Commit Graph

  • 837ed5d980 Merge pull request #648 from exo-explore/modelasyncload Alex Cheema 2025-01-28 23:39:11 +00:00
  • 9c1bea97e8 fix embed_tokens for last layer in qwen models modelasyncload Alex Cheema 2025-01-28 23:09:45 +00:00
  • af171f06fa propagate prompts to other nodes so they can display them, cleaner prompt/output output Alex Cheema 2025-01-28 21:50:49 +00:00
  • edfa53a4c2 Merge pull request #646 from exo-explore/modelasyncload Alex Cheema 2025-01-28 18:56:19 +00:00
  • 4a5b80a958 make sure mlx stuff is on separate thread non blocking Alex Cheema 2025-01-28 18:56:00 +00:00
  • 92d1bc01de Merge pull request #645 from exo-explore/modelasyncload Alex Cheema 2025-01-28 18:49:47 +00:00
  • 6662d5668c load mlx model shard on mlx thread so it doesnt block Alex Cheema 2025-01-28 18:49:19 +00:00
  • a0d673fa3a Merge pull request #640 from exo-explore/simpledownload Alex Cheema 2025-01-27 19:38:11 +00:00
  • 7c649085a1 fix eta/speed for resuming an existing download, using the session downloaded bytes simpledownload Alex Cheema 2025-01-27 19:23:18 +00:00
  • 90e0e2761f ignore not_started progress updates Alex Cheema 2025-01-27 06:05:59 +00:00
  • 265586f7b4 set timeout on get too Alex Cheema 2025-01-27 06:05:40 +00:00
  • 4748bb7dc7 increase file download timeout to 30min Alex Cheema 2025-01-27 05:49:17 +00:00
  • ae770db4f3 increase download chunks to 1MB Alex Cheema 2025-01-27 05:37:50 +00:00
  • 82f75d0ccf increase hf download http timeout 15 mins for large downloads Alex Cheema 2025-01-27 05:20:30 +00:00
  • 295f41c5cc increase bench job timeout to give enough time to download Alex Cheema 2025-01-27 05:03:35 +00:00
  • 19a27c5bfd HF_HOME -> EXO_HOME Alex Cheema 2025-01-27 02:59:23 +00:00
  • d7ca9b7732 show each node id in the tinychat topology viz Alex Cheema 2025-01-27 02:20:22 +00:00
  • b349e48b0d fix visual bug where frontend would show the full hf repo size, but in some cases that includes redundant files so we should use the model index in those cases too Alex Cheema 2025-01-27 02:13:05 +00:00
  • 21586063f6 use llama-3.2-1b in tinygrad test Alex Cheema 2025-01-27 01:35:33 +00:00
  • 277d63d860 special case when a model doesnt have a model index file, then use wildcard for allow_patterns Alex Cheema 2025-01-27 01:26:15 +00:00
  • 74379ef671 log download logs with DEBUG>=6 very verbose Alex Cheema 2025-01-27 01:11:54 +00:00
  • 3c7bd48aa3 get rid of some more hf bloat Alex Cheema 2025-01-27 01:08:46 +00:00
  • 1df023023e remove a lot of hf bloat Alex Cheema 2025-01-27 01:06:47 +00:00
  • b89495f444 rewrite ShardDownloader, simplify significantly Alex Cheema 2025-01-27 00:37:57 +00:00
  • 903950f64e Merge pull request #638 from exo-explore/deepseekv3fix Alex Cheema 2025-01-26 20:33:22 +00:00
  • a3766f538a add exception for mlx-community/DeepSeek-R1-3bit and mlx-community/DeepSeek-V3-3bit in tokenizers test deepseekv3fix Alex Cheema 2025-01-26 20:32:48 +00:00
  • 9711d632e0 Merge pull request #637 from exo-explore/deepseekv3fix Alex Cheema 2025-01-26 20:31:53 +00:00
  • 82ef086010 add deepseek-v3-3bit and deepseek-r1-3bit Alex Cheema 2025-01-26 20:31:28 +00:00
  • 55ea366932 fix post_init deepseek v3 Alex Cheema 2025-01-26 20:27:31 +00:00
  • 63318983de Merge pull request #631 from sigseg5/main Alex Cheema 2025-01-26 19:20:58 +00:00
  • fb841a1f50 Adjust truncate size in history list for text without any spaces sigseg5 2025-01-26 00:38:58 +03:00
