Anurag1734/cuda-error-resolution-analysis
07
1[2 {3 "post_stream": {4 "posts": [5 {6 "id": 285550,7 "name": "Hadi Nayebi",8 "username": "Hadi_Nayebi",9 "avatar_template": "/user_avatar/discuss.pytorch.org/hadi_nayebi/{size}/38617_2.png",10 "created_at": "2021-05-23T00:48:33.946Z",11 "cooked": "<p>Hey, I hope everyone is having a great day.<br>\nI have the following setup:</p>\n<p>state = netA(input)<br>\noutput1=netB(state)<br>\noutput2=netC(state)</p>\n<p>I can define target values for output1 and output2. How would you approach training the netA?<br>\nI am trying to combine the gradient on the input of netB and netC (state) as the loss to train netA.<br>\nor should I use them separately?!!!<br>\nhow can I get the gradient on the input of netB and netC?</p>\n<p>Best,<br>\nhn</p>",12 "post_number": 1,13 "post_type": 1,14 "posts_count": 3,15 "updated_at": "2021-05-23T00:48:33.946Z",16 "reply_count": 0,17 "reply_to_post_number": null,18 "quote_count": 0,19 "incoming_link_count": 29,20 "reads": 9,21 "readers_count": 8,22 "score": 156.8,23 "yours": false,24 "topic_id": 122078,25 "topic_slug": "joining-multiple-neural-nets-in-non-sequential-model",26 "display_username": "Hadi Nayebi",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 "read": true,41 "user_title": null,42 "bookmarked": false,43 "actions_summary": [44 {45 "id": 2,46 "count": 147 }48 ],49 "moderator": false,50 "admin": false,51 "staff": false,52 "user_id": 45640,53 "hidden": false,54 "trust_level": 1,55 "deleted_at": null,56 "user_deleted": false,57 "edit_reason": null,58 "can_view_edit_history": true,59 "wiki": false,60 "post_url": "/t/joining-multiple-neural-nets-in-non-sequential-model/122078/1",61 "can_accept_answer": false,62 "can_unaccept_answer": false,63 "accepted_answer": false,64 "topic_accepted_answer": null,65 "can_vote": false66 },67 {68 "id": 285603,69 "name": "",70 "username": "ptrblck",71 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",72 "created_at": "2021-05-23T08:11:09.677Z",73 "cooked": "<p>You could calculate the losses from using <code>output1</code> and <code>output2</code>, accumulate them, and call <code>loss.backward()</code> to calculate the gradients for all 3 models. If you don’t want to train <code>netB</code> and <code>netC</code>, you could set the <code>.requires_grad</code> attribute if their parameters to <code>False</code>.</p>\n<p>In case you want to get the gradients in <code>state</code>, you could call <code>state.retrain_grad()</code>, which would then allow you to access its <code>.grad</code> after the <code>backward</code> call.</p>",74 "post_number": 2,75 "post_type": 1,76 "posts_count": 3,77 "updated_at": "2021-05-23T08:12:52.713Z",78 "reply_count": 1,79 "reply_to_post_number": null,80 "quote_count": 0,81 "incoming_link_count": 1,82 "reads": 8,83 "readers_count": 7,84 "score": 26.6,85 "yours": false,86 "topic_id": 122078,87 "topic_slug": "joining-multiple-neural-nets-in-non-sequential-model",88 "display_username": "",89 "primary_group_name": null,90 "flair_name": null,91 "flair_url": null,92 "flair_bg_color": null,93 "flair_color": null,94 "flair_group_id": null,95 "badges_granted": [],96 "version": 1,97 "can_edit": false,98 "can_delete": false,99 "can_recover": false,100 "can_see_hidden_post": false,101 "can_wiki": false,102 "read": true,103 "user_title": "",104 "bookmarked": false,105 "actions_summary": [106 {107 "id": 2,108 "count": 1109 }110 ],111 "moderator": true,112 "admin": true,113 "staff": true,114 "user_id": 3534,115 "hidden": false,116 "trust_level": 2,117 "deleted_at": null,118 "user_deleted": false,119 "edit_reason": null,120 "can_view_edit_history": true,121 "wiki": false,122 "post_url": "/t/joining-multiple-neural-nets-in-non-sequential-model/122078/2",123 "can_accept_answer": false,124 "can_unaccept_answer": false,125 "accepted_answer": false,126 "topic_accepted_answer": null127 },128 {129 "id": 287127,130 "name": "Hadi Nayebi",131 "username": "Hadi_Nayebi",132 "avatar_template": "/user_avatar/discuss.pytorch.org/hadi_nayebi/{size}/38617_2.png",133 "created_at": "2021-05-31T02:28:27.914Z",134 "cooked": "<p>Thank you ptrblck. I will try your suggestions and get back to you if I had more questions, I am a newbie, and things are slow on my side.