Anurag1734/cuda-error-resolution-analysis
07
1[2 {3 "post_stream": {4 "posts": [5 {6 "id": 402593,7 "name": "Vo Van Tu",8 "username": "tuvovan",9 "avatar_template": "/user_avatar/discuss.pytorch.org/tuvovan/{size}/60611_2.png",10 "created_at": "2023-05-18T08:00:31.254Z",11 "cooked": "<p>Hi, I’m trying to implement the block matching algorithm using torch.</p>\n<p>The basic idea is to take a patch say 8 by 8, then define a search window say 50 by 50 and we need to find the top 10 most similar patches inside that search window.</p>\n<p>I had idea to search for top 10 most similar patches in the whole image using unfold + calculate distance + sort function but not quite sure how to do if there is a search window.</p>\n<p>Thanks!</p>",12 "post_number": 1,13 "post_type": 1,14 "posts_count": 7,15 "updated_at": "2023-05-18T08:01:54.597Z",16 "reply_count": 0,17 "reply_to_post_number": null,18 "quote_count": 0,19 "incoming_link_count": 462,20 "reads": 14,21 "readers_count": 13,22 "score": 2312.8,23 "yours": false,24 "topic_id": 180212,25 "topic_slug": "block-matching-algorithm",26 "display_username": "Vo Van Tu",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 "moderator": false,45 "admin": false,46 "staff": false,47 "user_id": 66286,48 "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/block-matching-algorithm/180212/1",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": 402605,64 "name": "Arul",65 "username": "InnovArul",66 "avatar_template": "/user_avatar/discuss.pytorch.org/innovarul/{size}/5282_2.png",67 "created_at": "2023-05-18T10:23:36.906Z",68 "cooked": "<p>This might be of some help:</p>\n<aside class=\"onebox allowlistedgeneric\" data-onebox-src=\"https://github.com/ClementPinard/Pytorch-Correlation-extension\">\n <header class=\"source\">\n <img src=\"https://github.githubassets.com/favicons/favicon.svg\" class=\"site-icon\" width=\"32\" height=\"32\">\n\n <a href=\"https://github.com/ClementPinard/Pytorch-Correlation-extension\" target=\"_blank\" rel=\"noopener\">GitHub</a>\n </header>\n\n <article class=\"onebox-body\">\n <div class=\"aspect-image\" style=\"--aspect-ratio:690/345;\"><img src=\"https://opengraph.githubassets.com/bbb0f77b8198961684284730d4c8f8572ca2477f81cf000ba63ebed242c05bc3/ClementPinard/Pytorch-Correlation-extension\" class=\"thumbnail\" width=\"690\" height=\"345\"></div>\n\n<h3><a href=\"https://github.com/ClementPinard/Pytorch-Correlation-extension\" target=\"_blank\" rel=\"noopener\">GitHub - ClementPinard/Pytorch-Correlation-extension: Custom implementation...</a></h3>\n\n <p>Custom implementation of Corrleation Module. Contribute to ClementPinard/Pytorch-Correlation-extension development by creating an account on GitHub.</p>\n\n\n </article>\n\n <div class=\"onebox-metadata\">\n \n \n </div>\n\n <div style=\"clear: both\"></div>\n</aside>\n",69 "post_number": 2,70 "post_type": 1,71 "posts_count": 7,72 "updated_at": "2023-05-18T10:23:36.906Z",73 "reply_count": 1,74 "reply_to_post_number": null,75 "quote_count": 0,76 "incoming_link_count": 2,77 "reads": 14,78 "readers_count": 13,79 "score": 17.8,80 "yours": false,81 "topic_id": 180212,82 "topic_slug": "block-matching-algorithm",83 "display_username": "Arul",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 "link_counts": [98 {99 "url": "https://github.com/ClementPinard/Pytorch-Correlation-extension",100 "internal": false,101 "reflection": false,102 "title": "GitHub - ClementPinard/Pytorch-Correlation-extension: Custom implementation of Corrleation Module",103 "clicks": 60104 }105 ],106 "read": true,107 "user_title": "",108 "bookmarked": false,109 "actions_summary": [],110 "moderator": false,111 "admin": false,112 "staff": false,113 "user_id": 998,114 "hidden": false,115 "trust_level": 2,116 "deleted_at": null,117 "user_deleted": false,118 "edit_reason": null,119 "can_view_edit_history": true,120 "wiki": false,121 "post_url": "/t/block-matching-algorithm/180212/2",122 "can_accept_answer": false,123 "can_unaccept_answer": false,124 "accepted_answer": false,125 "topic_accepted_answer": null126 },127 {128 "id": 402626,129 "name": "Vo Van Tu",130 "username": "tuvovan",131 "avatar_template": "/user_avatar/discuss.pytorch.org/tuvovan/{size}/60611_2.png",132 "created_at": "2023-05-18T14:29:37.473Z",133 "cooked": "<p>you mean this might help with the distance step?