Anurag1734/cuda-error-resolution-analysis
07
1[2 {3 "post_stream": {4 "posts": [5 {6 "id": 358563,7 "name": "Mahdi Amrollahi",8 "username": "Mahdi_Amrollahi",9 "avatar_template": "/user_avatar/discuss.pytorch.org/mahdi_amrollahi/{size}/62481_2.png",10 "created_at": "2022-07-27T16:21:24.242Z",11 "cooked": "<p>Here in this document we have a vector v which is multiplied by jacobian matrix.<br>\nWhat is v vector?<br>\n<a href=\"https://pytorch.org/tutorials/beginner/basics/autogradqs_tutorial.html\" rel=\"noopener nofollow ugc\">Pytorch doc</a></p>",12 "post_number": 1,13 "post_type": 1,14 "posts_count": 1,15 "updated_at": "2022-07-27T16:21:24.242Z",16 "reply_count": 0,17 "reply_to_post_number": null,18 "quote_count": 0,19 "incoming_link_count": 5,20 "reads": 6,21 "readers_count": 5,22 "score": 26.2,23 "yours": false,24 "topic_id": 157632,25 "topic_slug": "what-is-v-vector-in-autograd-and-jacobian-matrix",26 "display_username": "Mahdi Amrollahi",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://pytorch.org/tutorials/beginner/basics/autogradqs_tutorial.html",43 "internal": false,44 "reflection": false,45 "title": "Automatic Differentiation with torch.autograd — PyTorch Tutorials 1.12.0+cu102 documentation",46 "clicks": 347 }48 ],49 "read": true,50 "user_title": "",51 "bookmarked": false,52 "actions_summary": [],53 "moderator": false,54 "admin": false,55 "staff": false,56 "user_id": 49415,57 "hidden": false,58 "trust_level": 1,59 "deleted_at": null,60 "user_deleted": false,61 "edit_reason": null,62 "can_view_edit_history": true,63 "wiki": false,64 "post_url": "/t/what-is-v-vector-in-autograd-and-jacobian-matrix/157632/1",65 "can_accept_answer": false,66 "can_unaccept_answer": false,67 "accepted_answer": false,68 "topic_accepted_answer": null,69 "can_vote": false70 }71 ],72 "stream": [73 35856374 ]75 },76 "timeline_lookup": [77 [78 1,79 118680 ]81 ],82 "suggested_topics": [83 {84 "fancy_title": "Gradient computation with PyTorch autograd with 1th and 2th order derivatives does not work",85 "id": 216580,86 "title": "Gradient computation with PyTorch autograd with 1th and 2th order derivatives does not work",87 "slug": "gradient-computation-with-pytorch-autograd-with-1th-and-2th-order-derivatives-does-not-work",88 "posts_count": 2,89 "reply_count": 0,90 "highest_post_number": 2,91 "image_url": null,92 "created_at": "2025-02-12T13:12:01.888Z",93 "last_posted_at": "2025-02-15T01:35:32.749Z",94 "bumped": true,95 "bumped_at": "2025-02-15T01:35:32.749Z",96 "archetype": "regular",97 "unseen": false,98 "pinned": false,99 "unpinned": null,100 "visible": true,101 "closed": false,102 "archived": false,103 "bookmarked": null,104 "liked": null,105 "tags_descriptions": {},106 "like_count": 1,107 "views": 67,108 "category_id": 7,109 "featured_link": null,110 "has_accepted_answer": true,111 "posters": [112 {113 "extras": null,114 "description": "Original Poster",115 "user": {116 "id": 82645,117 "username": "woutr",118 "name": "",119 "avatar_template": "/letter_avatar_proxy/v4/letter/w/6f9a4e/{size}.png",120 "trust_level": 1121 }122 },123 {124 "extras": "latest",125 "description": "Most Recent Poster, Accepted Answer",126 "user": {127 "id": 18088,128 "username": "KFrank",129 "name": "K. Frank",130 "avatar_template": "/letter_avatar_proxy/v4/letter/k/ecb155/{size}.png",131 "trust_level": 2132 }133 }134 ]135 },136 {137 "fancy_title": "`compile` a function with `autograd`",138 "id": 217385,139 "title": "`compile` a function with `autograd`",140 "slug": "compile-a-function-with-autograd",141 "posts_count": 2,142 "reply_count": 0,143 "highest_post_number": 2,144 "image_url": null,145 "created_at": "2025-03-03T15:38:12.611Z",146 "last_posted_at": "2025-03-05T22:47:12.081Z",147 "bumped": true,148 "bumped_at": "2025-03-05T22:47:12.081Z",149 "archetype": "regular",150 "unseen": false,151 "pinned": false,152 "unpinned": null,153 "visible": true,154 "closed": false,155 "archived": false,156 "bookmarked": null,157 "liked": null,158 "tags_descriptions": {},159 "like_count": 0,160 "views": 68,161 "category_id": 7,162 "featured_link": null,163 "has_accepted_answer": false,164 "posters": [165 {166 "extras": null,167 "description": "Original Poster",168 "user": {169 "id": 