{"users":[{"id":8066,"username":"Bhack","name":"Stefano Fabri","avatar_template":"/user_avatar/discuss.ai.google.dev/bhack/{size}/6388_2.png","trust_level":3},{"id":11779,"username":"mihaimaruseac","name":"Mihai Maruseac","avatar_template":"/user_avatar/discuss.ai.google.dev/mihaimaruseac/{size}/6253_2.png","primary_group_name":"tfteam","flair_name":"google","flair_url":"https://d3qe71uytubmmx.cloudfront.net/original/2X/c/c44b315468d82ff8c59c3dd831508fccc157ad2b.avif","flair_group_id":41,"trust_level":4},{"id":355,"username":"Mark_Daoust","name":"Mark Daoust","avatar_template":"/user_avatar/discuss.ai.google.dev/mark_daoust/{size}/411_2.png","trust_level":2},{"id":15077,"username":"Jacques_Pienaar","name":"Jacques Pienaar","avatar_template":"/user_avatar/discuss.ai.google.dev/jacques_pienaar/{size}/6539_2.png","primary_group_name":"tfteam","trust_level":4},{"id":9967,"username":"innat","name":"M.Innat","avatar_template":"/user_avatar/discuss.ai.google.dev/innat/{size}/13542_2.png","trust_level":2},{"id":9953,"username":"Charlie_Barber","name":"Charlie Barber","avatar_template":"/user_avatar/discuss.ai.google.dev/charlie_barber/{size}/14928_2.png","trust_level":0},{"id":40757,"username":"Divya_Sree_Kayyuri","name":"Divya Sree Kayyuri","avatar_template":"/user_avatar/discuss.ai.google.dev/divya_sree_kayyuri/{size}/29791_2.png","moderator":true,"trust_level":3},{"id":20482,"username":"Jason_Xie","name":"Jason Xie","avatar_template":"/user_avatar/discuss.ai.google.dev/jason_xie/{size}/24458_2.png","trust_level":1},{"id":15803,"username":"Kiran_Sai_Ramineni","name":"Kiran Sai","avatar_template":"/user_avatar/discuss.ai.google.dev/kiran_sai_ramineni/{size}/8859_2.png","moderator":true,"trust_level":3},{"id":143250,"username":"Mohammed_Dhilshad","name":"Mohammed Dhilshad","avatar_template":"/user_avatar/discuss.ai.google.dev/mohammed_dhilshad/{size}/31947_2.png","trust_level":0},{"id":134178,"username":"Alvis1","name":"Alvis ","avatar_template":"/user_avatar/discuss.ai.google.dev/alvis1/{size}/31627_2.png","trust_level":0},{"id":95688,"username":"Brian_Provost","name":"Brian Provost","avatar_template":"/user_avatar/discuss.ai.google.dev/brian_provost/{size}/90212_2.png","trust_level":1},{"id":133643,"username":"Kavin_Sai","name":"Kavin Sai","avatar_template":"/user_avatar/discuss.ai.google.dev/kavin_sai/{size}/35785_2.png","trust_level":0},{"id":127963,"username":"Graham_Smith","name":"Graham Smith","avatar_template":"/user_avatar/discuss.ai.google.dev/graham_smith/{size}/119439_2.png","trust_level":0},{"id":92611,"username":"Joe_Smith3","name":"","avatar_template":"/user_avatar/discuss.ai.google.dev/joe_smith3/{size}/87554_2.png","trust_level":0},{"id":127606,"username":"BASU_PATIL","name":"BASU PATIL","avatar_template":"/user_avatar/discuss.ai.google.dev/basu_patil/{size}/119167_2.png","trust_level":0},{"id":41479,"username":"l1234567j222","name":"","avatar_template":"/user_avatar/discuss.ai.google.dev/l1234567j222/{size}/45518_2.png","trust_level":1},{"id":121819,"username":"brokencog","name":"Daniel","avatar_template":"/user_avatar/discuss.ai.google.dev/brokencog/{size}/114099_2.png","trust_level":0},{"id":11553,"username":"Rafael_Scatena","name":"Rafael Scatena","avatar_template":"/user_avatar/discuss.ai.google.dev/rafael_scatena/{size}/16875_2.png","trust_level":1},{"id":116548,"username":"DefinitlyNotAModel","name":"Heiko ","avatar_template":"/user_avatar/discuss.ai.google.dev/definitlynotamodel/{size}/24344_2.png","trust_level":0},{"id":11865,"username":"Quan_Nie","name":"Quan Nie","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/q/977dab/{size}.png","trust_level":0},{"id":15753,"username":"Brempong_Appiah_Dank","name":"Brempong Appiah Dankwah","avatar_template":"/user_avatar/discuss.ai.google.dev/brempong_appiah_dank/{size}/17878_2.png","trust_level":1},{"id":105023,"username":"satadru","name":"Satadru Bag","avatar_template":"/user_avatar/discuss.ai.google.dev/satadru/{size}/98963_2.png","trust_level":0},{"id":13722,"username":"Mahrukh","name":"mahrukhS","avatar_template":"/user_avatar/discuss.ai.google.dev/mahrukh/{size}/17694_2.png","trust_level":0},{"id":9470,"username":"Gayathri","name":"Ninja","avatar_template":"/user_avatar/discuss.ai.google.dev/gayathri/{size}/6846_2.png","trust_level":0},{"id":10667,"username":"IUSR","name":"Jerome J","avatar_template":"/user_avatar/discuss.ai.google.dev/iusr/{size}/9822_2.png","trust_level":0},{"id":112327,"username":"M_Husnain","name":"M Husnain ","avatar_template":"/user_avatar/discuss.ai.google.dev/m_husnain/{size}/32147_2.png","trust_level":0},{"id":36485,"username":"DriaanJ","name":"Driaan Jansen","avatar_template":"/user_avatar/discuss.ai.google.dev/driaanj/{size}/31251_2.png","trust_level":0},{"id":38009,"username":"Vadym_Matsishevskyi","name":"Vadym Matsishevskyi","avatar_template":"/user_avatar/discuss.ai.google.dev/vadym_matsishevskyi/{size}/42428_2.png","trust_level":3},{"id":7131,"username":"affectsai","name":"Timothy