</p>",135 "post_number": 3,136 "post_type": 1,137 "posts_count": 3,138 "updated_at": "2021-05-31T02:28:27.914Z",139 "reply_count": 0,140 "reply_to_post_number": 2,141 "quote_count": 0,142 "incoming_link_count": 0,143 "reads": 6,144 "readers_count": 5,145 "score": 1.2,146 "yours": false,147 "topic_id": 122078,148 "topic_slug": "joining-multiple-neural-nets-in-non-sequential-model",149 "display_username": "Hadi Nayebi",150 "primary_group_name": null,151 "flair_name": null,152 "flair_url": null,153 "flair_bg_color": null,154 "flair_color": null,155 "flair_group_id": null,156 "badges_granted": [],157 "version": 1,158 "can_edit": false,159 "can_delete": false,160 "can_recover": false,161 "can_see_hidden_post": false,162 "can_wiki": false,163 "read": true,164 "user_title": null,165 "reply_to_user": {166 "id": 3534,167 "username": "ptrblck",168 "name": "",169 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png"170 },171 "bookmarked": false,172 "actions_summary": [],173 "moderator": false,174 "admin": false,175 "staff": false,176 "user_id": 45640,177 "hidden": false,178 "trust_level": 1,179 "deleted_at": null,180 "user_deleted": false,181 "edit_reason": null,182 "can_view_edit_history": true,183 "wiki": false,184 "post_url": "/t/joining-multiple-neural-nets-in-non-sequential-model/122078/3",185 "can_accept_answer": false,186 "can_unaccept_answer": false,187 "accepted_answer": false,188 "topic_accepted_answer": null189 }190 ],191 "stream": [192 285550,193 285603,194 287127195 ]196 },197 "timeline_lookup": [198 [199 1,200 1617201 ],202 [203 3,204 1609205 ]206 ],207 "suggested_topics": [208 {209 "fancy_title": "How to run PyTorch DDP with Hydra + Optuna?",210 "id": 213316,211 "title": "How to run PyTorch DDP with Hydra + Optuna?",212 "slug": "how-to-run-pytorch-ddp-with-hydra-optuna",213 "posts_count": 1,214 "reply_count": 0,215 "highest_post_number": 1,216 "image_url": null,217 "created_at": "2024-11-22T17:40:03.638Z",218 "last_posted_at": "2024-11-22T17:40:03.696Z",219 "bumped": true,220 "bumped_at": "2024-11-22T17:40:03.696Z",221 "archetype": "regular",222 "unseen": false,223 "pinned": false,224 "unpinned": null,225 "visible": true,226 "closed": false,227 "archived": false,228 "bookmarked": null,229 "liked": null,230 "tags_descriptions": {},231 "like_count": 0,232 "views": 235,233 "category_id": 1,234 "featured_link": null,235 "has_accepted_answer": false,236 "posters": [237 {238 "extras": "latest single",239 "description": "Original Poster, Most Recent Poster",240 "user": {241 "id": 81072,242 "username": "deepakpokkalla",243 "name": "Deepak Pokkalla",244 "avatar_template": "/user_avatar/discuss.pytorch.org/deepakpokkalla/{size}/74136_2.png",245 "trust_level": 1246 }247 }248 ]249 },250 {251 "fancy_title": "Pytorch not compatible with sm_86 CUDA Capability",252 "id": 215283,253 "title": "Pytorch not compatible with sm_86 CUDA Capability",254 "slug": "pytorch-not-compatible-with-sm-86-cuda-capability",255 "posts_count": 5,256 "reply_count": 3,257 "highest_post_number": 5,258 "image_url": null,259 "created_at": "2025-01-11T23:30:42.751Z",260 "last_posted_at": "2025-01-12T17:56:09.703Z",261 "bumped": true,262 "bumped_at": "2025-01-12T17:56:09.703Z",263 "archetype": "regular",264 "unseen": false,265 "pinned": false,266 "unpinned": null,267 "visible": true,268 "closed": false,269 "archived": false,270 "bookmarked": null,271 "liked": null,272 "tags_descriptions": {},273 "like_count": 0,274 "views": 308,275 "category_id": 1,276 "featured_link": null,277 "has_accepted_answer": false,278 "posters": [279 {280 "extras": "latest",281 "description": "Original Poster, Most Recent Poster",282 "user": {283 "id": 82028,284 "username": "the-ml-bulldog",285 "name": "",286 "avatar_template": "/user_avatar/discuss.pytorch.org/the-ml-bulldog/{size}/75057_2.png",287 "trust_level": 1288 }289 },290 {291 "extras": null,292 "description": "Frequent Poster",293 "user": {294 "id": 3534,295 "username": "ptrblck",296 "name": "",297 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",298 "admin": true,299 "moderator": true,300 "trust_level": 2301 }302 }303 ]304 },305 {306 "fancy_title": "Why dose TorchDispatchMode miss aten::_index_put_impl_",307 "id": 215461,308 "title": "Why dose TorchDispatchMode miss aten::_index_put_impl_",309 "slug": "why-dose-torchdispatchmode-miss-aten-index-put-impl",310 "posts_count": 1,311 "reply_count": 0,312 "highest_post_number": 1,313 "image_url": "https://discuss.pytorch.org/uploads/default/optimized/3X/f/a/fa85623e202f4c6665e11f8cd7dc6fc19706c5dc_2_1024x292.png",314 "created_at": "2025-01-16T09:09:12.460Z",315 "last_posted_at": "2025-01-16T09:09:12.503Z",316 "bumped": true,317 "bumped_at": "2025-01-17T03:01:27.936Z",318 "archetype": "regular",319 "unseen": false,320 "pinned": false,321 "unpinned": null,322 "visible": true,323 "closed": false,324 "archived": false,325 "bookmarked": null,326 "liked": null,327 "tags_descriptions": {},328 "like_count": 0,329 "views": 122,330 "category_id": 1,331 "featured_link": null,332 "has_accepted_answer": false,333 "posters": [334 {335 "extras": "latest single",336 "description": "Original Poster, Most Recent Poster",337 "user": {338 "id": 82119,339 "username": "Cookie_Yang",340 "name": "Cookie Yang",341 "avatar_template": "/user_avatar/discuss.pytorch.org/cookie_yang/{size}/75125_2.png",342 "trust_level": 0343 }344 }345 ]346 },347 {348 "fancy_title": "Pytorch on 3.10 python version of jetson orin",349 "id": 216032,350 "title": "Pytorch on 3.10 python version of jetson orin",351 "slug": "pytorch-on-3-10-python-version-of-jetson-orin",352 "posts_count": 3,353 "reply_count": 1,354 "highest_post_number": 3,355 "image_url": null,356 "created_at": "2025-01-29T19:13:23.937Z",357 "last_posted_at": "2025-01-29T21:26:40.385Z",358 "bumped": true,359 "bumped_at": "2025-01-29T21:37:17.202Z",360 "archetype": "regular",361 "unseen": false,362 "pinned": false,363 "unpinned": null,364 "visible": true,365 "closed": false,366 "archived": false,367 "bookmarked": null,368 "liked": null,369 "tags_descriptions": {},370 "like_count": 0,371 "views": 399,372 "category_id": 1,373 "featured_link": null,374 "has_accepted_answer": false,375 "posters": [376 {377 "extras": "latest",378 "description": "Original Poster, Most Recent Poster",379 "user": {380 "id": 82374,381 "username": "Vaibhavi_Kavathekar",382 "name": "Vaibhavi Kavathekar",383 "avatar_template": "/user_avatar/discuss.pytorch.org/vaibhavi_kavathekar/{size}/75361_2.png",384 "trust_level": 0385 }386 },387 {388 "extras": null,389 "description": "Frequent Poster",390 "user": {391 "id": 3534,392 "username": "ptrblck",393 "name": "",394 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",395 "admin": true,396 "moderator": true,397 "trust_level": 2398 }399 }400 ]401 },402 {403 "fancy_title": "API domain – pytorch.sh",404 "id": 216277,405 "title": "API domain -- pytorch.sh",406 "slug": "api-domain-pytorch-sh",407 "posts_count": 5,408 "reply_count": 2,409 "highest_post_number": 6,410 "image_url": null,411 "created_at": "2025-02-05T15:30:28.291Z",412 "last_posted_at": "2025-02-05T22:15:58.826Z",413 "bumped": true,414 "bumped_at": "2025-02-05T22:15:58.826Z",415 "archetype": "regular",416 "unseen": false,417 "pinned": false,418 "unpinned": null,419 "visible": true,420 "closed": false,421 "archived": false,422 "bookmarked": null,423 "liked": null,424 "tags_descriptions": {},425 "like_count": 2,426 "views": 231,427 "category_id": 1,428 "featured_link": null,429 "has_accepted_answer": false,430 "posters": [431 {432 "extras": "latest",433 "description": "Original Poster, Most Recent Poster",434 "user": {435 "id": 82506,436 "username": "rosecloud",437 "name": "Frederick",438 "avatar_template": "/letter_avatar_proxy/v4/letter/r/76d3ee/{size}.png",439 "trust_level": 0440 }441 },442 {443 "extras": null,444 "description": "Frequent Poster",445 "user": {446 "id": 1,447 "username": "smth",448 "name": "",449 "avatar_template": "/user_avatar/discuss.pytorch.org/smth/{size}/13_2.png",450 "admin": true,451 "moderator": true,452 "trust_level": 2453 }454 },455 {456 "extras": null,457 "description": "Frequent Poster",458 "user": {459 "id": 75469,460 "username": "Matt_White",461 "name": "Matt White",462 "avatar_template": "/user_avatar/discuss.pytorch.org/matt_white/{size}/69693_2.png",463 "admin": true,464 "trust_level": 1465 }466 }467 ]468 }469 ],470 "tags_descriptions": {},471 "fancy_title": "Joining multiple neural nets in non-sequential model",472 "id": 122078,473 "title": "Joining multiple neural nets in non-sequential model",474 "posts_count": 3,475 "created_at": "2021-05-23T00:48:33.886Z",476 "views": 