83042,170 "username": "madirbah",171 "name": "madirbah",172 "avatar_template": "/user_avatar/discuss.pytorch.org/madirbah/{size}/74610_2.png",173 "trust_level": 1174 }175 },176 {177 "extras": "latest",178 "description": "Most Recent Poster",179 "user": {180 "id": 41396,181 "username": "soulitzer",182 "name": "",183 "avatar_template": "/letter_avatar_proxy/v4/letter/s/839c29/{size}.png",184 "trust_level": 2185 }186 }187 ]188 },189 {190 "fancy_title": "Loss.backward() called after torch.nograd()",191 "id": 215480,192 "title": "Loss.backward() called after torch.nograd()",193 "slug": "loss-backward-called-after-torch-nograd",194 "posts_count": 5,195 "reply_count": 3,196 "highest_post_number": 5,197 "image_url": null,198 "created_at": "2025-01-16T15:42:16.612Z",199 "last_posted_at": "2025-01-19T16:44:10.276Z",200 "bumped": true,201 "bumped_at": "2025-01-19T16:44:10.276Z",202 "archetype": "regular",203 "unseen": false,204 "pinned": false,205 "unpinned": null,206 "visible": true,207 "closed": false,208 "archived": false,209 "bookmarked": null,210 "liked": null,211 "tags_descriptions": {},212 "like_count": 0,213 "views": 81,214 "category_id": 7,215 "featured_link": null,216 "has_accepted_answer": true,217 "posters": [218 {219 "extras": "latest",220 "description": "Original Poster, Most Recent Poster",221 "user": {222 "id": 82127,223 "username": "hanxix",224 "name": "",225 "avatar_template": "/letter_avatar_proxy/v4/letter/h/8797f3/{size}.png",226 "trust_level": 0227 }228 },229 {230 "extras": null,231 "description": "Frequent Poster, Accepted Answer",232 "user": {233 "id": 3534,234 "username": "ptrblck",235 "name": "",236 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",237 "admin": true,238 "moderator": true,239 "trust_level": 2240 }241 }242 ]243 },244 {245 "fancy_title": "How Does model.eval() Affect Gradient Descent in PyTorch and How to Handle Frequent Evaluations?",246 "id": 214671,247 "title": "How Does model.eval() Affect Gradient Descent in PyTorch and How to Handle Frequent Evaluations?",248 "slug": "how-does-model-eval-affect-gradient-descent-in-pytorch-and-how-to-handle-frequent-evaluations",249 "posts_count": 4,250 "reply_count": 2,251 "highest_post_number": 4,252 "image_url": null,253 "created_at": "2024-12-26T16:52:24.941Z",254 "last_posted_at": "2024-12-31T03:02:16.027Z",255 "bumped": true,256 "bumped_at": "2024-12-31T03:02:16.027Z",257 "archetype": "regular",258 "unseen": false,259 "pinned": false,260 "unpinned": null,261 "visible": true,262 "closed": false,263 "archived": false,264 "bookmarked": null,265 "liked": null,266 "tags_descriptions": {},267 "like_count": 0,268 "views": 55,269 "category_id": 7,270 "featured_link": null,271 "has_accepted_answer": false,272 "posters": [273 {274 "extras": null,275 "description": "Original Poster",276 "user": {277 "id": 81154,278 "username": "5o1",279 "name": "",280 "avatar_template": "/user_avatar/discuss.pytorch.org/5o1/{size}/74221_2.png",281 "trust_level": 1282 }283 },284 {285 "extras": "latest",286 "description": "Most Recent Poster",287 "user": {288 "id": 41396,289 "username": "soulitzer",290 "name": "",291 "avatar_template": "/letter_avatar_proxy/v4/letter/s/839c29/{size}.png",292 "trust_level": 2293 }294 }295 ]296 },297 {298 "fancy_title": "Making autograd saved tensors hooks specific to certain arguments",299 "id": 218433,300 "title": "Making autograd saved tensors hooks specific to certain arguments",301 "slug": "making-autograd-saved-tensors-hooks-specific-to-certain-arguments",302 "posts_count": 9,303 "reply_count": 6,304 "highest_post_number": 9,305 "image_url": null,306 "created_at": "2025-03-31T10:12:52.310Z",307 "last_posted_at": "2025-09-17T07:44:20.765Z",308 "bumped": true,309 "bumped_at": "2025-09-17T07:44:20.765Z",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": 192,322 "category_id": 7,323 "featured_link": null,324 "has_accepted_answer": true,325 "posters": [326 {327 "extras": "latest",328 "description": "Original Poster, Most Recent Poster, Accepted Answer",329 "user": {330 "id": 81725,331 "username": "mseeger",332 "name": null,333 "avatar_template": "/letter_avatar_proxy/v4/letter/m/6bbea6/{size}.png",334 "trust_level": 1335 }336 },337 {338 "extras": null,339 "description": "Frequent Poster",340 "user": {341 "id": 41396,342 "username": "soulitzer",343 "name": "",344 "avatar_template": "/letter_avatar_proxy/v4/letter/s/839c29/{size}.png",345 "trust_level": 2346 }347 }348 ]349 }350 ],351 "tags_descriptions": {},352 "fancy_title": "What is v vector in autograd and Jacobian matrix?",353 "id": 157632,354 "title": "What is v vector in autograd and Jacobian matrix?",355 "posts_count": 1,356 "created_at": "2022-07-27T16:21:24.182Z",357 "views": 365,358 "reply_count": 0,359 "like_count": 0,360 "last_posted_at": "2022-07-27T16:21:24.242Z",361 "visible": true,362 "closed": false,363 "archived": false,364 "has_summary": false,365 "archetype": "regular",366 "slug": "what-is-v-vector-in-autograd-and-jacobian-matrix",367 "category_id": 7,368 "word_count": 29,369 "deleted_at": null,370 "user_id": 49415,371 "featured_link": null,372 "pinned_globally": false,373 "pinned_at": null,374 "pinned_until": null,375 "image_url": null,376 "slow_mode_seconds": 0,377 "draft": null,378 "draft_key": "topic_157632",379 "draft_sequence": null,380 "unpinned": null,381 "pinned": false,382 "current_post_number": 1,383 "highest_post_number": 1,384 "deleted_by": null,385 "actions_summary": [386 {387 "id": 4,388 "count": 0,389 "hidden": false,390 "can_act": false391 },392 {393 "id": 8,394 "count": 0,395 "hidden": false,396 "can_act": false397 },398 {399 "id": 10,400 "count": 0,401 "hidden": false,402 "can_act": false403 },404 {405 "id": 7,406 "count": 0,407 "hidden": false,408 "can_act": false409 }410 ],411 "chunk_size": 20,412 "bookmarked": false,413 "topic_timer": null,414 "message_bus_last_id": 0,415 "participant_count": 1,416 "show_read_indicator": false,417 "thumbnails": null,418 "slow_mode_enabled_until": null,419 "can_vote": false,420 "vote_count": 0,421 "user_voted": false,422 "discourse_zendesk_plugin_zendesk_id": null,423 "discourse_zendesk_plugin_zendesk_url": "https://your-url.zendesk.com/agent/tickets/",424 "details": {425 "can_edit": false,426 "notification_level": 1,427 "participants": [428 {429 "id": 49415,430 "username": "Mahdi_Amrollahi",431 "name": "Mahdi Amrollahi",432 "avatar_template": "/user_avatar/discuss.pytorch.org/mahdi_amrollahi/{size}/62481_2.png",433 "post_count": 1,434 "primary_group_name": null,435 "flair_name": null,436 "flair_url": null,437 "flair_color": null,438 "flair_bg_color": null,439 "flair_group_id": null,440 "trust_level": 1441 }442 ],443 "created_by": {444 "id": 49415,445 "username": "Mahdi_Amrollahi",446 "name": "Mahdi Amrollahi",447 "avatar_template": "/user_avatar/discuss.pytorch.org/mahdi_amrollahi/{size}/62481_2.png"448 },449 "last_poster": {450 "id": 49415,451 "username": "Mahdi_Amrollahi",452 "name": "Mahdi Amrollahi",453 "avatar_template": "/user_avatar/discuss.pytorch.org/mahdi_amrollahi/{size}/62481_2.png"454 },455 "links": [456 {457 "url": "https://pytorch.org/tutorials/beginner/basics/autogradqs_tutorial.html",458 "title": "Automatic Differentiation with torch.autograd — PyTorch Tutorials 1.12.0+cu102 documentation",459 "internal": false,460 "attachment": false,461 "reflection": false,462 "clicks": 3,463 "user_id": 49415,464 "domain": "pytorch.org",465 "root_domain": "pytorch.org"466 }467 ]468 },469 "bookmarks": []470 },471 {472 "post_stream": {473 "posts": [474 {475 "id": 358534,476 "name": "Yuancheng Xu",477 "username": "Yuancheng_Xu",478 "avatar_template": "/user_avatar/discuss.pytorch.org/yuancheng_xu/{size}/45613_2.png",479 "created_at": "2022-07-27T13:28:42.977Z",480 "cooked": "<p>Given a neural network classifier with 10 classes (the final layer logits have dimension 10). Is there a way to compute the gradients of each of the logit w.r.t. the input in one backward propagation?</p>\n<p>The only way I know to do this is by setting a for-loop like this</p>\n<pre><code class=\"lang-auto\">input_grad = torch.zeros(10,input_size)\nlogit = NN(x)\nfor i in range(10):\n x.grad.zero_()\n logit[i].backward(retain_graph=True)\n input_grad[i] = x.grad\n</code></pre>\n<p>However, this for loop is much slower than one backprop. Another minor question: is it correct that the above for-loop has 10 times computational cost as a single backprop, if the network is large enough?