</p>",134 "post_number": 3,135 "post_type": 1,136 "posts_count": 7,137 "updated_at": "2023-05-18T14:29:37.473Z",138 "reply_count": 1,139 "reply_to_post_number": 2,140 "quote_count": 0,141 "incoming_link_count": 5,142 "reads": 11,143 "readers_count": 10,144 "score": 32.2,145 "yours": false,146 "topic_id": 180212,147 "topic_slug": "block-matching-algorithm",148 "display_username": "Vo Van Tu",149 "primary_group_name": null,150 "flair_name": null,151 "flair_url": null,152 "flair_bg_color": null,153 "flair_color": null,154 "flair_group_id": null,155 "badges_granted": [],156 "version": 1,157 "can_edit": false,158 "can_delete": false,159 "can_recover": false,160 "can_see_hidden_post": false,161 "can_wiki": false,162 "read": true,163 "user_title": null,164 "reply_to_user": {165 "id": 998,166 "username": "InnovArul",167 "name": "Arul",168 "avatar_template": "/user_avatar/discuss.pytorch.org/innovarul/{size}/5282_2.png"169 },170 "bookmarked": false,171 "actions_summary": [],172 "moderator": false,173 "admin": false,174 "staff": false,175 "user_id": 66286,176 "hidden": false,177 "trust_level": 1,178 "deleted_at": null,179 "user_deleted": false,180 "edit_reason": null,181 "can_view_edit_history": true,182 "wiki": false,183 "post_url": "/t/block-matching-algorithm/180212/3",184 "can_accept_answer": false,185 "can_unaccept_answer": false,186 "accepted_answer": false,187 "topic_accepted_answer": null188 },189 {190 "id": 402636,191 "name": "Arul",192 "username": "InnovArul",193 "avatar_template": "/user_avatar/discuss.pytorch.org/innovarul/{size}/5282_2.png",194 "created_at": "2023-05-18T15:27:50.965Z",195 "cooked": "<p>Yes. This package may help to match patches between feature maps.<br>\nI assumed the matching should happen at every pixel, where you consider the <code>8x8</code> patch at every pixel position and match it within the <code>50x50</code> window around it. Maybe this is not the case?</p>\n<p>Not sure of your use case. If its matching one <code>8x8</code> patch with a particular <code>50x50</code> window, you could achieve it by convolving (<code>F.conv2d</code>?) this <code>8x8</code> patch over the <code>50x50</code> window, then sort the responses.</p>",196 "post_number": 4,197 "post_type": 1,198 "posts_count": 7,199 "updated_at": "2023-05-18T15:33:27.775Z",200 "reply_count": 0,201 "reply_to_post_number": 3,202 "quote_count": 0,203 "incoming_link_count": 0,204 "reads": 9,205 "readers_count": 8,206 "score": 1.8,207 "yours": false,208 "topic_id": 180212,209 "topic_slug": "block-matching-algorithm",210 "display_username": "Arul",211 "primary_group_name": null,212 "flair_name": null,213 "flair_url": null,214 "flair_bg_color": null,215 "flair_color": null,216 "flair_group_id": null,217 "badges_granted": [],218 "version": 2,219 "can_edit": false,220 "can_delete": false,221 "can_recover": false,222 "can_see_hidden_post": false,223 "can_wiki": false,224 "read": true,225 "user_title": "",226 "reply_to_user": {227 "id": 66286,228 "username": "tuvovan",229 "name": "Vo Van Tu",230 "avatar_template": "/user_avatar/discuss.pytorch.org/tuvovan/{size}/60611_2.png"231 },232 "bookmarked": false,233 "actions_summary": [],234 "moderator": false,235 "admin": false,236 "staff": false,237 "user_id": 998,238 "hidden": false,239 "trust_level": 2,240 "deleted_at": null,241 "user_deleted": false,242 "edit_reason": null,243 "can_view_edit_history": true,244 "wiki": false,245 "post_url": "/t/block-matching-algorithm/180212/4",246 "can_accept_answer": false,247 "can_unaccept_answer": false,248 "accepted_answer": false,249 "topic_accepted_answer": null250 },251 {252 "id": 402703,253 "name": "Vo Van Tu",254 "username": "tuvovan",255 "avatar_template": "/user_avatar/discuss.pytorch.org/tuvovan/{size}/60611_2.png",256 "created_at": "2023-05-19T01:12:01.842Z",257 "cooked": "<p>yeah right! the matching should happen at every pixel.</p>\n<p>Let say there is a 256x256 image, for each pixel of that image, take a patch of 8x8 around it and start looking for similar patches inside a search window of 50x50.<br>\n2 nested for loops might work, not sure if there is a better solution using vector or tensor…</p>",258 "post_number": 5,259 "post_type": 1,260 "posts_count": 7,261 "updated_at": "2023-05-19T01:12:01.842Z",262 "reply_count": 1,263 "reply_to_post_number": null,264 "quote_count": 0,265 "incoming_link_count": 7,266 "reads": 7,267 "readers_count": 6,268 "score": 41.4,269 "yours": false,270 "topic_id": 180212,271 "topic_slug": "block-matching-algorithm",272 "display_username": "Vo Van Tu",273 "primary_group_name": null,274 "flair_name": null,275 "flair_url": null,276 "flair_bg_color": null,277 "flair_color": null,278 "flair_group_id": null,279 "badges_granted": [],280 "version": 1,281 "can_edit": false,282 "can_delete": false,283 "can_recover": false,284 "can_see_hidden_post": false,285 "can_wiki": false,286 "read": true,287 "user_title": null,288 "bookmarked": false,289 "actions_summary": [],290 "moderator": false,291 "admin": false,292 "staff": false,293 "user_id": 66286,294 "hidden": false,295 "trust_level": 1,296 "deleted_at": null,297 "user_deleted": false,298 "edit_reason": null,299 "can_view_edit_history": true,300 "wiki": false,301 "post_url": "/t/block-matching-algorithm/180212/5",302 "can_accept_answer": false,303 "can_unaccept_answer": false,304 "accepted_answer": false,305 "topic_accepted_answer": null306 },307 {308 "id": 402885,309 "name": "Arul",310 "username": "InnovArul",311 "avatar_template": "/user_avatar/discuss.pytorch.org/innovarul/{size}/5282_2.png",312 "created_at": "2023-05-21T07:45:05.497Z",313 "cooked": "<p>Then my assumption was correct. Thanks for clarifying.<br>\nThe <code>Pytorch-Correlation-extension</code> package I linked above provides the necessary functionality.<br>\nYou can look at the examples provided in the Github and see if it works for you.</p><aside class=\"onebox allowlistedgeneric\" data-onebox-src=\"https://github.com/ClementPinard/Pytorch-Correlation-extension#example\">\n <header class=\"source\">\n <img src=\"https://github.githubassets.com/favicons/favicon.svg\" class=\"site-icon\" width=\"32\" height=\"32\">\n\n <a href=\"https://github.com/ClementPinard/Pytorch-Correlation-extension#example\" target=\"_blank\" rel=\"noopener\">GitHub</a>\n </header>\n\n <article class=\"onebox-body\">\n <div class=\"aspect-image\" style=\"--aspect-ratio:690/345;\"><img src=\"https://opengraph.githubassets.com/b83d96e07786cd9be7946fcb2bf60c15c7a27265e936f9713d086b8a1f5d6ce9/ClementPinard/Pytorch-Correlation-extension\" class=\"thumbnail\" width=\"690\" height=\"345\"></div>\n\n<h3><a href=\"https://github.com/ClementPinard/Pytorch-Correlation-extension#example\" target=\"_blank\" rel=\"noopener\">GitHub - ClementPinard/Pytorch-Correlation-extension: Custom implementation...</a></h3>\n\n <p>Custom implementation of Corrleation Module. Contribute to ClementPinard/Pytorch-Correlation-extension development by creating an account on GitHub.</p>\n\n\n </article>\n\n <div class=\"onebox-metadata\">\n \n \n </div>\n\n <div style=\"clear: both\"></div>\n</aside>\n",314 "post_number": 6,315 "post_type": 1,316 "posts_count": 7,317 "updated_at": "2023-05-21T07:45:05.497Z",318 "reply_count": 0,319 "reply_to_post_number": 5,320 "quote_count": 0,321 "incoming_link_count": 6,322 "reads": 5,323 "readers_count": 4,324 "score": 31.0,325 "yours": false,326 "topic_id": 180212,327 "topic_slug": "block-matching-algorithm",328 "display_username": "Arul",329 "primary_group_name": null,330 "flair_name": null,331 "flair_url": null,332 "flair_bg_color": null,333 "flair_color": null,334 "flair_group_id": null,335 "badges_granted": [],336 "version": 1,337 "can_edit": false,338 "can_delete": false,339 "can_recover": false,340 "can_see_hidden_post": false,341 "can_wiki": false,342 "link_counts": [343 {344 "url": "https://github.com/ClementPinard/Pytorch-Correlation-extension#example",345 "internal": false,346 "reflection": false,347 "title": "GitHub - ClementPinard/Pytorch-Correlation-extension: Custom implementation of Corrleation Module",348 "clicks": 30349 }350 ],351 "read": true,352 "user_title": "",353 "reply_to_user": {354 "id": 66286,355 "username": "tuvovan",356 "name": "Vo Van Tu",357 "avatar_template": "/user_avatar/discuss.pytorch.org/tuvovan/{size}/60611_2.png"358 },359 "bookmarked": false,360 "actions_summary": [],361 "moderator": false,362 "admin": false,363 "staff": false,364 "user_id": 998,365 "hidden": false,366 "trust_level": 2,367 "deleted_at": null,368 "user_deleted": false,369 "edit_reason": null,370 "can_view_edit_history": true,371 "wiki": false,372 "post_url": "/t/block-matching-algorithm/180212/6",373 "can_accept_answer": false,374 "can_unaccept_answer": false,375 "accepted_answer": false,376 "topic_accepted_answer": null377 },378 {379 "id": 402897,380 "name": "J Johnson",381 "username": "J_Johnson",382 "avatar_template": "/user_avatar/discuss.pytorch.org/j_johnson/{size}/55494_2.png",383 "created_at": "2023-05-21T11:39:15.141Z",384 "cooked": "<p>Here is one way you could approach the problem in a parallelized way. This assumes you want to compare “patches” with l1loss(though you could easily substitute MSE).