Sweeney-Fanelli","avatar_template":"/user_avatar/discuss.ai.google.dev/affectsai/{size}/8894_2.png","trust_level":0},{"id":113113,"username":"ChivitiH","name":"","avatar_template":"/user_avatar/discuss.ai.google.dev/chivitih/{size}/38051_2.png","trust_level":0},{"id":21882,"username":"Pascal567","name":"","avatar_template":"/user_avatar/discuss.ai.google.dev/pascal567/{size}/24580_2.png","trust_level":0},{"id":19057,"username":"dencoder","name":"","avatar_template":"/user_avatar/discuss.ai.google.dev/dencoder/{size}/22556_2.png","trust_level":0},{"id":10426,"username":"Peter_Bleackley","name":"Peter Bleackley","avatar_template":"https://avatars.discourse-cdn.com/v4/letter/p/898d66/{size}.png","trust_level":1},{"id":15098,"username":"Bah","name":"Bahram","avatar_template":"/user_avatar/discuss.ai.google.dev/bah/{size}/16701_2.png","trust_level":0},{"id":20237,"username":"Sai_Rishik","name":"Sai Rishik","avatar_template":"/user_avatar/discuss.ai.google.dev/sai_rishik/{size}/24038_2.png","trust_level":1},{"id":14615,"username":"Zsolt_Szakaly","name":"Zsolt Szakaly","avatar_template":"/user_avatar/discuss.ai.google.dev/zsolt_szakaly/{size}/9255_2.png","trust_level":1},{"id":58312,"username":"Monty_Hall","name":"Monty Hall","avatar_template":"/user_avatar/discuss.ai.google.dev/monty_hall/{size}/31969_2.png","trust_level":0},{"id":108282,"username":"kishor","name":"K Kr. T","avatar_template":"/user_avatar/discuss.ai.google.dev/kishor/{size}/32069_2.png","trust_level":0},{"id":78081,"username":"Brian_Parbhu","name":"Brian Parbhu","avatar_template":"/user_avatar/discuss.ai.google.dev/brian_parbhu/{size}/75198_2.png","trust_level":0},{"id":10804,"username":"B1t_F3n1x","name":"","avatar_template":"/user_avatar/discuss.ai.google.dev/b1t_f3n1x/{size}/16123_2.png","trust_level":0},{"id":14788,"username":"Harry_Tye","name":"Harry Tye","avatar_template":"/user_avatar/discuss.ai.google.dev/harry_tye/{size}/14604_2.png","trust_level":0}],"primary_groups":[{"id":49,"name":"tfteam"}],"flair_groups":[{"id":41,"name":"google","flair_url":"https://d3qe71uytubmmx.cloudfront.net/original/2X/c/c44b315468d82ff8c59c3dd831508fccc157ad2b.avif","flair_bg_color":"","flair_color":""}],"topic_list":{"can_create_topic":false,"more_topics_url":"/c/tensorflow/21?page=1","per_page":30,"top_tags":[{"id":318,"name":"help_request","slug":"318-tag"},{"id":20,"name":"models","slug":"models"},{"id":137,"name":"keras","slug":"keras"},{"id":218,"name":"tflite","slug":"tflite"},{"id":52,"name":"datasets","slug":"datasets"},{"id":269,"name":"gpu","slug":"gpu"},{"id":362,"name":"tensorflow","slug":"tensorflow"},{"id":31,"name":"tfjs","slug":"tfjs"},{"id":306,"name":"education","slug":"education"},{"id":8,"name":"api","slug":"api"},{"id":196,"name":"install","slug":"install"},{"id":212,"name":"tfdata","slug":"tfdata"},{"id":353,"name":"tfcore","slug":"tfcore"},{"id":278,"name":"learning","slug":"learning"},{"id":126,"name":"android","slug":"android"},{"id":271,"name":"model_garden","slug":"model-garden"},{"id":292,"name":"model","slug":"model"},{"id":210,"name":"python","slug":"python"},{"id":242,"name":"model_maker","slug":"242-tag"},{"id":29,"name":"tfx","slug":"tfx"},{"id":299,"name":"tfhub","slug":"tfhub"},{"id":46,"name":"nlp","slug":"nlp"},{"id":273,"name":"tflite_micro","slug":"tflite-micro"},{"id":16,"name":"build","slug":"build"},{"id":268,"name":"getting_started","slug":"268-tag"},{"id":91,"name":"object-detection","slug":"object-detection"},{"id":12,"name":"docs","slug":"docs"},{"id":96,"name":"distributed-training","slug":"distributed-training"},{"id":209,"name":"tfkeras","slug":"tfkeras"},{"id":287,"name":"model_optimization","slug":"287-tag"},{"id":313,"name":"training","slug":"training"}],"topics":[{"fancy_title":"Deformable convolution and other custom ops","fancy_title_localized":false,"locale":null,"id":32546,"title":"Deformable convolution and other custom ops","slug":"deformable-convolution-and-other-custom-ops","posts_count":28,"reply_count":17,"highest_post_number":42,"image_url":null,"created_at":"2021-06-14T11:36:08.165Z","last_posted_at":"2026-07-11T22:07:24.611Z","bumped":true,"bumped_at":"2026-07-11T22:07:24.611Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":10,"name":"contributing","slug":"contributing"},{"id":274,"name":"xla","slug":"xla"},{"id":240,"name":"help_dev","slug":"help-dev"},{"id":249,"name":"sig_addons","slug":"sig-addons"},{"id":158,"name":"sig_keras","slug":"sig-keras"}],"tags_descriptions":{},"views":6331,"like_count":7,"has_summary":false,"last_poster_username":"innat","category_id":57,"op_like_count":1,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Users are seeking updates on the integration of deformable convolution in TensorFlow Addons, highlighting ongoing challenges regarding the maintenance overhead and hardware support of custom C++ operations.