316,477 "reply_count": 1,478 "like_count": 2,479 "last_posted_at": "2021-05-31T02:28:27.914Z",480 "visible": true,481 "closed": false,482 "archived": false,483 "has_summary": false,484 "archetype": "regular",485 "slug": "joining-multiple-neural-nets-in-non-sequential-model",486 "category_id": 1,487 "word_count": 183,488 "deleted_at": null,489 "user_id": 45640,490 "featured_link": null,491 "pinned_globally": false,492 "pinned_at": null,493 "pinned_until": null,494 "image_url": null,495 "slow_mode_seconds": 0,496 "draft": null,497 "draft_key": "topic_122078",498 "draft_sequence": null,499 "unpinned": null,500 "pinned": false,501 "current_post_number": 1,502 "highest_post_number": 3,503 "deleted_by": null,504 "actions_summary": [505 {506 "id": 4,507 "count": 0,508 "hidden": false,509 "can_act": false510 },511 {512 "id": 8,513 "count": 0,514 "hidden": false,515 "can_act": false516 },517 {518 "id": 10,519 "count": 0,520 "hidden": false,521 "can_act": false522 },523 {524 "id": 7,525 "count": 0,526 "hidden": false,527 "can_act": false528 }529 ],530 "chunk_size": 20,531 "bookmarked": false,532 "topic_timer": null,533 "message_bus_last_id": 0,534 "participant_count": 2,535 "show_read_indicator": false,536 "thumbnails": null,537 "slow_mode_enabled_until": null,538 "can_vote": false,539 "vote_count": 0,540 "user_voted": false,541 "discourse_zendesk_plugin_zendesk_id": null,542 "discourse_zendesk_plugin_zendesk_url": "https://your-url.zendesk.com/agent/tickets/",543 "details": {544 "can_edit": false,545 "notification_level": 1,546 "participants": [547 {548 "id": 45640,549 "username": "Hadi_Nayebi",550 "name": "Hadi Nayebi",551 "avatar_template": "/user_avatar/discuss.pytorch.org/hadi_nayebi/{size}/38617_2.png",552 "post_count": 2,553 "primary_group_name": null,554 "flair_name": null,555 "flair_url": null,556 "flair_color": null,557 "flair_bg_color": null,558 "flair_group_id": null,559 "trust_level": 1560 },561 {562 "id": 3534,563 "username": "ptrblck",564 "name": "",565 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",566 "post_count": 1,567 "primary_group_name": null,568 "flair_name": null,569 "flair_url": null,570 "flair_color": null,571 "flair_bg_color": null,572 "flair_group_id": null,573 "admin": true,574 "moderator": true,575 "trust_level": 2576 }577 ],578 "created_by": {579 "id": 45640,580 "username": "Hadi_Nayebi",581 "name": "Hadi Nayebi",582 "avatar_template": "/user_avatar/discuss.pytorch.org/hadi_nayebi/{size}/38617_2.png"583 },584 "last_poster": {585 "id": 45640,586 "username": "Hadi_Nayebi",587 "name": "Hadi Nayebi",588 "avatar_template": "/user_avatar/discuss.pytorch.org/hadi_nayebi/{size}/38617_2.png"589 }590 },591 "bookmarks": []592 },593 {594 "post_stream": {595 "posts": [596 {597 "id": 287070,598 "name": "ADONAI TZEVAOT",599 "username": "ADONAI_TZEVAOT",600 "avatar_template": "/user_avatar/discuss.pytorch.org/adonai_tzevaot/{size}/31089_2.png",601 "created_at": "2021-05-30T14:44:51.043Z",602 "cooked": "<p>I have an input x of dimension 1x 2, batch size = 128, … Hence input gets passed in batches as 128x2. I have 2 parameters L_p(dimension = 1 x 1) and R_p (dimension = 2 x 2). The operation <code>x @ R_p</code> works but the operation <code>L_p @ x</code> is throwing error that matrix dimensions are not satisfied since x is passed as 128 x 2. But actually x is 1x2 , so how do I make it work? Please help!</p>",603 "post_number": 1,604 "post_type": 1,605 "posts_count": 2,606 "updated_at": "2021-05-30T14:44:51.043Z",607 "reply_count": 0,608 "reply_to_post_number": null,609 "quote_count": 0,610 "incoming_link_count": 10,611 "reads": 8,612 "readers_count": 7,613 "score": 51.6,614 "yours": false,615 "topic_id": 122784,616 "topic_slug": "left-multiplying-using-parameter-not-possible",617 "display_username": "ADONAI TZEVAOT",618 "primary_group_name": null,619 "flair_name": null,620 "flair_url": null,621 "flair_bg_color": null,622 "flair_color": null,623 "flair_group_id": null,624 "badges_granted": [],625 "version": 1,626 "can_edit": false,627 "can_delete": false,628 "can_recover": false,629 "can_see_hidden_post": false,630 "can_wiki": false,631 "read": true,632 "user_title": null,633 "bookmarked": false,634 "actions_summary": [],635 "moderator": false,636 "admin": false,637 "staff": false,638 "user_id": 