</p>\n<pre><code class=\"lang-auto\">import torch\nimport torch.nn.functional as F\n\ndef l1loss(patch1, patch2, dim=1):\n return torch.mean(torch.abs((patch1-patch2)), dim=dim)\n\ndef get_patches(images, kernel_size=(8,8)):\n return F.unfold(images, kernel_size) # output returns size (batch, flattened patch, patches)\n\nimages=torch.randn((1,1,50, 50))\n\npatches=get_patches(images)\nb, hw, p = patches.shape\npatches_exp=patches.unsqueeze(3).expand(b, hw, p, p)\n\nz=torch.triu_indices(p,p,1) #mapping of the triangular upper matrix\n\nlosses = l1loss(patches_exp.triu(), patches_exp.rot90(k=1, dims=[2,3]).triu())\nlosses=losses[losses!=0]\nprint(losses.size()) #size should be p*((p-1)/2) which represents non-zeros for triu when diagonals are zero\nvalues, indices = torch.topk(losses, k=10, largest=False) # top 10 values and their indices\nprint(values, indices)\nindex=0\nx_val=z[0][indices[index]] # get the indices from the triu_indices mapping at the selected topk index\ny_val=z[1][indices[index]]\n\nprint(x_val, y_val, values[index]) #check the indices and value\nprint(l1loss(patches[:,:,x_val],patches[:,:,y_val])) #check that the loss matches when the indices are applied to the original unfolded patches\n</code></pre>\n<p><code>.triu()</code> is used to eliminate duplicate calculations or getting the losses between the same patch.</p>\n<p>Updated to correct an error and include a usage example with <code>topk()</code>.</p>\n<p><a href=\"https://pytorch.org/docs/stable/generated/torch.topk.html\" class=\"onebox\" target=\"_blank\" rel=\"noopener nofollow ugc\">https://pytorch.org/docs/stable/generated/torch.topk.html</a></p>\n<p>Note that with larger images, you’re going to run into some major memory problems and may need to split up the operation into an iterable.</p>",385 "post_number": 7,386 "post_type": 1,387 "posts_count": 7,388 "updated_at": "2023-05-21T17:25:08.902Z",389 "reply_count": 0,390 "reply_to_post_number": null,391 "quote_count": 0,392 "incoming_link_count": 41,393 "reads": 5,394 "readers_count": 4,395 "score": 206.0,396 "yours": false,397 "topic_id": 180212,398 "topic_slug": "block-matching-algorithm",399 "display_username": "J Johnson",400 "primary_group_name": null,401 "flair_name": null,402 "flair_url": null,403 "flair_bg_color": null,404 "flair_color": null,405 "flair_group_id": null,406 "badges_granted": [],407 "version": 3,408 "can_edit": false,409 "can_delete": false,410 "can_recover": false,411 "can_see_hidden_post": false,412 "can_wiki": false,413 "link_counts": [414 {415 "url": "https://pytorch.org/docs/stable/generated/torch.topk.html",416 "internal": false,417 "reflection": false,418 "title": "torch.topk — PyTorch 2.0 documentation",419 "clicks": 4420 }421 ],422 "read": true,423 "user_title": null,424 "bookmarked": false,425 "actions_summary": [],426 "moderator": false,427 "admin": false,428 "staff": false,429 "user_id": 41458,430 "hidden": false,431 "trust_level": 2,432 "deleted_at": null,433 "user_deleted": false,434 "edit_reason": null,435 "can_view_edit_history": true,436 "wiki": false,437 "post_url": "/t/block-matching-algorithm/180212/7",438 "can_accept_answer": false,439 "can_unaccept_answer": false,440 "accepted_answer": false,441 "topic_accepted_answer": null442 }443 ],444 "stream": [445 402593,446 402605,447 402626,448 402636,449 402703,450 402885,451 402897452 ]453 },454 "timeline_lookup": [455 [456 1,457 891458 ],459 [460 6,461 888462 ]463 ],464 "suggested_topics": [465 {466 "fancy_title": "ResUnet implementation in pytorch for medical image segmentation",467 "id": 212702,468 "title": "ResUnet implementation in pytorch for medical image segmentation",469 "slug": "resunet-implementation-in-pytorch-for-medical-image-segmentation",470 "posts_count": 2,471 "reply_count": 0,472 "highest_post_number": 2,473 "image_url": null,474 "created_at": "2024-11-08T11:32:30.168Z",475 "last_posted_at": "2024-11-11T19:57:35.383Z",476 "bumped": true,477 "bumped_at": "2024-11-11T19:57:35.383Z",478 "archetype": "regular",479 "unseen": false,480 "pinned": false,481 "unpinned": null,482 "visible": true,483 "closed": false,484 "archived": false,485 "bookmarked": null,486 "liked": null,487 "tags_descriptions": {},488 "like_count": 0,489 "views": 106,490 "category_id": 5,491 "featured_link": null,492 "has_accepted_answer": false,493 "posters": [494 {495 "extras": null,496 "description": "Original Poster",497 "user": {498 "id": 80772,499 "username": "Saurav_K_Mhptra",500 "name": "Saurav Kausik Mahapatra",501 "avatar_template": "/user_avatar/discuss.pytorch.org/saurav_k_mhptra/{size}/73231_2.png",502 "trust_level": 0503 }504 },505 {506 "extras": "latest",507 "description": "Most