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":8066,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":11779,"primary_group_id":49,"flair_group_id":41},{"extras":null,"description":"Frequent Poster","user_id":355,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":15077,"primary_group_id":49,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":9967,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Tensorflow And Ubuntu 26","fancy_title_localized":false,"locale":"en","id":171573,"title":"Tensorflow And Ubuntu 26","slug":"tensorflow-and-ubuntu-26","posts_count":2,"reply_count":0,"highest_post_number":2,"image_url":null,"created_at":"2026-06-17T05:42:36.358Z","last_posted_at":"2026-07-10T16:02:34.974Z","bumped":true,"bumped_at":"2026-07-10T16:02:34.974Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":332,"name":"ubuntu","slug":"ubuntu"}],"tags_descriptions":{},"views":61,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":21,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A user experiencing difficulties installing TensorFlow on Ubuntu was advised to upgrade pip within a virtual environment and provide error messages if the installation continues to fail.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":9953,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Face Detector in MediaPipe model maker","fancy_title_localized":false,"locale":null,"id":39959,"title":"Face Detector in MediaPipe model maker","slug":"face-detector-in-mediapipe-model-maker","posts_count":6,"reply_count":2,"highest_post_number":11,"image_url":null,"created_at":"2024-09-21T16:20:01.359Z","last_posted_at":"2026-06-15T07:47:51.916Z","bumped":true,"bumped_at":"2026-06-15T07:47:51.916Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":242,"name":"model_maker","slug":"242-tag"},{"id":189,"name":"mediapipe","slug":"mediapipe"},{"id":154,"name":"roadmap","slug":"roadmap"}],"tags_descriptions":{},"views":163,"like_count":0,"has_summary":false,"last_poster_username":"Mohammed_Dhilshad","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Following inquiries about face detection support in MediaPipe, users are questioning the legality of retraining and commercially using a recreated BlazeFace model with Google's pre-trained weights.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":20482,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":15803,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":143250,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"What is the maximum model size to run on Google Coral Dev Board Micro? or How to run FullSubNet on Coral mirco?","fancy_title_localized":false,"locale":"en","id":147283,"title":"What is the maximum model size to run on Google Coral Dev Board Micro? or How to run FullSubNet on Coral mirco?","slug":"what-is-the-maximum-model-size-to-run-on-google-coral-dev-board-micro-or-how-to-run-fullsubnet-on-coral-mirco","posts_count":1,"reply_count":0,"highest_post_number":1,"image_url":null,"created_at":"2026-05-22T08:18:29.486Z","last_posted_at":"2026-05-22T08:18:29.577Z","bumped":true,"bumped_at":"2026-05-22T08:18:29.577Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":20,"name":"models","slug":"models"}],"tags_descriptions":{},"views":79,"like_count":0,"has_summary":false,"last_poster_username":"Alvis1","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A user is encountering an issue running a 6MB int8 FullSubNet model on the Coral Dev Board Micro, which gets stuck, despite successfully running other converted models like MobileNet and YAMNet.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest single","description":"Original Poster, Most Recent Poster","user_id":134178,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Mixed rounding behavior in quantized model using TFL Interpreter","fancy_title_localized":false,"locale":null,"id":138662,"title":"Mixed rounding behavior in quantized model using TFL Interpreter","slug":"mixed-rounding-behavior-in-quantized-model-using-tfl-interpreter","posts_count":3,"reply_count":1,"highest_post_number":3,"image_url":null,"created_at":"2026-04-08T17:55:38.951Z","last_posted_at":"2026-05-21T17:58:17.181Z","bumped":true,"bumped_at":"2026-05-21T17:58:17.181Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":8,"name":"api","slug":"api"},{"id":131,"name":"tf-lite","slug":"tf-lite"}],"tags_descriptions":{},"views":72,"like_count":0,"has_summary":false,"last_poster_username":"Brian_Provost","category_id":21,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Unable to enforce a single global rounding mode in the TFLite interpreter for mixed rounding behavior, Brian_Provost modified CMSIS-NN source code per layer type to match TensorFlow's specific rounding algorithms.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest","description":"Original Poster, Most Recent Poster","user_id":95688,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":133643,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Import tensorflow &ldquo;Could not find cuda drivers on your machine, GPU will not be used&rdquo;","fancy_title_localized":false,"locale":null,"id":143227,"title":"Import tensorflow \"Could not find cuda drivers on your machine, GPU will not be used\"","slug":"import-tensorflow-could-not-find-cuda-drivers-on-your-machine-gpu-will-not-be-used","posts_count":4,"reply_count":2,"highest_post_number":4,"image_url":null,"created_at":"2026-05-02T20:23:13.210Z","last_posted_at":"2026-05-05T09:12:55.369Z","bumped":true,"bumped_at":"2026-05-05T09:12:55.369Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":25,"name":"core","slug":"core"}],"tags_descriptions":{},"views":132,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":21,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"TensorFlow fails to find CUDA drivers due to Python 3.14's incompatibility with TensorFlow 2.21, with a recommendation to use Python 3.12 or 3.13 for compatible GPU utilization.