</p>",481 "post_number": 1,482 "post_type": 1,483 "posts_count": 4,484 "updated_at": "2022-07-27T13:29:49.110Z",485 "reply_count": 0,486 "reply_to_post_number": null,487 "quote_count": 0,488 "incoming_link_count": 158,489 "reads": 5,490 "readers_count": 4,491 "score": 791.0,492 "yours": false,493 "topic_id": 157613,494 "topic_slug": "compute-the-gradients-of-all-logins-with-respect-to-input",495 "display_username": "Yuancheng Xu",496 "primary_group_name": null,497 "flair_name": null,498 "flair_url": null,499 "flair_bg_color": null,500 "flair_color": null,501 "flair_group_id": null,502 "badges_granted": [],503 "version": 1,504 "can_edit": false,505 "can_delete": false,506 "can_recover": false,507 "can_see_hidden_post": false,508 "can_wiki": false,509 "read": true,510 "user_title": null,511 "bookmarked": false,512 "actions_summary": [],513 "moderator": false,514 "admin": false,515 "staff": false,516 "user_id": 52233,517 "hidden": false,518 "trust_level": 1,519 "deleted_at": null,520 "user_deleted": false,521 "edit_reason": null,522 "can_view_edit_history": true,523 "wiki": false,524 "post_url": "/t/compute-the-gradients-of-all-logins-with-respect-to-input/157613/1",525 "can_accept_answer": false,526 "can_unaccept_answer": false,527 "accepted_answer": false,528 "topic_accepted_answer": null,529 "can_vote": false530 },531 {532 "id": 358536,533 "name": "",534 "username": "AlphaBetaGamma96",535 "avatar_template": "/letter_avatar_proxy/v4/letter/a/3da27b/{size}.png",536 "created_at": "2022-07-27T13:35:24.848Z",537 "cooked": "<p>Could you try the following code and see if you get the same values?</p>\n<pre><code class=\"lang-auto\">logit = NN(x)\ninput_grad, = torch.autograd.grad(logit, x, torch.ones_like(logit)) #note the comma after input_grad\n</code></pre>\n<p>Could try something like this to directly compare actually,</p>\n<pre><code class=\"lang-auto\">input_grad = torch.zeros(10,input_size)\nlogit = NN(x)\nfor i in range(10):\n x.grad.zero_()\n logit[i].backward(retain_graph=True)\n input_grad[i] = x.grad\n\nlogit_all = NN(x)\ninput_grad_all, = torch.autograd.grad(logit_all, x, torch.ones_like(logit_all))\n\nprint(torch.allclose(input_grad, input_grad_all))\n</code></pre>",538 "post_number": 2,539 "post_type": 1,540 "posts_count": 4,541 "updated_at": "2022-07-27T13:35:38.200Z",542 "reply_count": 1,543 "reply_to_post_number": null,544 "quote_count": 0,545 "incoming_link_count": 1,546 "reads": 5,547 "readers_count": 4,548 "score": 11.0,549 "yours": false,550 "topic_id": 157613,551 "topic_slug": "compute-the-gradients-of-all-logins-with-respect-to-input",552 "display_username": "",553 "primary_group_name": null,554 "flair_name": null,555 "flair_url": null,556 "flair_bg_color": null,557 "flair_color": null,558 "flair_group_id": null,559 "badges_granted": [],560 "version": 1,561 "can_edit": false,562 "can_delete": false,563 "can_recover": false,564 "can_see_hidden_post": false,565 "can_wiki": false,566 "read": true,567 "user_title": "",568 "bookmarked": false,569 "actions_summary": [],570 "moderator": false,571 "admin": false,572 "staff": false,573 "user_id": 34294,574 "hidden": false,575 "trust_level": 2,576 "deleted_at": null,577 "user_deleted": false,578 "edit_reason": null,579 "can_view_edit_history": true,580 "wiki": false,581 "post_url": "/t/compute-the-gradients-of-all-logins-with-respect-to-input/157613/2",582 "can_accept_answer": false,583 "can_unaccept_answer": false,584 "accepted_answer": false,585 "topic_accepted_answer": null586 },587 {588 "id": 358541,589 "name": "Yuancheng Xu",590 "username": "Yuancheng_Xu",591 "avatar_template": "/user_avatar/discuss.pytorch.org/yuancheng_xu/{size}/45613_2.png",592 "created_at": "2022-07-27T14:22:50.747Z",593 "cooked": "<p>Thank you for the reply!</p>\n<p>Let’s say x is of size 5 * 3 * 32 * 32 (5 is the batch size) and num_classes = 10.<br>\nI think in your example, input_grad_all has the same size of the input x (both of them are 5 * 3 * 32 * 32). However, I would like to compute the gradients of each logit w.r.t. the input. That is, I am expecting the gradient to be of size 10 * 5 * 3 * 32 * 32.