38974,639 "hidden": false,640 "trust_level": 1,641 "deleted_at": null,642 "user_deleted": false,643 "edit_reason": null,644 "can_view_edit_history": true,645 "wiki": false,646 "post_url": "/t/left-multiplying-using-parameter-not-possible/122784/1",647 "can_accept_answer": false,648 "can_unaccept_answer": false,649 "accepted_answer": false,650 "topic_accepted_answer": null,651 "can_vote": false652 },653 {654 "id": 287117,655 "name": "",656 "username": "ptrblck",657 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",658 "created_at": "2021-05-31T00:12:46.739Z",659 "cooked": "<p><code>x @ R_p</code> will apply a matrix multiplication with the shapes:</p>\n<pre><code class=\"lang-python\">[128, 2] @ [2, 2] = [128, 2]\n</code></pre>\n<p>which are the expected shapes.<br>\nHowever, <code>L_p @ x</code> tries to execute:</p>\n<pre><code class=\"lang-python\">[1, 1] @ [128, 2]\n</code></pre>\n<p>which is invalid for a matmul.<br>\nI’m not sure what the expected output shape is, but assuming that you would like to broadcast <code>L_p</code> in the batch dimension, you could use:</p>\n<pre><code class=\"lang-python\">out = L_p.expand(x.size(0), 1, -1) @ x.unsqueeze(1)\n</code></pre>\n<p>which would create an output of <code>[128, 1, 2]</code>.</p>",660 "post_number": 2,661 "post_type": 1,662 "posts_count": 2,663 "updated_at": "2021-05-31T00:12:46.739Z",664 "reply_count": 0,665 "reply_to_post_number": null,666 "quote_count": 0,667 "incoming_link_count": 0,668 "reads": 7,669 "readers_count": 6,670 "score": 16.4,671 "yours": false,672 "topic_id": 122784,673 "topic_slug": "left-multiplying-using-parameter-not-possible",674 "display_username": "",675 "primary_group_name": null,676 "flair_name": null,677 "flair_url": null,678 "flair_bg_color": null,679 "flair_color": null,680 "flair_group_id": null,681 "badges_granted": [],682 "version": 1,683 "can_edit": false,684 "can_delete": false,685 "can_recover": false,686 "can_see_hidden_post": false,687 "can_wiki": false,688 "read": true,689 "user_title": "",690 "bookmarked": false,691 "actions_summary": [692 {693 "id": 2,694 "count": 1695 }696 ],697 "moderator": true,698 "admin": true,699 "staff": true,700 "user_id": 3534,701 "hidden": false,702 "trust_level": 2,703 "deleted_at": null,704 "user_deleted": false,705 "edit_reason": null,706 "can_view_edit_history": true,707 "wiki": false,708 "post_url": "/t/left-multiplying-using-parameter-not-possible/122784/2",709 "can_accept_answer": false,710 "can_unaccept_answer": false,711 "accepted_answer": false,712 "topic_accepted_answer": null713 }714 ],715 "stream": [716 287070,717 287117718 ]719 },720 "timeline_lookup": [721 [722 1,723 1609724 ]725 ],726 "suggested_topics": [727 {728 "fancy_title": "Question about default allowed globals",729 "id": 219716,730 "title": "Question about default allowed globals",731 "slug": "question-about-default-allowed-globals",732 "posts_count": 1,733 "reply_count": 0,734 "highest_post_number": 1,735 "image_url": null,736 "created_at": "2025-05-03T16:51:58.037Z",737 "last_posted_at": "2025-05-03T16:51:58.085Z",738 "bumped": true,739 "bumped_at": "2025-05-03T16:53:22.180Z",740 "archetype": "regular",741 "unseen": false,742 "pinned": false,743 "unpinned": null,744 "visible": true,745 "closed": false,746 "archived": false,747 "bookmarked": null,748 "liked": null,749 "tags_descriptions": {},750 "like_count": 0,751 "views": 182,752 "category_id": 1,753 "featured_link": null,754 "has_accepted_answer": false,755 "posters": [756 {757 "extras": "latest single",758 "description": "Original Poster, Most Recent Poster",759 "user": {760 "id": 22299,761 "username": "cwindolf",762 "name": "Charlie",763 "avatar_template": "/letter_avatar_proxy/v4/letter/c/ecd19e/{size}.png",764 "trust_level": 1765 }766 }767 ]768 },769 {770 "fancy_title": "Help: DDP hangs when using multi-threaded CUDA stream parallel loss computation",771 "id": 220193,772 "title": "Help: DDP hangs when using multi-threaded CUDA stream parallel loss computation",773 "slug": "help-ddp-hangs-when-using-multi-threaded-cuda-stream-parallel-loss-computation",774 "posts_count": 1,775 "reply_count": 0,776 "highest_post_number": 1,777 "image_url": null,778 "created_at": "2025-05-20T13:18:07.311Z",779 "last_posted_at": "2025-05-20T13:18:07.342Z",780 "bumped": true,781 "bumped_at": "2025-05-20T13:18:07.342Z",782 "archetype": "regular",783 "unseen": false,784 "pinned": false,785 "unpinned": null,786 "visible": true,787 "closed": false,788 "archived": false,789 "bookmarked": null,790 "liked": null,791 "tags_descriptions": {},792 "like_count": 0,793 "views": 37,794 "category_id": 1,795 "featured_link": null,796 "has_accepted_answer": false,797 "posters": [798 {799 "extras": "latest single",800 "description": "Original Poster, Most Recent Poster",801 "user": {802 "id": 84181,803 "username": "hantianyou",804 "name": "",805 "avatar_template": "/user_avatar/discuss.pytorch.org/hantianyou/{size}/77087_2.png",806 "trust_level": 1807 }808 }809 ]810 },811 {812 "fancy_title": "No errors before saving, but an error during testing",813 "id": 212388,814 "title": "No errors before saving, but an error during testing",815 "slug": "no-errors-before-saving-but-an-error-during-testing",816 "posts_count": 8,817 "reply_count": 2,818 "highest_post_number": 8,819 "image_url": null,820 "created_at": "2024-11-01T03:04:03.812Z",821 "last_posted_at": "2024-11-10T14:54:52.243Z",822 "bumped": true,823 "bumped_at": "2024-11-10T14:54:52.243Z",824 "archetype": "regular",825 "unseen": false,826 "pinned": false,827 "unpinned": null,828 "visible": true,829 "closed": false,830 "archived": false,831 "bookmarked": null,832 "liked": null,833 "tags_descriptions": {},834 "like_count": 0,835 "views": 60,836 "category_id": 1,837 "featured_link": null,838 "has_accepted_answer": true,839 "posters": [840 {841 "extras": "latest",842 "description": "Original Poster, Most Recent Poster, Accepted Answer",843 "user": {844 "id": 80199,845 "username": "Mathews_Vinoy",846 "name": "Mathews Vinoy",847 "avatar_template": "/user_avatar/discuss.pytorch.org/mathews_vinoy/{size}/73319_2.png",848 "trust_level": 1849 }850 },851 {852 "extras": null,853 "description": "Frequent Poster",854 "user": {855 "id": 3534,856 "username": "ptrblck",857 "name": "",858 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",859 "admin": true,860 "moderator": true,861 "trust_level": 2862 }863 }864 ]865 },866 {867 "fancy_title": "Torch svd grad show all zero, when only use Vt[-1]",868 "id": 215307,869 "title": "Torch svd grad show all zero, when only use Vt[-1]",870 "slug": "torch-svd-grad-show-all-zero-when-only-use-vt-1",871 "posts_count": 2,872 "reply_count": 0,873 "highest_post_number": 2,874 "image_url": null,875 "created_at": "2025-01-13T04:25:23.822Z",876 "last_posted_at": "2025-01-14T19:59:03.630Z",877 "bumped": true,878 "bumped_at": "2025-01-14T19:59:03.630Z",879 "archetype": "regular",880 "unseen": false,881 "pinned": false,882 "unpinned": null,883 "visible": true,884 "closed": false,885 "archived": false,886 "bookmarked": null,887 "liked": null,888 "tags_descriptions": {},889 "like_count": 0,890 "views": 61,891 "category_id": 1,892 "featured_link": null,893 "has_accepted_answer": false,894 "posters": [895 {896 "extras": null,897 "description": "Original Poster",898 "user": {899 "id": 82045,900 "username": "yangtaodummt",901 "name": "Yangtaodummt",902 "avatar_template": "/user_avatar/discuss.pytorch.org/yangtaodummt/{size}/75071_2.png",903 "trust_level": 0904 }905 },906 {907 "extras": "latest",908 "description": "Most Recent Poster",909 "user": {910 "id": 18088,911 "username": "KFrank",912 "name": "K. Frank",913 "avatar_template": "/letter_avatar_proxy/v4/letter/k/ecb155/{size}.png",914 "trust_level": 2915 }916 }917 ]918 },919 {920 "fancy_title": "Help with pytorch running on Pop_OS!",921 "id": 220408,922 "title": "Help with pytorch running on Pop_OS!",923 "slug": "help-with-pytorch-running-on-pop-os",924 "posts_count": 3,925 "reply_count": 1,926 "highest_post_number": 3,927 "image_url": null,928 "created_at": "2025-05-28T13:56:18.702Z",929 "last_posted_at": "2025-05-28T17:26:47.663Z",930 "bumped": true,931 "bumped_at": "2025-05-28T17:26:47.663Z",932 "archetype": "regular",933 "unseen": false,934 "pinned": false,935 "unpinned": null,936 "visible": true,937 "closed": false,938 "archived": false,939 "bookmarked": null,940 "liked": null,941 "tags_descriptions": {},942 "like_count": 0,943 "views": 79,944 "category_id": 1,945 "featured_link": null,946 "has_accepted_answer": false,947 "posters": [948 {949 "extras": "latest",950 "description": "Original