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":127963,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Can tensorflow lite object detection model be trained with AMD grpahics card","fancy_title_localized":false,"locale":null,"id":142512,"title":"Can tensorflow lite object detection model be trained with AMD grpahics card","slug":"can-tensorflow-lite-object-detection-model-be-trained-with-amd-grpahics-card","posts_count":2,"reply_count":0,"highest_post_number":2,"image_url":null,"created_at":"2026-04-27T14:36:26.582Z","last_posted_at":"2026-05-03T18:17:19.839Z","bumped":true,"bumped_at":"2026-05-03T18:17:19.839Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":385,"name":"tensorflow-lite","slug":"tensorflow-lite"}],"tags_descriptions":{},"views":48,"like_count":2,"has_summary":false,"last_poster_username":"BASU_PATIL","category_id":21,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Training TensorFlow Lite object detection models with an AMD graphics card is supported, but only through ROCm-enabled TensorFlow.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":92611,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":127606,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Getting 0 nodes delegated while using TFLite C Library with TFLite Flex Delegate","fancy_title_localized":false,"locale":null,"id":72789,"title":"Getting 0 nodes delegated while using TFLite C Library with TFLite Flex Delegate","slug":"getting-0-nodes-delegated-while-using-tflite-c-library-with-tflite-flex-delegate","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":"https://d3qe71uytubmmx.cloudfront.net/original/3X/d/1/d14f032bc36087adbfc992b3b016447a0b6f33e9.png","created_at":"2025-03-17T08:30:51.152Z","last_posted_at":"2026-04-23T03:27:47.278Z","bumped":true,"bumped_at":"2026-04-23T03:27:47.278Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":218,"name":"tflite","slug":"tflite"}],"tags_descriptions":{},"views":138,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":21,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"The issue of TFLite Flex Delegate not delegating `tf.unpack` operations, resulting in 0 delegated nodes, may be resolved by updating to the latest versions as suggested in a GitHub issue.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":41479,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Does not contain symbol &lsquo;cc_autoconf_impl&rsquo;","fancy_title_localized":false,"locale":null,"id":139439,"title":"Does not contain symbol 'cc_autoconf_impl'","slug":"does-not-contain-symbol-cc-autoconf-impl","posts_count":3,"reply_count":1,"highest_post_number":3,"image_url":null,"created_at":"2026-04-13T04:23:17.573Z","last_posted_at":"2026-04-20T14:43:56.539Z","bumped":true,"bumped_at":"2026-04-20T14:43:56.539Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":354,"name":"tf-install","slug":"tf-install"}],"tags_descriptions":{},"views":53,"like_count":0,"has_summary":false,"last_poster_username":"brokencog","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"User provided CUDA compiler release 13.2 and driver version 595.58.03 details to assist in troubleshooting TensorFlow's GPU detection failure after both pip installation and a local build attempt.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest","description":"Original Poster, Most Recent Poster","user_id":121819,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Regression in a dataset with flattened images and target reals","fancy_title_localized":false,"locale":null,"id":28301,"title":"Regression in a dataset with flattened images and target reals","slug":"regression-in-a-dataset-with-flattened-images-and-target-reals","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2024-05-30T22:05:33.164Z","last_posted_at":"2026-04-13T12:03:27.929Z","bumped":true,"bumped_at":"2026-04-13T12:03:27.929Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":52,"name":"datasets","slug":"datasets"},{"id":181,"name":"tfimage","slug":"tfimage"}],"tags_descriptions":{},"views":160,"like_count":1,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":37,"op_like_count":1,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A user is encountering problems setting up a regression dataset with flattened lung images and real targets, with the latest advice suggesting verifying image file paths and CSV format as the cause of the error.