Recent Poster",508 "user": {509 "id": 75871,510 "username": "qq-me",511 "name": "Ivan Nikishev",512 "avatar_template": "/user_avatar/discuss.pytorch.org/qq-me/{size}/70055_2.png",513 "trust_level": 2514 }515 }516 ]517 },518 {519 "fancy_title": "Adding softmax layer with crossentropy",520 "id": 214246,521 "title": "Adding softmax layer with crossentropy",522 "slug": "adding-softmax-layer-with-crossentropy",523 "posts_count": 2,524 "reply_count": 0,525 "highest_post_number": 2,526 "image_url": null,527 "created_at": "2024-12-15T15:31:16.823Z",528 "last_posted_at": "2024-12-15T17:04:48.668Z",529 "bumped": true,530 "bumped_at": "2024-12-15T17:04:48.668Z",531 "archetype": "regular",532 "unseen": false,533 "pinned": false,534 "unpinned": null,535 "visible": true,536 "closed": false,537 "archived": false,538 "bookmarked": null,539 "liked": null,540 "tags_descriptions": {},541 "like_count": 0,542 "views": 174,543 "category_id": 5,544 "featured_link": null,545 "has_accepted_answer": false,546 "posters": [547 {548 "extras": null,549 "description": "Original Poster",550 "user": {551 "id": 46221,552 "username": "mathwseg",553 "name": "mathwseg",554 "avatar_template": "/user_avatar/discuss.pytorch.org/mathwseg/{size}/39221_2.png",555 "trust_level": 1556 }557 },558 {559 "extras": "latest",560 "description": "Most Recent Poster",561 "user": {562 "id": 41396,563 "username": "soulitzer",564 "name": "",565 "avatar_template": "/letter_avatar_proxy/v4/letter/s/839c29/{size}.png",566 "trust_level": 2567 }568 }569 ]570 },571 {572 "fancy_title": "Getting “Could not initialize NNPACK! Reason: Unsupported hardware.” warning even though NNPACK is enabled",573 "id": 214397,574 "title": "Getting \"Could not initialize NNPACK! Reason: Unsupported hardware.\" warning even though NNPACK is enabled",575 "slug": "getting-could-not-initialize-nnpack-reason-unsupported-hardware-warning-even-though-nnpack-is-enabled",576 "posts_count": 1,577 "reply_count": 0,578 "highest_post_number": 1,579 "image_url": null,580 "created_at": "2024-12-19T09:39:58.123Z",581 "last_posted_at": "2024-12-19T09:39:58.160Z",582 "bumped": true,583 "bumped_at": "2024-12-19T09:39:58.160Z",584 "archetype": "regular",585 "unseen": false,586 "pinned": false,587 "unpinned": null,588 "visible": true,589 "closed": false,590 "archived": false,591 "bookmarked": null,592 "liked": null,593 "tags_descriptions": {},594 "like_count": 0,595 "views": 238,596 "category_id": 5,597 "featured_link": null,598 "has_accepted_answer": false,599 "posters": [600 {601 "extras": "latest single",602 "description": "Original Poster, Most Recent Poster",603 "user": {604 "id": 81586,605 "username": "KirilloCirillo",606 "name": "",607 "avatar_template": "/user_avatar/discuss.pytorch.org/kirillocirillo/{size}/74611_2.png",608 "trust_level": 1609 }610 }611 ]612 },613 {614 "fancy_title": "Help :frowning: Auto encoder predicts shapes ‘ok’ but lacks colours",615 "id": 214163,616 "title": "Help :( Auto encoder predicts shapes 'ok' but lacks colours",617 "slug": "help-auto-encoder-predicts-shapes-ok-but-lacks-colours",618 "posts_count": 3,619 "reply_count": 2,620 "highest_post_number": 5,621 "image_url": "https://discuss.pytorch.org/uploads/default/original/3X/5/0/501da039c27c2a1804a96da5f766dbf4ec662e9a.png",622 "created_at": "2024-12-12T18:31:06.027Z",623 "last_posted_at": "2024-12-19T14:51:00.132Z",624 "bumped": true,625 "bumped_at": "2024-12-19T14:51:00.132Z",626 "archetype": "regular",627 "unseen": false,628 "pinned": false,629 "unpinned": null,630 "visible": true,631 "closed": false,632 "archived": false,633 "bookmarked": null,634 "liked": null,635 "tags_descriptions": {},636 "like_count": 1,637 "views": 82,638 "category_id": 5,639 "featured_link": null,640 "has_accepted_answer": false,641 "posters": [642 {643 "extras": "latest",644 "description": "Original Poster, Most Recent Poster",645 "user": {646 "id": 81089,647 "username": "Aknw_Fen",648 "name": "Aknw Fen",649 "avatar_template": "/user_avatar/discuss.pytorch.org/aknw_fen/{size}/74156_2.png",650 "trust_level": 2651 }652 },653 {654 "extras": null,655 "description": "Frequent Poster",656 "user": {657 "id": 77908,658 "username": "mycul",659 "name": "",660 "avatar_template": "/user_avatar/discuss.pytorch.org/mycul/{size}/72394_2.png",661 "trust_level": 2662 }663 }664 ]665 },666 {667 "fancy_title": "Premultiplied alpha",668 "id": 220230,669 "title": "Premultiplied alpha",670 "slug": "premultiplied-alpha",671 "posts_count": 4,672 "reply_count": 2,673 "highest_post_number": 4,674 "image_url": null,675 "created_at": "2025-05-21T16:40:18.345Z",676 "last_posted_at": "2025-05-25T18:54:50.339Z",677 "bumped": true,678 "bumped_at": "2025-05-25T18:54:50.339Z",679 "archetype": "regular",680 "unseen": false,681 "pinned": false,682 "unpinned": null,683 "visible": true,684 "closed": false,685 "archived": false,686 "bookmarked": null,687 "liked": null,688 "tags_descriptions": {},689 "like_count": 1,690 "views": 81,691 "category_id": 5,692 "featured_link": null,693 "has_accepted_answer": true,694 "posters": [695 {696 "extras": null,697 "description": "Original Poster",698 "user": {699 "id": 84391,700 "username": "rdgq",701 "name": "",702 "avatar_template": "/user_avatar/discuss.pytorch.org/rdgq/{size}/77106_2.png",703 "trust_level": 1704 }705 },706 {707 "extras": "latest",708 "description": "Most Recent Poster, Accepted Answer",709 "user": {710 "id": 75120,711 "username": "ajayrkumar",712 "name": "Ajay Rajendra Kumar",713 "avatar_template": "/user_avatar/discuss.pytorch.org/ajayrkumar/{size}/77139_2.png",714 "trust_level": 1715 }716 }717 ]718 }719 ],720 "tags_descriptions": {},721 "fancy_title": "Block Matching algorithm?",722 "id": 180212,723 "title": "Block Matching algorithm?",724 "posts_count": 7,725 "created_at": "2023-05-18T08:00:31.173Z",726 "views": 979,727 "reply_count": 3,728 "like_count": 0,729 "last_posted_at": "2023-05-21T11:39:15.141Z",730 "visible": true,731 "closed": false,732 "archived": false,733 "has_summary": false,734 "archetype": "regular",735 "slug": "block-matching-algorithm",736 "category_id": 5,737 "word_count": 566,738 "deleted_at": null,739 "user_id": 66286,740 "featured_link": null,741 "pinned_globally": false,742 "pinned_at": null,743 "pinned_until": null,744 "image_url": null,745 "slow_mode_seconds": 0,746 "draft": null,747 "draft_key": "topic_180212",748 "draft_sequence": null,749 "unpinned": null,750 "pinned": false,751 "current_post_number": 1,752 "highest_post_number": 7,753 "deleted_by": null,754 "actions_summary": [755 {756 "id": 4,757 "count": 0,758 "hidden": false,759 "can_act": false760 },761 {762 "id": 8,763 "count": 0,764 "hidden": false,765 "can_act": false766 },767 {768 "id": 10,769 "count": 0,770 "hidden": false,771 "can_act": false772 },773 {774 "id": 7,775 "count": 0,776 "hidden": false,777 "can_act": false778 }779 ],780 "chunk_size": 20,781 "bookmarked": false,782 "topic_timer": null,783 "message_bus_last_id": 0,784 "participant_count": 3,785 "show_read_indicator": false,786 "thumbnails": null,787 "slow_mode_enabled_until": null,788 "can_vote": false,789 "vote_count": 0,790 "user_voted": false,791 "discourse_zendesk_plugin_zendesk_id": null,792 "discourse_zendesk_plugin_zendesk_url": "https://your-url.zendesk.com/agent/tickets/",793 "details": {794 "can_edit": false,795 "notification_level": 1,796 "participants": [797 {798 "id": 998,799 "username": "InnovArul",800 "name": "Arul",801 "avatar_template": "/user_avatar/discuss.pytorch.org/innovarul/{size}/5282_2.png",802 "post_count": 3,803 "primary_group_name": null,804 "flair_name": null,805 "flair_url": null,806 "flair_color": null,807 "flair_bg_color": null,808 "flair_group_id": null,809 "trust_level": 2810 },811 {812 "id": 66286,813 "username": "tuvovan",814 "name": "Vo Van Tu",815 "avatar_template": "/user_avatar/discuss.pytorch.org/tuvovan/{size}/60611_2.png",816 "post_count": 3,817 "primary_group_name": null,818 "flair_name": null,819 "flair_url": null,820 "flair_color": null,821 "flair_bg_color": null,822 "flair_group_id": null,823 "trust_level": 1824 },825 {826 "id": 41458,827 "username": "J_Johnson",828 "name": "J Johnson",829 "avatar_template": "/user_avatar/discuss.pytorch.org/j_johnson/{size}/55494_2.png",830 "post_count": 1,831 "primary_group_name": null,832 "flair_name": null,833 "flair_url": null,834 "flair_color": null,835 "flair_bg_color": null,836 "flair_group_id": null,837 "trust_level": 2838 }839 ],840 "created_by": {841 "id": 66286,842 "username": "tuvovan",843 "name": "Vo Van Tu",844 "avatar_template": "/user_avatar/discuss.pytorch.org/tuvovan/{size}/60611_2.png"845 },846 "last_poster": {847 "id": 41458,848 "username": "J_Johnson",849 "name": "J Johnson",850 "avatar_template": "/user_avatar/discuss.pytorch.org/j_johnson/{size}/55494_2.png"851 },852 "links": [853 {854 "url": "https://github.com/ClementPinard/Pytorch-Correlation-extension",855 "title": "GitHub - ClementPinard/Pytorch-Correlation-extension: Custom implementation of Corrleation Module",856 "internal": false,857 "attachment": false,858 "reflection": false,859 "clicks": 60,860 "user_id": 