1,389 "can_edit": false,390 "can_delete": false,391 "can_recover": false,392 "can_see_hidden_post": false,393 "can_wiki": false,394 "link_counts": [395 {396 "url": "https://github.com/pytorch/pytorch/issues/61032",397 "internal": false,398 "reflection": false,399 "title": "Deterministic indexing operation fails in indices size check / missing broadcast · Issue #61032 · pytorch/pytorch · GitHub",400 "clicks": 224401 },402 {403 "url": "https://pytorch.org/docs/stable/notes/randomness.html",404 "internal": false,405 "reflection": false,406 "title": "Reproducibility — PyTorch 1.9.0 documentation",407 "clicks": 35408 }409 ],410 "read": true,411 "user_title": "",412 "reply_to_user": {413 "id": 46670,414 "username": "Nexkira",415 "name": "Giuseppe Torrisi",416 "avatar_template": "/letter_avatar_proxy/v4/letter/n/9fc29f/{size}.png"417 },418 "bookmarked": false,419 "actions_summary": [],420 "moderator": true,421 "admin": true,422 "staff": true,423 "user_id": 3534,424 "hidden": false,425 "trust_level": 2,426 "deleted_at": null,427 "user_deleted": false,428 "edit_reason": null,429 "can_view_edit_history": true,430 "wiki": false,431 "post_url": "/t/torch-deterministic-algorithms-error/125200/7",432 "can_accept_answer": false,433 "can_unaccept_answer": false,434 "accepted_answer": false,435 "topic_accepted_answer": null436 },437 {438 "id": 293232,439 "name": "Giuseppe Torrisi",440 "username": "Nexkira",441 "avatar_template": "/letter_avatar_proxy/v4/letter/n/9fc29f/{size}.png",442 "created_at": "2021-06-30T09:31:33.239Z",443 "cooked": "<p>oh ok thank you for making clear, I’ll wait when the fix is ready</p>",444 "post_number": 8,445 "post_type": 1,446 "posts_count": 9,447 "updated_at": "2021-06-30T09:31:33.239Z",448 "reply_count": 0,449 "reply_to_post_number": null,450 "quote_count": 0,451 "incoming_link_count": 8,452 "reads": 26,453 "readers_count": 25,454 "score": 45.2,455 "yours": false,456 "topic_id": 125200,457 "topic_slug": "torch-deterministic-algorithms-error",458 "display_username": "Giuseppe Torrisi",459 "primary_group_name": null,460 "flair_name": null,461 "flair_url": null,462 "flair_bg_color": null,463 "flair_color": null,464 "flair_group_id": null,465 "badges_granted": [],466 "version": 1,467 "can_edit": false,468 "can_delete": false,469 "can_recover": false,470 "can_see_hidden_post": false,471 "can_wiki": false,472 "read": true,473 "user_title": null,474 "bookmarked": false,475 "actions_summary": [],476 "moderator": false,477 "admin": false,478 "staff": false,479 "user_id": 46670,480 "hidden": false,481 "trust_level": 1,482 "deleted_at": null,483 "user_deleted": false,484 "edit_reason": null,485 "can_view_edit_history": true,486 "wiki": false,487 "post_url": "/t/torch-deterministic-algorithms-error/125200/8",488 "can_accept_answer": false,489 "can_unaccept_answer": false,490 "accepted_answer": false,491 "topic_accepted_answer": null492 },493 {494 "id": 334716,495 "name": "Chuyang Zhao",496 "username": "Chuyang",497 "avatar_template": "/user_avatar/discuss.pytorch.org/chuyang/{size}/34516_2.png",498 "created_at": "2022-03-07T05:31:35.653Z",499 "cooked": "<p>Thanks! Solved my issue too</p>",500 "post_number": 9,501 "post_type": 1,502 "posts_count": 9,503 "updated_at": "2022-03-07T05:31:35.653Z",504 "reply_count": 0,505 "reply_to_post_number": 5,506 "quote_count": 0,507 "incoming_link_count": 4,508 "reads": 14,509 "readers_count": 13,510 "score": 22.8,511 "yours": false,512 "topic_id": 125200,513 "topic_slug": "torch-deterministic-algorithms-error",514 "display_username": "Chuyang Zhao",515 "primary_group_name": null,516 "flair_name": null,517 "flair_url": null,518 "flair_bg_color": null,519 "flair_color": null,520 "flair_group_id": null,521 "badges_granted": [],522 "version": 1,523 "can_edit": false,524 "can_delete": false,525 "can_recover": false,526 "can_see_hidden_post": false,527 "can_wiki": false,528 "read": true,529 "user_title": null,530 "reply_to_user": {531 "id": 46756,532 "username": "ssykiotis",533 "name": "Stavros Sykiotis",534 "avatar_template": "/user_avatar/discuss.pytorch.org/ssykiotis/{size}/39814_2.png"535 },536 "bookmarked": false,537 "actions_summary": [],538 "moderator": false,539 "admin": false,540 "staff": false,541 "user_id": 41953,542 "hidden": false,543 "trust_level": 1,544 "deleted_at": null,545 "user_deleted": false,546 "edit_reason": null,547 "can_view_edit_history": true,548 "wiki": false,549 "post_url": "/t/torch-deterministic-algorithms-error/125200/9",550 "can_accept_answer": false,551 "can_unaccept_answer": false,552 "accepted_answer": false,553 "topic_accepted_answer": null554 }555 ],556 "stream": [557 292603,558 292727,559 292736,560 292963,561 293221,562 293222,563 293227,564 293232,565 334716566 ]567 },568 "timeline_lookup": [569 [570 1,571 1581572 ],573 [574 3,575 1580576 ],577 [578 5,579 1579580 ],581 [582 7,583 1578584 ],585 [586 9,587 1329588 ]589 ],590 "suggested_topics": [591 {592 "fancy_title": "Why did my customized KL divergence cause NAN?",593 "id": 214339,594 "title": "Why did my customized KL divergence cause NAN?",595 "slug": "why-did-my-customized-kl-divergence-cause-nan",596 "posts_count": 2,597 "reply_count": 0,598 "highest_post_number": 2,599 "image_url": null,600 "created_at": "2024-12-18T00:23:20.532Z",601 "last_posted_at": "2024-12-22T03:13:56.957Z",602 "bumped": true,603 "bumped_at": "2024-12-22T03:13:56.957Z",604 "archetype": "regular",605 "unseen": false,606 "pinned": false,607 "unpinned": null,608 "visible": true,609 "closed": false,610 "archived": false,611 "bookmarked": null,612 "liked": null,613 "tags_descriptions": {},614 "like_count": 2,615 "views": 206,616 "category_id": 5,617 "featured_link": null,618 "has_accepted_answer": true,619 "posters": [620 {621 "extras": null,622 "description": "Original Poster",623 "user": {624 "id": 45116,625 "username": "JimW",626 "name": "",627 "avatar_template": "/user_avatar/discuss.pytorch.org/jimw/{size}/38000_2.png",628 "trust_level": 1629 }630 },631 {632 "extras": "latest",633 "description": "Most Recent Poster, Accepted Answer",634 "user": {635 "id": 18088,636 "username": "KFrank",637 "name": "K. Frank",638 "avatar_template": "/letter_avatar_proxy/v4/letter/k/ecb155/{size}.png",639 "trust_level": 2640 }641 }642 ]643 },644 {645 "fancy_title": "Extra GPU usage on custom Qwen2-VL",646 "id": 212198,647 "title": "Extra GPU usage on custom Qwen2-VL",648 "slug": "extra-gpu-usage-on-custom-qwen2-vl",649 "posts_count": 1,650 "reply_count": 0,651 "highest_post_number": 1,652 "image_url": "https://discuss.pytorch.org/uploads/default/optimized/3X/1/2/1284e7c5c2e5ddde6a34f371a02bf4c24b2f01f6_2_1023x660.png",653 "created_at": "2024-10-28T09:05:40.136Z",654 "last_posted_at": "2024-10-28T09:05:40.194Z",655 "bumped": true,656 "bumped_at": "2024-10-28T09:05:40.194Z",657 "archetype": "regular",658 "unseen": false,659 "pinned": false,660 "unpinned": null,661 "visible": true,662 "closed": false,663 "archived": false,664 "bookmarked": null,665 "liked": null,666 "tags_descriptions": {},667 "like_count": 0,668 "views": 237,669 "category_id": 5,670 "featured_link": null,671 "has_accepted_answer": false,672 "posters": [673 {674 "extras": "latest single",675 "description": "Original Poster, Most Recent Poster",676 "user": {677 "id": 68592,678 "username": "ChanYalcn",679 "name": "Cihan Yalçın",680 "avatar_template": "/user_avatar/discuss.pytorch.org/chanyalcn/{size}/63089_2.png",681 "trust_level": 1682 }683 }684 ]685 },686 {687 "fancy_title": "Issue in the curve of training",688 "id": 214245,689 "title": "Issue in the curve of training",690 "slug": "issue-in-the-curve-of-training",691 "posts_count": 6,692 "reply_count": 4,693 "highest_post_number": 6,694 "image_url": "https://discuss.pytorch.org/uploads/default/optimized/3X/c/a/cacc24171efe3828c29ec1dabee719d95fec2ebb_2_1024x588.png",695 "created_at": "2024-12-15T15:23:43.751Z",696 "last_posted_at": "2024-12-15T21:16:13.480Z",697 "bumped": true,698 "bumped_at": "2024-12-15T21:16:13.480Z",699 "archetype": "regular",700 "unseen": false,701 "pinned": false,702 "unpinned": null,703 "visible": true,704 "closed": false,705 "archived": false,706 "bookmarked": null,707 "liked": null,708 "tags_descriptions": {},709 "like_count": 2,710 "views": 37,711 "category_id": 5,712 "featured_link": null,713 "has_accepted_answer": true,714 "posters": [715 {716 "extras": null,717 "description": "Original Poster",718 "user": {719 "id": 43052,720 "username": "samm",721 "name": "",722 "avatar_template": "/user_avatar/discuss.pytorch.org/samm/{size}/36169_2.png",723 "trust_level": 1724 }725 },726 {727 "extras": "latest",728 "description": "Most Recent Poster, Accepted Answer",729 "user": {730 "id": 3534,731 "username": "ptrblck",732 "name": "",733 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",734 "admin": true,735 "moderator": true,736 "trust_level": 2737 }738 }739 ]740 },741 {742 "fancy_title": "How can I implement operations on queue with supported MIL ops?",743 "id": 213763,744 "title": "How can I implement operations on queue with supported MIL ops?",745 "slug": "how-can-i-implement-operations-on-queue-with-supported-mil-ops",746 "posts_count": 1,747 "reply_count": 0,748 "highest_post_number": 1,749 "image_url": null,750 "created_at": "2024-12-03T22:35:04.894Z",751 "last_posted_at": "2024-12-03T22:35:04.945Z",752 "bumped": true,753 "bumped_at": "2024-12-03T22:35:04.945Z",754 "archetype": "regular",755 "unseen": false,756 "pinned": false,757 "unpinned": null,758 "visible": true,759 "closed": false,760 "archived": false,761 "bookmarked": null,762 "liked": null,763 "tags_descriptions": {},764 "like_count": 0,765 "views": 18,766 "category_id": 5,767 "featured_link": null,768 "has_accepted_answer": false,769 "posters": [770 {771 "extras": "latest single",772 "description": "Original Poster, Most Recent Poster",773 "user": {774 "id": 45116,775 "username": "JimW",776 "name": "",777 "avatar_template": "/user_avatar/discuss.pytorch.org/jimw/{size}/38000_2.png",778 "trust_level": 1779 }780 }781 ]782 },783 {784 "fancy_title": "Change vit_b_16 input size",785 "id": 214515,786 "title": "Change vit_b_16 input size",787 "slug": "change-vit-b-16-input-size",788 "posts_count": 2,789 "reply_count": 0,790 "highest_post_number": 2,791 "image_url": null,792 "created_at": "2024-12-21T18:00:08.693Z",793 "last_posted_at": "2024-12-21T19:00:07.503Z",794 "bumped": true,795 "bumped_at": "2024-12-21T19:00:07.503Z",796 "archetype": "regular",797 "unseen": false,798 "pinned": false,799 "unpinned": null,800 "visible": true,801 "closed": false,802 "archived": false,803 "bookmarked": null,804 "liked": null,805 "tags_descriptions": {},806 "like_count": 1,807 "views": 193,808 "category_id": 5,809 "featured_link": null,810 "has_accepted_answer": false,811 "posters": [812 {813 "extras": null,814 "description": "Original Poster",815 "user": {816 "id": 81646,817 "username": "bruhnugget-nice",818 "name": "bruhnugget",819 "avatar_template": "/user_avatar/discuss.pytorch.org/bruhnugget-nice/{size}/74667_2.png",820 "trust_level": 1821 }822 },823 {824 "extras": "latest",825 "description": "Most Recent Poster",826 "user": {827 "id": 3534,828 "username": "ptrblck",829 "name": "",830 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",831 "admin": true,832 "moderator": true,833 "trust_level": 2834 }835 }836 ]837 }838 ],839 "tags_descriptions": {},840 "fancy_title": "Torch deterministic algorithms error",841 "id": 125200,842 "title": "Torch deterministic algorithms error",843 "posts_count": 9,844 "created_at": "2021-06-27T21:32:23.393Z",845 "views": 3244,846 "reply_count": 5,847 "like_count": 2,848 "last_posted_at": "2022-03-07T05:31:35.653Z",849 "visible": true,850 "closed": false,851 "archived": false,852 "has_summary": false,853 "archetype": "regular",854 "slug": "torch-deterministic-algorithms-error",855 "category_id": 5,856 "word_count": 704,857 "deleted_at": null,858 "user_id": 46670,859 "featured_link": null,860 "pinned_globally": false,861 "pinned_at": null,862 "pinned_until": null,863 "image_url": null,864 "slow_mode_seconds": 0,865 "draft": null,866 "draft_key": "topic_125200",867 "draft_sequence": null,868 "unpinned": null,869 "pinned": false,870 "current_post_number": 1,871 "highest_post_number": 9,872 "deleted_by": null,873 "actions_summary": [874 {875 "id": 4,876 "count": 0,877 "hidden": false,878 "can_act": false879 },880 {881 "id": 8,882 "count": 0,883 "hidden": false,884 "can_act": false885 },886 {887 "id": 10,888 "count": 0,889 "hidden": false,890 "can_act": false891 },892 {893 "id": 7,894 "count": 0,895 "hidden": false,896 "can_act": false897 }898 ],899 "chunk_size": 20,900 "bookmarked": false,901 "topic_timer": null,902 "message_bus_last_id": 0,903 "participant_count": 4,904 "show_read_indicator": false,905 "thumbnails": null,906 "slow_mode_enabled_until": null,907 "can_vote": false,908 "vote_count": 0,909 "user_voted": false,910 "discourse_zendesk_plugin_zendesk_id": null,911 "discourse_zendesk_plugin_zendesk_url": "https://your-url.zendesk.com/agent/tickets/",912 "details": {913 "can_edit": false,914 "notification_level": 1,915 "participants": [916 {917 "id": 46670,918 "username": "Nexkira",919 "name": "Giuseppe Torrisi",920 "avatar_template": "/letter_avatar_proxy/v4/letter/n/9fc29f/{size}.png",921 "post_count": 5,922 "primary_group_name": null,923 "flair_name": null,924 "flair_url": null,925 "flair_color": null,926 "flair_bg_color": null,927 "flair_group_id": null,928 "trust_level": 1929 },930 {931 "id": 3534,932 "username": "ptrblck",933 "name": "",934 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",935 "post_count": 2,936 "primary_group_name": null,937 "flair_name": null,938 "flair_url": null,939 "flair_color": null,940 "flair_bg_color": null,941 "flair_group_id": null,942 "admin": true,943 "moderator": true,944 "trust_level": 2945 },946 {947 "id": 41953,948 "username": "Chuyang",949 "name": "Chuyang Zhao",950 "avatar_template": "/user_avatar/discuss.pytorch.org/chuyang/{size}/34516_2.png",951 "post_count": 1,952 "primary_group_name": null,953 "flair_name": null,954 "flair_url": null,955 "flair_color": null,956 "flair_bg_color": null,957 "flair_group_id": null,958 "trust_level": 1959 },960 {961 "id": 46756,962 "username": "ssykiotis",963 "name": "Stavros Sykiotis",964 "avatar_template": "/user_avatar/discuss.pytorch.org/ssykiotis/{size}/39814_2.png",965 "post_count": 1,966 "primary_group_name": null,967 "flair_name": null,968 "flair_url": null,969 "flair_color": null,970 "flair_bg_color": null,971 "flair_group_id": null,972 "trust_level": 0973 }974 ],975 "created_by": {976 "id": 46670,977 "username": "Nexkira",978 "name": "Giuseppe Torrisi",979 "avatar_template": "/letter_avatar_proxy/v4/letter/n/9fc29f/{size}.png"980 },981 "last_poster": {982 "id": 41953,983 "username": "Chuyang",984 "name": "Chuyang Zhao",985 "avatar_template": "/user_avatar/discuss.pytorch.org/chuyang/{size}/34516_2.png"986 },987 "links": [988 {989 "url": "https://github.com/pytorch/pytorch/issues/61032",990 "title": "Deterministic indexing operation fails in indices size check / missing broadcast · Issue #61032 · pytorch/pytorch · GitHub",991 "internal": false,992 "attachment": false,993 "reflection": false,994 "clicks": 224,995 "user_id": 3534,996 "domain": "github.com",997 "root_domain": "github.com"998 },999 {1000 "url": "https://pytorch.org/docs/stable/notes/randomness.html",1001 "title": "Reproducibility — PyTorch 1.9.0 documentation",1002 "internal": false,1003 "attachment": false,1004 "reflection": false,1005 "clicks": 35,1006 "user_id": 3534,1007 "domain": "pytorch.org",1008 "root_domain": "pytorch.org"1009 }1010 ]1011 },1012 "bookmarks": []1013 },1014 {1015 "post_stream": {1016 "posts": [1017 {1018 "id": 334628,1019 "name": "FeHng",1020 "username": "FeHng",1021 "avatar_template": "/user_avatar/discuss.pytorch.org/fehng/{size}/47275_2.png",1022 "created_at": "2022-03-06T07:42:18.066Z",1023 "cooked": "<p>I have a conv layer with input_channel=10 and output_channel=20. Usually I need take feature maps with shape [B,H,W,C=10] as input. But I want use different channel at different time, the number of channels will not be greater than 10. For example, I want take feature maps with shape [B,H,W,C=5] as input. That means, in theory, only half of the kernel will be used, which makes a speed up.<br>\nThe easiest way is to pad the feature maps to full shape,but I don’t want do this. Because this means I can’t get acceleration.</p>",1024 "post_number": 1,1025 "post_type": 1,1026 "posts_count": 2,1027 "updated_at": "2022-03-06T07:42:18.066Z",1028 "reply_count": 0,1029 "reply_to_post_number": null,1030 "quote_count": 0,1031 "incoming_link_count": 13,1032 "reads": 4,1033 "readers_count": 3,1034 "score": 65.8,1035 "yours": false,1036 "topic_id": 145726,1037 "topic_slug": "use-some-of-the-input-channel-of-the-convolution-kernel",1038 "display_username": "FeHng",1039 "primary_group_name": null,1040 "flair_name": null,1041 "flair_url": null,1042 "flair_bg_color": null,1043 "flair_color": null,1044 "flair_group_id": null,1045 "badges_granted": [],1046 "version": 1,1047 "can_edit": false,1048 "can_delete": false,1049 "can_recover": false,1050 "can_see_hidden_post": false,1051 "can_wiki": false,1052 "read": true,1053 "user_title": null,1054 "bookmarked": false,1055 "actions_summary": [],1056 "moderator": false,1057 "admin": false,1058 "staff": false,1059 "user_id": 53781,1060 "hidden": false,1061 "trust_level": 1,1062 "deleted_at": null,1063 "user_deleted": false,1064 "edit_reason": null,1065 "can_view_edit_history": true,1066 "wiki": false,1067 "post_url": "/t/use-some-of-the-input-channel-of-the-convolution-kernel/145726/1",1068 "can_accept_answer": false,1069 "can_unaccept_answer": false,1070 "accepted_answer": false,1071 "topic_accepted_answer": null,1072 "can_vote": false1073 },1074 {1075 "id": 334707,1076 "name": "",1077 "username": "ptrblck",1078 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",1079 "created_at": "2022-03-07T02:09:59.956Z",1080 "cooked": "<p>You could slice the conv kernel and use the functional API via <code>F.conv2d</code>.</p>",1081 "post_number": 2,1082 "post_type": 1,1083 "posts_count": 2,1084 "updated_at": "2022-03-07T02:09:59.956Z",1085 "reply_count": 0,1086 "reply_to_post_number": null,1087 "quote_count": 0,1088 "incoming_link_count": 0,1089 "reads": 3,1090 "readers_count": 2,1091 "score": 0.6,1092 "yours": false,1093 "topic_id": 145726,1094 "topic_slug": "use-some-of-the-input-channel-of-the-convolution-kernel",1095 "display_username": "",1096 "primary_group_name": null,1097 "flair_name": null,1098 "flair_url": null,1099 "flair_bg_color": null,1100 "flair_color": null,1101 "flair_group_id": null,1102 "badges_granted": [],1103 "version": 1,1104 "can_edit": false,1105 "can_delete": false,1106 "can_recover": false,1107 "can_see_hidden_post": false,1108 "can_wiki": false,1109 "read": true,1110 "user_title": "",1111 "bookmarked": false,1112 "actions_summary": [],1113 "moderator": true,1114 "admin": true,1115 "staff": true,1116 "user_id": 3534,1117 "hidden": false,1118 "trust_level": 2,1119 "deleted_at": null,1120 "user_deleted": false,1121 "edit_reason": null,1122 "can_view_edit_history": true,1123 "wiki": false,1124 "post_url": "/t/use-some-of-the-input-channel-of-the-convolution-kernel/145726/2",1125 "can_accept_answer": false,1126 "can_unaccept_answer": false,1127 "accepted_answer": false,1128 "topic_accepted_answer": null1129 }1130 ],1131 "stream": [1132 334628,1133 3347071134 ]1135 },1136 "timeline_lookup": [1137 [1138 1,1139 13301140 ],1141 [1142 2,1143 13291144 ]1145 ],1146 "suggested_topics": [1147 {1148 "fancy_title": "Premultiplied alpha",1149 "id": 220230,1150 "title": "Premultiplied alpha",1151 "slug": "premultiplied-alpha",1152 "posts_count": 4,1153 "reply_count": 2,1154 "highest_post_number": 4,1155 "image_url": null,1156 "created_at": "2025-05-21T16:40:18.345Z",1157 "last_posted_at": "2025-05-25T18:54:50.339Z",1158 "bumped": true,1159 "bumped_at": "2025-05-25T18:54:50.339Z",1160 "archetype": "regular",1161 "unseen": false,1162 "pinned": false,1163 "unpinned": null,1164 "visible": true,1165 "closed": false,1166 "archived": false,1167 "bookmarked": null,1168 "liked": null,1169 "tags_descriptions": {},1170 "like_count": 1,1171 "views": 81,1172 "category_id": 5,1173 "featured_link": null,1174 "has_accepted_answer": true,1175 "posters": [1176 {1177 "extras": null,1178 "description": "Original Poster",1179 "user": {1180 "id": 84391,1181 "username": "rdgq",1182 "name": "",1183 "avatar_template": "/user_avatar/discuss.pytorch.org/rdgq/{size}/77106_2.png",1184 "trust_level": 11185 }1186 },1187 {1188 "extras": "latest",1189 "description": "Most Recent Poster, Accepted Answer",1190 "user": {1191 "id": 75120,1192 "username": "ajayrkumar",1193 "name": "Ajay Rajendra Kumar",1194 "avatar_template": "/user_avatar/discuss.pytorch.org/ajayrkumar/{size}/77139_2.png",1195 "trust_level": 11196 }1197 }1198 ]1199 },1200 {