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":11553,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Memory leak with training different models in loops","fancy_title_localized":false,"locale":null,"id":136413,"title":"Memory leak with training different models in loops","slug":"memory-leak-with-training-different-models-in-loops","posts_count":4,"reply_count":0,"highest_post_number":4,"image_url":null,"created_at":"2026-03-27T08:45:06.141Z","last_posted_at":"2026-04-07T07:47:13.949Z","bumped":true,"bumped_at":"2026-04-07T07:47:13.949Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":20,"name":"models","slug":"models"},{"id":7,"name":"bug","slug":"bug"},{"id":52,"name":"datasets","slug":"datasets"}],"tags_descriptions":{},"views":182,"like_count":1,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A progressive memory growth issue during TensorFlow model training loops, attributed to unreleased C++ caches, is resolved by wrapping each model's training in a separate process to ensure memory is freed.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":116548,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Why tensorflow switches to clang suddenly?","fancy_title_localized":false,"locale":null,"id":31149,"title":"Why tensorflow switches to clang suddenly?","slug":"why-tensorflow-switches-to-clang-suddenly","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2023-09-06T06:57:01.426Z","last_posted_at":"2026-04-02T12:51:20.549Z","bumped":true,"bumped_at":"2026-04-02T12:51:20.549Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":318,"name":"help_request","slug":"318-tag"}],"tags_descriptions":{},"views":337,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"TensorFlow switched to Clang as the default compiler for OSS Linux CPU and GPU builds to improve framework efficiency and reduce bugs.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":11865,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Error running tuner component in tfx pipeline","fancy_title_localized":false,"locale":null,"id":27711,"title":"Error running tuner component in tfx pipeline","slug":"error-running-tuner-component-in-tfx-pipeline","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2024-05-27T04:25:08.901Z","last_posted_at":"2026-04-02T08:24:07.002Z","bumped":true,"bumped_at":"2026-04-02T08:24:07.002Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":29,"name":"tfx","slug":"tfx"},{"id":51,"name":"pipelines","slug":"pipelines"}],"tags_descriptions":{},"views":137,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"The TFX Tuner component is encountering a TypeError due to a None value, with the recommended solution being to explicitly set `max_trials` and share reproducible code for debugging.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":15753,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Transformer for asynchronous multi-stream image time-series with online prediction?","fancy_title_localized":false,"locale":null,"id":137033,"title":"Transformer for asynchronous multi-stream image time-series with online prediction?","slug":"transformer-for-asynchronous-multi-stream-image-time-series-with-online-prediction","posts_count":1,"reply_count":0,"highest_post_number":1,"image_url":null,"created_at":"2026-03-30T12:20:10.290Z","last_posted_at":"2026-03-30T12:20:10.369Z","bumped":true,"bumped_at":"2026-03-30T12:20:10.369Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":181,"name":"tfimage","slug":"tfimage"}],"tags_descriptions":{},"views":36,"like_count":0,"has_summary":false,"last_poster_username":"satadru","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"An architecture is sought for online classification of asynchronous multi-stream image time-series, proposing a ViT encoder and causal transformer with real-valued timestamp embeddings to handle irregular, unaligned data for improved causal predictions.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest single","description":"Original Poster, Most Recent Poster","user_id":105023,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Issues Running Sanity Check","fancy_title_localized":false,"locale":null,"id":25265,"title":"Issues Running Sanity Check","slug":"issues-running-sanity-check","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2024-01-08T06:15:44.260Z","last_posted_at":"2026-03-26T15:27:02.162Z","bumped":true,"bumped_at":"2026-03-26T15:27:02.162Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":10,"name":"contributing","slug":"contributing"}],"tags_descriptions":{},"views":333,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Mahrukh encountered an issue with sanity check opening `ci_build.sh` in VS Code, and the latest advice suggests running the command via WSL or Git Bash by prepending 'bash'.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":13722,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Unknown error: Graph execution error","fancy_title_localized":false,"locale":null,"id":32215,"title":"Unknown error: Graph execution error","slug":"unknown-error-graph-execution-error","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2024-04-10T03:24:30.139Z","last_posted_at":"2026-03-26T12:43:08.517Z","bumped":true,"bumped_at":"2026-03-26T12:43:08.517Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":348,"name":"model-training","slug":"model-training"},{"id":205,"name":"epoc","slug":"epoc"}],"tags_descriptions":{},"views":245,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":21,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A user encountered a \"Graph execution error\" (InvalidArgumentError) during model training, which is attributed to a shape mismatch requiring proper resizing; further debugging requires an accessible Colab link or reproducible code.