Poster, Most Recent Poster",951 "user": {952 "id": 84490,953 "username": "Alikebrahim",954 "name": "Ali K. Ebrahim",955 "avatar_template": "/user_avatar/discuss.pytorch.org/alikebrahim/{size}/77199_2.png",956 "trust_level": 0957 }958 },959 {960 "extras": null,961 "description": "Frequent Poster",962 "user": {963 "id": 3534,964 "username": "ptrblck",965 "name": "",966 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",967 "admin": true,968 "moderator": true,969 "trust_level": 2970 }971 }972 ]973 }974 ],975 "tags_descriptions": {},976 "fancy_title": "Left multiplying using parameter not possible?",977 "id": 122784,978 "title": "Left multiplying using parameter not possible?",979 "posts_count": 2,980 "created_at": "2021-05-30T14:44:50.988Z",981 "views": 361,982 "reply_count": 0,983 "like_count": 1,984 "last_posted_at": "2021-05-31T00:12:46.739Z",985 "visible": true,986 "closed": false,987 "archived": false,988 "has_summary": false,989 "archetype": "regular",990 "slug": "left-multiplying-using-parameter-not-possible",991 "category_id": 1,992 "word_count": 162,993 "deleted_at": null,994 "user_id": 38974,995 "featured_link": null,996 "pinned_globally": false,997 "pinned_at": null,998 "pinned_until": null,999 "image_url": null,1000 "slow_mode_seconds": 0,1001 "draft": null,1002 "draft_key": "topic_122784",1003 "draft_sequence": null,1004 "unpinned": null,1005 "pinned": false,1006 "current_post_number": 1,1007 "highest_post_number": 2,1008 "deleted_by": null,1009 "actions_summary": [1010 {1011 "id": 4,1012 "count": 0,1013 "hidden": false,1014 "can_act": false1015 },1016 {1017 "id": 8,1018 "count": 0,1019 "hidden": false,1020 "can_act": false1021 },1022 {1023 "id": 10,1024 "count": 0,1025 "hidden": false,1026 "can_act": false1027 },1028 {1029 "id": 7,1030 "count": 0,1031 "hidden": false,1032 "can_act": false1033 }1034 ],1035 "chunk_size": 20,1036 "bookmarked": false,1037 "topic_timer": null,1038 "message_bus_last_id": 0,1039 "participant_count": 2,1040 "show_read_indicator": false,1041 "thumbnails": null,1042 "slow_mode_enabled_until": null,1043 "can_vote": false,1044 "vote_count": 0,1045 "user_voted": false,1046 "discourse_zendesk_plugin_zendesk_id": null,1047 "discourse_zendesk_plugin_zendesk_url": "https://your-url.zendesk.com/agent/tickets/",1048 "details": {1049 "can_edit": false,1050 "notification_level": 1,1051 "participants": [1052 {1053 "id": 3534,1054 "username": "ptrblck",1055 "name": "",1056 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",1057 "post_count": 1,1058 "primary_group_name": null,1059 "flair_name": null,1060 "flair_url": null,1061 "flair_color": null,1062 "flair_bg_color": null,1063 "flair_group_id": null,1064 "admin": true,1065 "moderator": true,1066 "trust_level": 21067 },1068 {1069 "id": 38974,1070 "username": "ADONAI_TZEVAOT",1071 "name": "ADONAI TZEVAOT",1072 "avatar_template": "/user_avatar/discuss.pytorch.org/adonai_tzevaot/{size}/31089_2.png",1073 "post_count": 1,1074 "primary_group_name": null,1075 "flair_name": null,1076 "flair_url": null,1077 "flair_color": null,1078 "flair_bg_color": null,1079 "flair_group_id": null,1080 "trust_level": 11081 }1082 ],1083 "created_by": {1084 "id": 38974,1085 "username": "ADONAI_TZEVAOT",1086 "name": "ADONAI TZEVAOT",1087 "avatar_template": "/user_avatar/discuss.pytorch.org/adonai_tzevaot/{size}/31089_2.png"1088 },1089 "last_poster": {1090 "id": 3534,1091 "username": "ptrblck",1092 "name": "",1093 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png"1094 }1095 },1096 "bookmarks": []1097 },1098 {1099 "post_stream": {1100 "posts": [1101 {1102 "id": 287066,1103 "name": "Santhosh Dhaipule Chandrakanth",1104 "username": "Santhoshnumberone",1105 "avatar_template": "/user_avatar/discuss.pytorch.org/santhoshnumberone/{size}/7029_2.png",1106 "created_at": "2021-05-30T11:29:36.168Z",1107 "cooked": "<p>Running this file <a href=\"https://github.com/timesler/facenet-pytorch/blob/master/examples/finetune.ipynb\" class=\"inline-onebox\" rel=\"noopener nofollow ugc\">facenet-pytorch/finetune.ipynb at master · timesler/facenet-pytorch · GitHub</a></p>\n<p>I want to train only the final layer of</p>\n<pre><code class=\"lang-auto\">resnet = InceptionResnetV1(\n classify=True,\n pretrained='vggface2',\n num_classes=len(dataset.class_to_idx)\n).to(device)\n</code></pre>\n<p>I