998,861 "domain": "github.com",862 "root_domain": "github.com"863 },864 {865 "url": "https://github.com/ClementPinard/Pytorch-Correlation-extension#example",866 "title": "GitHub - ClementPinard/Pytorch-Correlation-extension: Custom implementation of Corrleation Module",867 "internal": false,868 "attachment": false,869 "reflection": false,870 "clicks": 30,871 "user_id": 998,872 "domain": "github.com",873 "root_domain": "github.com"874 },875 {876 "url": "https://pytorch.org/docs/stable/generated/torch.topk.html",877 "title": "torch.topk — PyTorch 2.0 documentation",878 "internal": false,879 "attachment": false,880 "reflection": false,881 "clicks": 4,882 "user_id": 41458,883 "domain": "pytorch.org",884 "root_domain": "pytorch.org"885 }886 ]887 },888 "bookmarks": []889 },890 {891 "post_stream": {892 "posts": [893 {894 "id": 402914,895 "name": "Rituja Pardhi",896 "username": "Rituja_Pardhi",897 "avatar_template": "/user_avatar/discuss.pytorch.org/rituja_pardhi/{size}/58679_2.png",898 "created_at": "2023-05-21T16:56:30.831Z",899 "cooked": "<p>Getting this error when trying to run GIN with pytorch lightning</p>",900 "post_number": 1,901 "post_type": 1,902 "posts_count": 1,903 "updated_at": "2023-05-21T16:56:30.831Z",904 "reply_count": 0,905 "reply_to_post_number": null,906 "quote_count": 0,907 "incoming_link_count": 18,908 "reads": 3,909 "readers_count": 2,910 "score": 90.6,911 "yours": false,912 "topic_id": 180405,913 "topic_slug": "mat2-is-sparse-internal-assert-failed-at-aten-src-aten-native-sparse-cuda-sparsematmul-cu-786",914 "display_username": "Rituja Pardhi",915 "primary_group_name": null,916 "flair_name": null,917 "flair_url": null,918 "flair_bg_color": null,919 "flair_color": null,920 "flair_group_id": null,921 "badges_granted": [],922 "version": 1,923 "can_edit": false,924 "can_delete": false,925 "can_recover": false,926 "can_see_hidden_post": false,927 "can_wiki": false,928 "read": true,929 "user_title": null,930 "bookmarked": false,931 "actions_summary": [],932 "moderator": false,933 "admin": false,934 "staff": false,935 "user_id": 66363,936 "hidden": false,937 "trust_level": 1,938 "deleted_at": null,939 "user_deleted": false,940 "edit_reason": null,941 "can_view_edit_history": true,942 "wiki": false,943 "post_url": "/t/mat2-is-sparse-internal-assert-failed-at-aten-src-aten-native-sparse-cuda-sparsematmul-cu-786/180405/1",944 "can_accept_answer": false,945 "can_unaccept_answer": false,946 "accepted_answer": false,947 "topic_accepted_answer": null,948 "can_vote": false949 }950 ],951 "stream": [952 402914953 ]954 },955 "timeline_lookup": [956 [957 1,958 888959 ]960 ],961 "suggested_topics": [962 {963 "fancy_title": "Problem with cuda/pytorch on Fedora",964 "id": 214122,965 "title": "Problem with cuda/pytorch on Fedora",966 "slug": "problem-with-cuda-pytorch-on-fedora",967 "posts_count": 3,968 "reply_count": 1,969 "highest_post_number": 3,970 "image_url": null,971 "created_at": "2024-12-11T20:38:53.217Z",972 "last_posted_at": "2024-12-12T11:58:37.395Z",973 "bumped": true,974 "bumped_at": "2024-12-12T11:58:37.395Z",975 "archetype": "regular",976 "unseen": false,977 "pinned": false,978 "unpinned": null,979 "visible": true,980 "closed": false,981 "archived": false,982 "bookmarked": null,983 "liked": null,984 "tags_descriptions": {},985 "like_count": 0,986 "views": 272,987 "category_id": 1,988 "featured_link": null,989 "has_accepted_answer": false,990 "posters": [991 {992 "extras": "latest",993 "description": "Original Poster, Most Recent Poster",994 "user": {995 "id": 81457,996 "username": "Michal_P",997 "name": "Michał P.",998 "avatar_template": "/user_avatar/discuss.pytorch.org/michal_p/{size}/74476_2.png",999 "trust_level": 01000 }1001 },1002 {1003 "extras": null,1004 "description": "Frequent Poster",1005 "user": {1006 "id": 3534,1007 "username": "ptrblck",1008 "name": "",1009 "avatar_template": "/user_avatar/discuss.pytorch.org/ptrblck/{size}/1823_2.png",1010 "admin": true,1011 "moderator": true,1012 "trust_level": 21013 }1014 }1015 ]1016 },1017 {1018 "fancy_title": "Need advice on how to modify the GRU to made it work with raw vibration data",1019 "id": 216261,1020 "title": "Need advice on how to modify the GRU to made it work with raw vibration data",1021 "slug": "need-advice-on-how-to-modify-the-gru-to-made-it-work-with-raw-vibration-data",1022 "posts_count": 4,1023 "reply_count": 1,1024 "highest_post_number": 4,1025 "image_url": null,1026 "created_at": "2025-02-05T09:42:18.194Z",1027 "last_posted_at": "2025-02-06T06:48:34.494Z",1028 "bumped": true,1029 "bumped_at": "2025-02-06T06:48:34.494Z",1030 "archetype": "regular",1031 "unseen": false,1032 "pinned": false,1033 "unpinned": null,1034 "visible": true,1035 "closed": false,1036 "archived": false,1037 "bookmarked": null,1038 "liked": null,1039 "tags_descriptions": {},1040 "like_count": 1,1041 "views": 58,1042 "category_id": 1,1043 "featured_link": null,1044 "has_accepted_answer": false,1045 "posters": [1046 {1047 "extras": null,1048 "description": "Original Poster",1049 "user": {1050 "id": 67255,1051 "username": "Mystify2823",1052 "name": "",1053 "avatar_template": "/user_avatar/discuss.pytorch.org/mystify2823/{size}/61564_2.png",1054 "trust_level": 11055 }1056 },1057 {1058 "extras": "latest",1059 "description": "Most Recent Poster",1060 "user": {1061 "id": 19553,1062 "username": "anantguptadbl",1063 "name": "Anant Gupta",1064 "avatar_template": "/user_avatar/discuss.pytorch.org/anantguptadbl/{size}/17784_2.png",1065 "trust_level": 21066 }1067 }1068 ]1069 },1070 {1071 "fancy_title": "Built-in 0-1 loss function?",1072 "id": 216747,1073 "title": "Built-in 0-1 loss function?",1074 "slug": "built-in-0-1-loss-function",1075 "posts_count": 3,1076 "reply_count": 0,1077 "highest_post_number": 3,1078 "image_url": null,1079 "created_at": "2025-02-16T16:14:21.754Z",1080 "last_posted_at": "2025-02-17T23:41:00.716Z",1081 "bumped": true,1082 "bumped_at": "2025-02-17T23:41:00.716Z",1083 "archetype": "regular",1084 "unseen": false,1085 "pinned": false,1086 "unpinned": null,1087 "visible": true,1088 "closed": false,1089 "archived": false,1090 "bookmarked": null,1091 "liked": null,1092 "tags_descriptions": {},1093 "like_count": 0,1094 "views": 156,1095 "category_id": 1,1096 "featured_link": null,1097 "has_accepted_answer": true,1098 "posters": [1099 {1100 "extras": "latest",1101 "description": "Original Poster, Most Recent Poster",1102 "user": {1103 "id": 75981,1104 "username": "Geremia",1105 "name": "Geremia",1106 "avatar_template": "/user_avatar/discuss.pytorch.org/geremia/{size}/70165_2.png",1107 "trust_level": 21108 }1109 },1110 {1111 "extras": null,1112 "description": "Frequent Poster, Accepted Answer",1113 "user": {1114 "id": 18088,1115 "username": "KFrank",1116 "name": "K. Frank",1117 "avatar_template": "/letter_avatar_proxy/v4/letter/k/ecb155/{size}.png",1118 "trust_level": 21119 }1120 }1121 ]1122 },1123 {1124 "fancy_title": "Is it possible to use `parametrizations.orthogonal` inside a parametrization?",1125 "id": 217203,1126 "title": "Is it possible to use `parametrizations.orthogonal` inside a parametrization?",1127 "slug": "is-it-possible-to-use-parametrizations-orthogonal-inside-a-parametrization",1128 "posts_count": 2,1129 "reply_count": 0,1130 "highest_post_number": 2,1131 "image_url": null,1132 "created_at": "2025-02-26T20:27:20.005Z",1133 "last_posted_at": "2025-02-27T02:25:07.644Z",1134 "bumped": true,1135 "bumped_at": "2025-02-27T02:25:07.644Z",1136 "archetype": "regular",1137 "unseen": false,1138 "pinned": false,1139 "unpinned": null,1140 "visible": true,1141 "closed": false,1142 "archived": false,1143 "bookmarked": null,1144 "liked": null,1145 "tags_descriptions": {},1146 "like_count": 0,1147 "views": 69,1148 "category_id": 1,1149 "featured_link": null,1150 "has_accepted_answer": false,1151 "posters": [1152 {1153 "extras": "latest single",1154 "description": "Original Poster, Most Recent Poster",1155 "user": {1156 "id": 67552,1157 "username": "dherrera1911",1158 "name": "Daniel Herrera",1159 "avatar_template": "/user_avatar/discuss.pytorch.org/dherrera1911/{size}/61905_2.png",1160 "trust_level": 11161 }1162 }1163 ]1164 },1165 {1166 "fancy_title": "Get int8 onnx format model",1167 "id": 213636,1168 "title": "Get int8 onnx format model",1169 "slug": "get-int8-onnx-format-model",1170 "posts_count": 3,1171 "reply_count": 1,1172 "highest_post_number": 3,1173 "image_url": null,1174 "created_at": "2024-11-30T14:00:20.834Z",1175 "last_posted_at": "2024-12-02T05:43:39.388Z",1176 "bumped": true,1177 "bumped_at": "2024-12-02T05:43:39.388Z",1178 "archetype": "regular",1179 "unseen": false,1180 "pinned": false,1181 "unpinned": null,1182 "visible": true,1183 "closed": false,1184 "archived": false,1185 "bookmarked": null,1186 "liked": null,1187 "tags_descriptions": {},1188 "like_count": 1,1189 "views": 227,1190 "category_id": 1,1191 "featured_link": null,1192 "has_accepted_answer": false,1193 "posters": [1194 {1195 "extras": "latest single",1196 "description": "Original Poster, Most Recent Poster",1197 "user": {1198 "id": 80928,1199 "username": "xiangyun",1200 "name": "",