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":9470,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Missing Python wheel package building from source","fancy_title_localized":false,"locale":null,"id":32204,"title":"Missing Python wheel package building from source","slug":"missing-python-wheel-package-building-from-source","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2024-04-08T10:23:53.596Z","last_posted_at":"2026-03-25T09:01:36.742Z","bumped":true,"bumped_at":"2026-03-25T09:01:36.742Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":362,"name":"tensorflow","slug":"tensorflow"},{"id":210,"name":"python","slug":"python"},{"id":16,"name":"build","slug":"build"}],"tags_descriptions":{},"views":210,"like_count":1,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"IUSR inquired about missing Python wheel package building from source for TensorFlow, and Divya_Sree_Kayyuri confirmed that manually running `build_pip_package` is the correct procedure as per the official documentation.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":10667,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Request for Review: Deterministic Sorting Fix in auto_control_deps_utils.py (PR #110822)","fancy_title_localized":false,"locale":null,"id":133472,"title":"Request for Review: Deterministic Sorting Fix in auto_control_deps_utils.py (PR #110822)","slug":"request-for-review-deterministic-sorting-fix-in-auto-control-deps-utils-py-pr-110822","posts_count":3,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2026-03-17T04:27:57.276Z","last_posted_at":"2026-03-19T18:36:00.911Z","bumped":true,"bumped_at":"2026-03-19T18:36:00.911Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":210,"name":"python","slug":"python"},{"id":25,"name":"core","slug":"core"}],"tags_descriptions":{},"views":68,"like_count":0,"has_summary":false,"last_poster_username":"M_Husnain","category_id":21,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"A fix for non-deterministic resource categorization in TensorFlow (PR #110822) is under review after a new reviewer was assigned, and the contributor is awaiting feedback after implementing final fixes and passing all tests.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest","description":"Original Poster, Most Recent Poster","user_id":112327,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Building tensorflow from source for RTX5000 GPU series","fancy_title_localized":false,"locale":null,"id":65171,"title":"Building tensorflow from source for RTX5000 GPU series","slug":"building-tensorflow-from-source-for-rtx5000-gpu-series","posts_count":9,"reply_count":6,"highest_post_number":17,"image_url":null,"created_at":"2025-02-09T07:14:11.998Z","last_posted_at":"2026-03-18T12:54:01.519Z","bumped":true,"bumped_at":"2026-03-18T12:54:01.519Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":16,"name":"build","slug":"build"}],"tags_descriptions":{},"views":2760,"like_count":0,"has_summary":false,"last_poster_username":"ChivitiH","category_id":21,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"TensorFlow 2.22 has been successfully built with native sm_120 support using Clang 20 and four source patches, achieving significant performance improvements, resolving the CUDA_ERROR_INVALID_PTX issue, and offering a pre-built wheel for Python 3.13.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":36485,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":38009,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":7131,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":113113,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Missing Method under c#","fancy_title_localized":false,"locale":null,"id":42435,"title":"Missing Method under c#","slug":"missing-method-under-c","posts_count":2,"reply_count":0,"highest_post_number":4,"image_url":null,"created_at":"2024-10-09T11:12:52.640Z","last_posted_at":"2026-03-17T06:36:31.723Z","bumped":true,"bumped_at":"2026-03-17T06:36:31.723Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":386,"name":"c","slug":"c"}],"tags_descriptions":{},"views":105,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Pascal567 seeks help with KerasNet in C# due to the missing `reset_states()` method for an LSTM model, and Divya_Sree_Kayyuri suggests managing state by creating a session and reinitializing the model.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":21882,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Getting Value Error while running the model prediction when it is deployed to Google Cloud Run","fancy_title_localized":false,"locale":null,"id":37487,"title":"Getting Value Error while running the model prediction when it is deployed to Google Cloud Run","slug":"getting-value-error-while-running-the-model-prediction-when-it-is-deployed-to-google-cloud-run","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2024-08-27T20:07:55.224Z","last_posted_at":"2026-03-16T10:43:51.926Z","bumped":true,"bumped_at":"2026-03-16T10:43:51.926Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":210,"name":"python","slug":"python"},{"id":291,"name":"google-cloud","slug":"google-cloud"},{"id":179,"name":"community_spotlight","slug":"179-tag"}],"tags_descriptions":{},"views":92,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":21,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Dencoder is encountering a ValueError during model prediction on Google Cloud Run, despite the code working locally, and Divya_Sree_Kayyuri suggests checking input tensor shapes and dependency conflicts.