did freeze all layers but unfroze last layer</p>\n<pre><code class=\"lang-auto\">#Freeze all layers\nfor param in resnet.parameters():\n param.requires_grad = False\n\n#Unfreeze the last layer\nfor param in resnet.logits.parameters():\n param.requires_grad = True \n</code></pre>\n<p>I tried running the code I get this error</p>\n<pre><code class=\"lang-auto\">Initial\n----------\nValid | 2/2 | loss: 6.2345 | fps: 182.8512 | acc: 0.0000 \n\nEpoch 1/100\n----------\n---------------------------------------------------------------------------\nRuntimeError Traceback (most recent call last)\n<ipython-input-45-49d240c83dad> in <module>()\n 19 resnet, loss_fn, train_loader, optimizer, scheduler,\n 20 batch_metrics=metrics, show_running=True, device=device,\n---> 21 writer=writer\n 22 )\n 23 \n\n2 frames\n/usr/local/lib/python3.7/dist-packages/torch/autograd/__init__.py in backward(tensors, grad_tensors, retain_graph, create_graph, grad_variables, inputs)\n 145 Variable._execution_engine.run_backward(\n 146 tensors, grad_tensors_, retain_graph, create_graph, inputs,\n--> 147 allow_unreachable=True, accumulate_grad=True) # allow_unreachable flag\n 148 \n 149 \n\nRuntimeError: element 0 of tensors does not require grad and does not have a grad_fn\n</code></pre>",1108 "post_number": 1,1109 "post_type": 1,1110 "posts_count": 2,1111 "updated_at": "2021-05-30T11:29:36.168Z",1112 "reply_count": 0,1113 "reply_to_post_number": null,1114 "quote_count": 0,1115 "incoming_link_count": 83,1116 "reads": 3,1117 "readers_count": 2,1118 "score": 415.6,1119 "yours": false,1120 "topic_id": 122781,1121 "topic_slug": "runtimeerror-on-running-transfer-learning",1122 "display_username": "Santhosh Dhaipule Chandrakanth",1123 "primary_group_name": null,1124 "flair_name": null,1125 "flair_url": null,1126 "flair_bg_color": null,1127 "flair_color": null,1128 "flair_group_id": null,1129 "badges_granted": [],1130 "version": 1,1131 "can_edit": false,1132 "can_delete": false,1133 "can_recover": false,1134 "can_see_hidden_post": false,1135 "can_wiki": false,1136 "link_counts": [1137 {1138 "url": "https://github.com/timesler/facenet-pytorch/blob/master/examples/finetune.ipynb",1139 "internal": false,1140 "reflection": false,1141 "title": "facenet-pytorch/finetune.ipynb at master · timesler/facenet-pytorch · GitHub",1142 "clicks": 21143 }1144 ],1145 "read": true,1146 "user_title": "",1147 "bookmarked": false,1148 "actions_summary": [],1149 "moderator": false,1150 "admin": false,1151 "staff": false,1152 "user_id": 11884,1153 "hidden": false,1154 "trust_level": 2,1155 "deleted_at": null,1156 "user_deleted": false,1157 "edit_reason": null,1158 "can_view_edit_history": true,1159 "wiki": false,1160 "post_url": "/t/runtimeerror-on-running-transfer-learning/122781/1",1161 "can_accept_answer": false,1162 "can_unaccept_answer": false,1163 "accepted_answer": false,1164 "topic_accepted_answer": null,1165 "can_vote": false1166 },1167 {1168 "id": 287116,1169 "name": "",1170 "username": "ptrblck",1171 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",1172 "created_at": "2021-05-30T23:31:10.278Z",1173 "cooked": "<p>The <code>logits</code> layer will only be used, if <code>self.classify</code> is set to <code>True</code> as seen <a href=\"https://github.com/timesler/facenet-pytorch/blob/54c869c51e0e3e12f7f92f551cdd2ecd164e2443/models/inception_resnet_v1.py#L298-L299\">here</a>, so I guess this might not be the case.<br>\nIf so, this layer won’t be used while all other parameters are frozen, which would yield this error.</p>",1174 "post_number": 2,1175 "post_type": 1,1176 "posts_count": 2,1177 "updated_at": "2021-05-30T23:31:10.278Z",1178 "reply_count": 0,1179 "reply_to_post_number": null,1180 "quote_count": 0,1181 "incoming_link_count": 3,1182 "reads": 3,1183 "readers_count": 2,1184 "score": 15.6,1185 "yours": false,1186 "topic_id": 122781,1187 "topic_slug": "runtimeerror-on-running-transfer-learning",1188 "display_username": "",1189 "primary_group_name": null,1190 "flair_name": null,1191 "flair_url": null,1192 "flair_bg_color": null,1193 "flair_color": null,1194 "flair_group_id": null,1195 "badges_granted": [],1196 "version": 1,1197 "can_edit": false,1198 "can_delete": false,1199 "can_recover": false,1200 "can_see_hidden_post": false,