</p>",594 "post_number": 3,595 "post_type": 1,596 "posts_count": 4,597 "updated_at": "2022-07-27T14:22:50.747Z",598 "reply_count": 1,599 "reply_to_post_number": 2,600 "quote_count": 0,601 "incoming_link_count": 1,602 "reads": 3,603 "readers_count": 2,604 "score": 10.6,605 "yours": false,606 "topic_id": 157613,607 "topic_slug": "compute-the-gradients-of-all-logins-with-respect-to-input",608 "display_username": "Yuancheng Xu",609 "primary_group_name": null,610 "flair_name": null,611 "flair_url": null,612 "flair_bg_color": null,613 "flair_color": null,614 "flair_group_id": null,615 "badges_granted": [],616 "version": 1,617 "can_edit": false,618 "can_delete": false,619 "can_recover": false,620 "can_see_hidden_post": false,621 "can_wiki": false,622 "read": true,623 "user_title": null,624 "reply_to_user": {625 "id": 34294,626 "username": "AlphaBetaGamma96",627 "name": "",628 "avatar_template": "/letter_avatar_proxy/v4/letter/a/3da27b/{size}.png"629 },630 "bookmarked": false,631 "actions_summary": [],632 "moderator": false,633 "admin": false,634 "staff": false,635 "user_id": 52233,636 "hidden": false,637 "trust_level": 1,638 "deleted_at": null,639 "user_deleted": false,640 "edit_reason": null,641 "can_view_edit_history": true,642 "wiki": false,643 "post_url": "/t/compute-the-gradients-of-all-logins-with-respect-to-input/157613/3",644 "can_accept_answer": false,645 "can_unaccept_answer": false,646 "accepted_answer": false,647 "topic_accepted_answer": null648 },649 {650 "id": 358562,651 "name": "",652 "username": "AlphaBetaGamma96",653 "avatar_template": "/letter_avatar_proxy/v4/letter/a/3da27b/{size}.png",654 "created_at": "2022-07-27T16:12:03.081Z",655 "cooked": "<p>Could you share a minimal reproducible example of <code>NN</code>? Just so there’s a complete example I can use to debug your problem.</p>",656 "post_number": 4,657 "post_type": 1,658 "posts_count": 4,659 "updated_at": "2022-07-27T16:12:03.081Z",660 "reply_count": 0,661 "reply_to_post_number": 3,662 "quote_count": 0,663 "incoming_link_count": 0,664 "reads": 3,665 "readers_count": 2,666 "score": 0.6,667 "yours": false,668 "topic_id": 157613,669 "topic_slug": "compute-the-gradients-of-all-logins-with-respect-to-input",670 "display_username": "",671 "primary_group_name": null,672 "flair_name": null,673 "flair_url": null,674 "flair_bg_color": null,675 "flair_color": null,676 "flair_group_id": null,677 "badges_granted": [],678 "version": 1,679 "can_edit": false,680 "can_delete": false,681 "can_recover": false,682 "can_see_hidden_post": false,683 "can_wiki": false,684 "read": true,685 "user_title": "",686 "reply_to_user": {687 "id": 52233,688 "username": "Yuancheng_Xu",689 "name": "Yuancheng Xu",690 "avatar_template": "/user_avatar/discuss.pytorch.org/yuancheng_xu/{size}/45613_2.png"691 },692 "bookmarked": false,693 "actions_summary": [],694 "moderator": false,695 "admin": false,696 "staff": false,697 "user_id": 34294,698 "hidden": false,699 "trust_level": 2,700 "deleted_at": null,701 "user_deleted": false,702 "edit_reason": null,703 "can_view_edit_history": true,704 "wiki": false,705 "post_url": "/t/compute-the-gradients-of-all-logins-with-respect-to-input/157613/4",706 "can_accept_answer": false,707 "can_unaccept_answer": false,708 "accepted_answer": false,709 "topic_accepted_answer": null710 }711 ],712 "stream": [713 358534,714 358536,715 358541,716 358562717 ]718 },719 "timeline_lookup": [720 [721 1,722 1186723 ]724 ],725 "suggested_topics": [726 {727 "fancy_title": "Get loss, gradient and hessian in one go",728 "id": 219601,729 "title": "Get loss, gradient and hessian in one go",730 "slug": "get-loss-gradient-and-hessian-in-one-go",731 "posts_count": 2,732 "reply_count": 0,733 "highest_post_number": 2,734 "image_url": null,735 "created_at": "2025-04-29T16:53:14.779Z",736 "last_posted_at": "2025-04-30T16:12:52.891Z",737 "bumped": true,738 "bumped_at": "2025-04-30T16:12:52.891Z",739 "archetype": "regular",740 "unseen": false,741 "pinned": false,742 "unpinned": null,743 "visible": true,744 "closed": false,745 "archived": false,746 "bookmarked": null,747 "liked": null,748 "tags_descriptions": {},749 "like_count": 0,750 "views": 109,751 "category_id": 7,752 "featured_link": null,753 "has_accepted_answer": true,754 "posters": [755 {756 "extras": null,757 "description": "Original Poster",758 "user": {759 "id": 75871,760 "username": "qq-me",761 "name": "Ivan Nikishev",762 "avatar_template": "/user_avatar/discuss.pytorch.org/qq-me/{size}/70055_2.png",763 "trust_level": 2764 }765 },766 {767 "extras": "latest",768 "description": "Most Recent Poster, Accepted Answer",769 "user": {770 "id": 18088,771 "username": "KFrank",772 "name": "K. Frank",773 "avatar_template": "/letter_avatar_proxy/v4/letter/k/ecb155/{size}.png",774 "trust_level": 2775 }776 }777 ]778 },779 {780 "fancy_title": "Free some saved tensors after partial backward",781 "id": 215486,782 "title": "Free some saved tensors after partial backward",783 "slug": "free-some-saved-tensors-after-partial-backward",784 "posts_count": 7,785 "reply_count": 5,786 "highest_post_number": 7,787 "image_url": "https://discuss.pytorch.org/uploads/default/original/3X/b/b/bbf6e39f3ea595f48225fb02741a0d7b481ac843.png",788 "created_at": "2025-01-16T17:45:37.159Z",789 "last_posted_at": "2025-02-14T16:47:34.353Z",790 "bumped": true,791 "bumped_at": "2025-02-14T16:47:34.353Z",792 "archetype": "regular",793 "unseen": false,794 "pinned": false,795 "unpinned": null,796 "visible": true,797 "closed": false,798 "archived": false,799 "bookmarked": null,800 "liked": null,801 "tags_descriptions": {},802 "like_count": 0,803 "views": 286,804 "category_id": 7,805 "featured_link": null,806 "has_accepted_answer": false,807 "posters": [808 {809 "extras": "latest",810 "description": "Original Poster, Most Recent Poster",811 "user": {812 "id": 80429,813 "username": "Neirda",814 "name": "",815 "avatar_template": "/user_avatar/discuss.pytorch.org/neirda/{size}/74356_2.png",816 "trust_level": 1817 }818 },819 {820 "extras": null,821 "description": "Frequent Poster",822 "user": {823 "id": 41396,824 "username": "soulitzer",825 "name": "",826 "avatar_template": "/letter_avatar_proxy/v4/letter/s/839c29/{size}.png",827 "trust_level": 2828 }829 }830 ]831 },832 {833 "fancy_title": "activation memory",834 "id": 213468,835 "title": "activation memory",836 "slug": "activation-memory",837 "posts_count": 2,838 "reply_count": 0,839 "highest_post_number": 2,840 "image_url": null,841 "created_at": "2024-11-26T13:48:42.606Z",842 "last_posted_at": "2024-12-12T16:08:06.342Z",843 "bumped": true,844 "bumped_at": "2024-12-12T16:08:06.342Z",845 "archetype": "regular",846 "unseen": false,847 "pinned": false,848 "unpinned": null,849 "visible": true,850 "closed": false,851 "archived": false,852 "bookmarked": null,853 "liked": null,854 "tags_descriptions": {},855 "like_count": 0,856 "views": 171,857 "category_id": 7,858 "featured_link": null,859 "has_accepted_answer": false,860 "posters": [861 {862 "extras": null,863 "description": "Original Poster",864 "user": {865 "id": 81127,866 "username": "amber",867 "name": "",868 "avatar_template": "/letter_avatar_proxy/v4/letter/a/9fc348/{size}.png",869 "trust_level": 0870 }871 },872 {873 "extras": "latest",874 "description": "Most Recent Poster",875 "user": {876 "id": 41396,877 "username": "soulitzer",878 "name": "",879 "avatar_template": "/letter_avatar_proxy/v4/letter/s/839c29/{size}.png",880 "trust_level": 2881 }882 }883 ]884 },885 {886 "fancy_title": "Gradient computation has been modified by an inplace operation. There is no inplace operation in my code",887 "id": 212157,888 "title": "Gradient computation has been modified by an inplace operation. There is no inplace operation in my code",889 "slug": "gradient-computation-has-been-modified-by-an-inplace-operation-there-is-no-inplace-operation-in-my-code",890 "posts_count": 2,891 "reply_count": 0,892 "highest_post_number": 2,893 "image_url": null,894 "created_at": "2024-10-27T06:13:54.768Z",895 "last_posted_at": "2024-10-29T00:11:42.243Z",896 "bumped": true,897 "bumped_at": "2024-10-29T00:11:42.243Z",898 "archetype": "regular",899 "unseen": false,900 "pinned": false,901 "unpinned": null,902 "visible": true,903 "closed": false,904 "archived": false,905 "bookmarked": null,906 "liked": null,907 "tags_descriptions": {},908 "like_count": 0,909 "views": 193,910 "category_id": 7,911 "featured_link": null,912 "has_accepted_answer": false,913 "posters": [914 {915 "extras": null,916 "description": "Original Poster",917 "user": {918 "id": 80510,919 "username": "Khushal_Sharma",920 "name": "Khushal Sharma",921 "avatar_template": "/user_avatar/discuss.pytorch.org/khushal_sharma/{size}/73593_2.png",922 "trust_level": 0923 }924 },925 {926 "extras": "latest",927 "description": "Most Recent Poster",928 "user": {929 "id": 3534,930 "username": "ptrblck",931 "name": "",932 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",933 "admin": true,934 "moderator": true,935 "trust_level": 2936 }937 }938 ]939 },940 {941 "fancy_title": "Segfault in autograd after using torch lightning",942 "id": 219723,943 "title": "Segfault in autograd after using torch lightning",944 "slug": "segfault-in-autograd-after-using-torch-lightning",945 "posts_count": 2,946 "reply_count": 0,947 "highest_post_number": 2,948 "image_url": null,949 "created_at": "2025-05-04T04:09:05.621Z",950 "last_posted_at": "2025-05-04T16:37:38.109Z",951 "bumped": true,952 "bumped_at": "2025-05-04T16:37:38.109Z",953 "archetype": "regular",954 "unseen": false,955 "pinned": false,956 "unpinned": null,957 "visible": true,958 "closed": false,959 "archived": false,960 "bookmarked": null,961 "liked": null,962 "tags_descriptions": {},963 "like_count": 0,964 "views": 80,965 "category_id": 7,966 "featured_link": null,967 "has_accepted_answer": false,968 "posters": [969 {970 "extras": null,971 "description": "Original Poster",972 "user": {973 "id": 84135,974 "username": "mxposed",975 "name": "Nick Markov",976 "avatar_template": "/user_avatar/discuss.pytorch.org/mxposed/{size}/76899_2.png",977 "trust_level": 0978 }979 },980 {981 "extras": "latest",982 "description": "Most Recent Poster",983 "user": {984 "id": 3534,985 "username": "ptrblck",986 "name": "",987 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",988 "admin": true,989 "moderator": true,990 "trust_level": 2991 }992 }993 ]994 }995 ],996 "tags_descriptions": {},997 "fancy_title": "Compute the gradients of all logins with respect to input",998 "id": 157613,999 "title": "Compute the gradients of all logins with respect to input",1000 "posts_count": 4,1001 "created_at": "2022-07-27T13:28:42.919Z",1002 "views": 537,1003 "reply_count": 2,1004 "like_count": 0,1005 "last_posted_at": "2022-07-27T16:12:03.081Z",1006 "visible": true,1007 "closed": false,1008 "archived": false,1009 "has_summary": false,1010 "archetype": "regular",1011 "slug": "compute-the-gradients-of-all-logins-with-respect-to-input",1012 "category_id": 7,1013 "word_count": 297,1014 "deleted_at": null,1015 "user_id": 52233,1016 "featured_link": null,1017 "pinned_globally": false,1018 "pinned_at": null,1019 "pinned_until": null,1020 "image_url": null,1021 "slow_mode_seconds": 0,1022 "draft": null,1023 "draft_key": "topic_157613",1024 "draft_sequence": null,1025 "unpinned": null,1026 "pinned": false,1027 "current_post_number": 1,1028 "highest_post_number": 4,1029 "deleted_by": null,1030 "actions_summary": [1031 {1032 "id": 4,1033 "count": 0,1034 "hidden": false,1035 "can_act": false1036 },1037 {1038 "id": 8,1039 "count": 0,1040 "hidden": false,1041 "can_act": false1042 },1043 {1044 "id": 10,1045 "count": 0,1046 "hidden": false,1047 "can_act": false1048 },1049 {1050 "id": 7,1051 "count": 0,1052 "hidden": false,1053 "can_act": false1054 }1055 ],1056 "chunk_size": 20,1057 "bookmarked": false,1058 "topic_timer": null,1059 "message_bus_last_id": 0,1060 "participant_count": 2,1061 "show_read_indicator": false,1062 "thumbnails": null,1063 "slow_mode_enabled_until": null,1064 "can_vote": false,1065 "vote_count": 0,1066 "user_voted": false,1067 "discourse_zendesk_plugin_zendesk_id": null,1068 "discourse_zendesk_plugin_zendesk_url": "https://your-url.zendesk.com/agent/tickets/",1069 "details": {1070 "can_edit": false,1071 "notification_level": 1,1072 "participants": [1073 {1074 "id": 34294,1075 "username": "AlphaBetaGamma96",1076 "name": "",1077 "avatar_template": "/letter_avatar_proxy/v4/letter/a/3da27b/{size}.png",1078 "post_count": 2,1079 "primary_group_name": null,1080 "flair_name": null,1081 "flair_url": null,1082 "flair_color": null,1083 "flair_bg_color": null,1084 "flair_group_id": null,1085 "trust_level": 21086 },1087 {1088 "id": 52233,1089 "username": "Yuancheng_Xu",1090 "name": "Yuancheng Xu",1091 "avatar_template": "/user_avatar/discuss.pytorch.org/yuancheng_xu/{size}/45613_2.png",1092 "post_count": 2,1093 "primary_group_name": null,1094 "flair_name": null,1095 "flair_url": null,1096 "flair_color": null,1097 "flair_bg_color": null,1098 "flair_group_id": null,1099 "trust_level": 11100 }1101 ],1102 "created_by": {1103 "id": 52233,1104 "username": "Yuancheng_Xu",1105 "name": "Yuancheng Xu",1106 "avatar_template": "/user_avatar/discuss.pytorch.org/yuancheng_xu/{size}/45613_2.png"1107 },1108 "last_poster": {1109 "id": 34294,1110 "username": "AlphaBetaGamma96",1111 "name": "",1112 "avatar_template": "/letter_avatar_proxy/v4/letter/a/3da27b/{size}.png"1113 }1114 },1115 "bookmarks": []1116 },1117 {1118 "post_stream": {1119 "posts": [1120 {1121 "id": 347752,1122 "name": "Peter Lorenz",1123 "username": "jS5t3r",1124 "avatar_template": "/user_avatar/discuss.pytorch.org/js5t3r/{size}/40545_2.png",1125 "created_at": "2022-05-21T19:22:23.036Z",1126 "cooked": "<p>I want to find a data structure to save intermediate results. Numpy and Pickle is not an option because I cannot load the data in chunks.</p>\n<p>The data is images, but saving <code>.jpeg</code> etc… is a loss of information.</p>\n<p>How can I save these intermediate results and which enables me to load them memory efficient with a data loader?</p>",1127 "post_number": 1,1128 "post_type": 1,1129 "posts_count": 5,1130 "updated_at": "2022-05-21T19:22:23.036Z",1131 "reply_count": 1,1132 "reply_to_post_number": null,1133 "quote_count": 0,1134 "incoming_link_count": 100,1135 "reads": 7,1136 "readers_count": 6,1137 "score": 506.4,1138 "yours": false,1139 "topic_id": 152288,1140 "topic_slug": "how-to-save-intermediate-results-and-load-it-later-with-a-datalaoder-again",1141 "display_username": "Peter Lorenz",1142 "primary_group_name": null,1143 "flair_name": null,1144 "flair_url": null,1145 "flair_bg_color": null,1146 "flair_color": null,1147 "flair_group_id": null,1148 "badges_granted": [],1149 "version": 1,1150 "can_edit": false,1151 "can_delete": false,1152 "can_recover": false,1153 "can_see_hidden_post": false,1154 "can_wiki": false,1155 "read": true,1156 "user_title": null,1157 "bookmarked": false,1158 "actions_summary": [],1159 "moderator": false,1160 "admin": false,1161 "staff": false,1162 "user_id": 43114,1163 "hidden": false,1164 "trust_level": 1,1165 "deleted_at": null,1166 "user_deleted": false,1167 "edit_reason": null,1168 "can_view_edit_history": true,1169 "wiki": false,1170 "post_url": "/t/how-to-save-intermediate-results-and-load-it-later-with-a-datalaoder-again/152288/1",1171 "can_accept_answer": false,1172 "can_unaccept_answer": false,1173 "accepted_answer": false,1174 "topic_accepted_answer": null,1175 "can_vote": false1176 },1177 {1178 "id": 347920,1179 "name": "",1180 "username": "ptrblck",1181 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",1182 "created_at": "2022-05-23T03:43:21.444Z",1183 "cooked": "<aside class=\"quote no-group\" data-username=\"jS5t3r\" data-post=\"1\" data-topic=\"152288\">\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/js5t3r/48/40545_2.png\" class=\"avatar\"> jS5t3r:</div>\n<blockquote>\n<p>Numpy and Pickle is not an option because I cannot load the data in chunks.</p>\n</blockquote>\n</aside>\n<p>I don’t quite understand the “cannot load the data in chunks” part, but given that I guess <code>torch.save</code> is not an option.</p>\n<aside class=\"quote no-group\" data-username=\"jS5t3r\" data-post=\"1\" data-topic=\"152288\">\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/js5t3r/48/40545_2.png\" class=\"avatar\"> jS5t3r:</div>\n<blockquote>\n<p>The data is images, but saving <code>.jpeg</code> etc… is a loss of information.</p>\n</blockquote>\n</aside>\n<p>You could use a lossless compression format such as PNG.</p>",1184 "post_number": 2,1185 "post_type": 1,1186 "posts_count": 5,1187 "updated_at": "2022-05-23T03:43:21.444Z",1188 "reply_count": 1,1189 "reply_to_post_number": null,1190 "quote_count": 1,1191 "incoming_link_count": 0,1192 "reads": 4,1193 "readers_count": 3,1194 "score": 5.8,1195 "yours": false,1196 "topic_id": 152288,1197 "topic_slug": "how-to-save-intermediate-results-and-load-it-later-with-a-datalaoder-again",1198 "display_username": "",1199 "primary_group_name": null,1200 "flair_name": null,