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":19057,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Buidling a network from a binary parse tree","fancy_title_localized":false,"locale":null,"id":29395,"title":"Buidling a network from a binary parse tree","slug":"buidling-a-network-from-a-binary-parse-tree","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2023-01-03T13:40:12.704Z","last_posted_at":"2026-03-13T07:53:21.287Z","bumped":true,"bumped_at":"2026-03-13T07:53:21.287Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":137,"name":"keras","slug":"keras"},{"id":46,"name":"nlp","slug":"nlp"}],"tags_descriptions":{},"views":512,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Peter_Bleackley is building a network from a binary parse tree using the Stanford Natural Language Inference Corpus, and Divya_Sree_Kayyuri suggests replacing `tf.raw_ops.Placeholder` with `tf.keras.layers.Input` to fix a related error.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":10426,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"TensorFlow Error. With Rust Failed to start training: InvalidArgument:","fancy_title_localized":false,"locale":null,"id":27782,"title":"TensorFlow Error. With Rust Failed to start training: InvalidArgument:","slug":"tensorflow-error-with-rust-failed-to-start-training-invalidargument","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2024-06-10T02:33:50.978Z","last_posted_at":"2026-03-12T17:33:09.376Z","bumped":true,"bumped_at":"2026-03-12T17:33:09.376Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":173,"name":"model-code","slug":"model-code"},{"id":99,"name":"tfgradient","slug":"tfgradient"}],"tags_descriptions":{},"views":94,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Bah is encountering an \"InvalidArgument\" error in Rust with TensorFlow when using ApplyGradientDescent, which Divya_Sree_Kayyuri clarified is due to providing only one input instead of the expected three, requiring a variable reference as the first input.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":15098,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Instance of flex delegate","fancy_title_localized":false,"locale":null,"id":46344,"title":"Instance of flex delegate","slug":"instance-of-flex-delegate","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2024-10-27T15:39:07.233Z","last_posted_at":"2026-03-12T04:15:44.607Z","bumped":true,"bumped_at":"2026-03-12T04:15:44.607Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":76,"name":"tflite-support","slug":"tflite-support"}],"tags_descriptions":{},"views":69,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":21,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":null,"has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":20237,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"TF 2.20 C API (libtensorflow.so)","fancy_title_localized":false,"locale":null,"id":128856,"title":"TF 2.20 C API (libtensorflow.so)","slug":"tf-2-20-c-api-libtensorflow-so","posts_count":3,"reply_count":1,"highest_post_number":3,"image_url":null,"created_at":"2026-03-05T10:11:11.437Z","last_posted_at":"2026-03-10T11:43:57.230Z","bumped":true,"bumped_at":"2026-03-10T11:43:57.230Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":212,"name":"tfdata","slug":"tfdata"},{"id":16,"name":"build","slug":"build"}],"tags_descriptions":{},"views":86,"like_count":3,"has_summary":false,"last_poster_username":"Zsolt_Szakaly","category_id":37,"op_like_count":2,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Zsolt_Szakaly seeks a reliable method to obtain or build the libtensorflow.so binary for TF 2.20, after facing challenges and random errors with older versions and failing to build from source despite following Gemini's suggestions.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest","description":"Original Poster, Most Recent Poster","user_id":14615,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Anybody get TF2 to build for RTX 6000 blackwell? (cuda 12.8 and cudnn 9.1 - forget tensor rt)","fancy_title_localized":false,"locale":null,"id":95930,"title":"Anybody get TF2 to build for RTX 6000 blackwell? (cuda 12.8 and cudnn 9.1 - forget tensor rt)","slug":"anybody-get-tf2-to-build-for-rtx-6000-blackwell-cuda-12-8-and-cudnn-9-1-forget-tensor-rt","posts_count":4,"reply_count":1,"highest_post_number":6,"image_url":null,"created_at":"2025-07-29T02:59:19.173Z","last_posted_at":"2026-03-09T04:52:32.206Z","bumped":true,"bumped_at":"2026-03-09T04:52:32.206Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":362,"name":"tensorflow","slug":"tensorflow"},{"id":104,"name":"cuda","slug":"cuda"}],"tags_descriptions":{},"views":311,"like_count":1,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":21,"op_like_count":1,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Users are encountering issues with TensorFlow detecting RTX 6000 Blackwell GPUs, even after following official instructions and trying stable and nightly versions, prompting suggestions to explore related GitHub issues.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":58312,"primary_group_id":null,"flair_group_id":null},{"extras":null,"description":"Frequent Poster","user_id":108282,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"ONNX Probabilistic Programming Working Group – Collaboration with the TensorFlow Probability Community","fancy_title_localized":false,"locale":null,"id":129528,"title":"ONNX Probabilistic Programming Working Group – Collaboration with the TensorFlow Probability Community","slug":"onnx-probabilistic-programming-working-group-collaboration-with-the-tensorflow-probability-community","posts_count":1,"reply_count":0,"highest_post_number":1,"image_url":null,"created_at":"2026-03-07T20:19:37.018Z","last_posted_at":"2026-03-07T20:19:37.122Z","bumped":true,"bumped_at":"2026-03-07T20:19:37.122Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":330,"name":"tf-probability","slug":"tf-probability"}],"tags_descriptions":{},"views":51,"like_count":0,"has_summary":false,"last_poster_username":"Brian_Parbhu","category_id":30,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"The ONNX Probabilistic Programming Working Group invites the TensorFlow Probability community to collaborate on bringing probabilistic modeling and Bayesian inference into the ONNX ecosystem, aiming to standardize operator domain and runtime semantics.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":"latest single","description":"Original Poster, Most Recent Poster","user_id":78081,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Keras is not saving models as intended","fancy_title_localized":false,"locale":null,"id":127353,"title":"Keras is not saving models as intended","slug":"keras-is-not-saving-models-as-intended","posts_count":4,"reply_count":2,"highest_post_number":4,"image_url":"https://d3qe71uytubmmx.cloudfront.net/optimized/2X/5/5ddfd10286ce16453e5a2d362d305dcbd8ec8991_2_1024x576.png","created_at":"2026-02-27T13:37:54.469Z","last_posted_at":"2026-03-05T09:17:06.054Z","bumped":true,"bumped_at":"2026-03-05T09:17:06.054Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":20,"name":"models","slug":"models"}],"tags_descriptions":{},"views":83,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Satadru is seeking guidance on how to properly save and resume training with the same learning schedule and optimizer state in Keras, while Divya_Sree_Kayyuri suggests referring to Keras documentation for understanding model saving and checkpoints.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":105023,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Optimizer Question","fancy_title_localized":false,"locale":null,"id":28122,"title":"Optimizer Question","slug":"optimizer-question","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2024-01-11T05:44:54.024Z","last_posted_at":"2026-02-25T16:00:38.993Z","bumped":true,"bumped_at":"2026-02-25T16:00:38.993Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":278,"name":"learning","slug":"learning"}],"tags_descriptions":{},"views":293,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"B1t_F3n1x inquired about implementing the GADAM optimizer in TensorFlow, and Divya_Sree_Kayyuri confirmed it's possible by subclassing `tf.keras.optimizers.Optimizer`.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":10804,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]},{"fancy_title":"Issues building and preparing binary for SparkFun Edge","fancy_title_localized":false,"locale":null,"id":24609,"title":"Issues building and preparing binary for SparkFun Edge","slug":"issues-building-and-preparing-binary-for-sparkfun-edge","posts_count":2,"reply_count":0,"highest_post_number":3,"image_url":null,"created_at":"2024-05-14T05:24:39.180Z","last_posted_at":"2026-02-25T15:05:04.115Z","bumped":true,"bumped_at":"2026-02-25T15:05:04.115Z","archetype":"regular","unseen":false,"pinned":false,"unpinned":null,"visible":true,"closed":false,"archived":false,"bookmarked":null,"liked":null,"tags":[{"id":273,"name":"tflite_micro","slug":"tflite-micro"}],"tags_descriptions":{},"views":130,"like_count":0,"has_summary":false,"last_poster_username":"Divya_Sree_Kayyuri","category_id":37,"op_like_count":0,"pinned_globally":false,"featured_link":null,"ai_topic_gist":"Harry_Tye is encountering issues building a binary for the SparkFun Edge device following a tutorial, and Divya_Sree_Kayyuri is asking for the specific error message and operating system to help resolve the problem.","has_accepted_answer":false,"can_vote":false,"posters":[{"extras":null,"description":"Original Poster","user_id":14788,"primary_group_id":null,"flair_group_id":null},{"extras":"latest","description":"Most Recent Poster","user_id":40757,"primary_group_id":null,"flair_group_id":null}]}]}}