  • 4512366580 Fix bubble behavior when user passes long text without any spaces sigseg5 2025-01-26 00:02:17 +03:00
  • 9525c0e7a7 Add adaptive padding for user and assistant messages on width <= 1480px sigseg5 2025-01-26 00:01:54 +03:00
  • fe368aa2b6 remove redundant frontend logs deepseek Alex Cheema 2025-01-25 18:45:03 +00:00
  • 66f73768cc Merge pull request #627 from exo-explore/deepseek Alex Cheema 2025-01-24 18:14:57 +00:00
  • fdd05baddb fix tokenizer tests Alex Cheema 2025-01-24 18:13:36 +00:00
  • 59174bdc62 we have a lot of models so group them nicely Alex Cheema 2025-01-24 18:02:00 +00:00
  • cfdaaef8e6 handle thinking outputs nicely, format latex beautifully Alex Cheema 2025-01-24 17:49:25 +00:00
  • d8ffa59dba add deepseek v1, v3 and all the distills Alex Cheema 2025-01-24 16:39:38 +00:00
  • aa1ce21f82 Merge pull request #625 from eltociear/patch-1 Alex Cheema 2025-01-23 16:51:32 +00:00
  • 4fb01f516d chore: update manual_discovery.py Ikko Eltociear Ashimine 2025-01-24 00:18:42 +09:00
  • a635b23044 Merge pull request #619 from exo-explore/runners2 v0.0.13-alpha Alex Cheema 2025-01-23 02:18:33 +00:00
  • ad0e0d02d8 fix readme images runners2 Alex Cheema 2025-01-23 02:17:58 +00:00
  • 2644fd02c8 Merge pull request #617 from exo-explore/runners2 Alex Cheema 2025-01-23 02:05:17 +00:00
  • 88ac12df6c install clang test Alex Cheema 2025-01-23 01:55:14 +00:00
  • dfd9d3eb48 linux install Alex Cheema 2025-01-23 01:44:57 +00:00
  • 200ff4d713 linux install Alex Cheema 2025-01-23 01:43:00 +00:00
  • b2764f177f linux install Alex Cheema 2025-01-23 01:40:59 +00:00
  • e57fa1dfa0 xlarge Alex Cheema 2025-01-23 01:40:13 +00:00
  • 209163c595 add linux tinygrad test Alex Cheema 2025-01-23 01:38:10 +00:00
  • 495987b50b beef up the instance Alex Cheema 2025-01-23 01:37:38 +00:00
  • 8484eb4165 fix config Alex Cheema 2025-01-23 01:37:01 +00:00
  • 790c08afd4 add linux tinygrad test Alex Cheema 2025-01-23 01:31:44 +00:00
  • a8a9e3ffa1 explicitly enable TOKENIZERS_PARALLELISM=true Alex Cheema 2025-01-23 01:26:27 +00:00
  • 5c9bcb8620 set GRPC_VERBOSITY=error; TRANSFORMERS_VERBOSITY=error Alex Cheema 2025-01-23 01:22:19 +00:00
  • d54e19c20a runners back Alex Cheema 2025-01-23 00:55:52 +00:00
  • cc78738e24 remove kern scan intervals Alex Cheema 2025-01-23 00:49:32 +00:00
  • 2391051c11 remove kern.timer.scan_interval from bootstrap.sh Alex Cheema 2025-01-23 00:41:40 +00:00
  • 112dea1582 add back the benchmarks baby Alex Cheema 2025-01-23 00:15:54 +00:00
  • dc5cdc4d78 add back opaque Alex Cheema 2025-01-22 23:59:39 +00:00
  • f8db4e131e fix check for sd2.1 Alex Cheema 2025-01-22 23:53:42 +00:00
  • bbb6856988 fix check for sd2.1 Alex Cheema 2025-01-22 23:51:09 +00:00
  • 9ba8bbbcf8 fix filter to include 169.254.* since thats what mac uses for ethernet Alex Cheema 2025-01-22 23:47:43 +00:00
  • 8ab9977f01 fix stable diffusion case for tui, make mlx run on its own thread again and non-blocking Alex Cheema 2025-01-22 23:22:53 +00:00
  • 3a4bae0dab fix issue with eos_token_id Alex Cheema 2025-01-22 22:58:09 +00:00
  • 87d1271d33 fix stream: false completion Alex Cheema 2025-01-22 22:46:04 +00:00
  • 55d1846f5e clean up DEBUG=2 logs, a few fixes for token Alex Cheema 2025-01-22 22:27:02 +00:00
  • 9954ce8e4d fix treating token as a list Alex Cheema 2025-01-22 22:13:13 +00:00
  • 09e12d8673 temporarily disable github runner benchmarks Alex Cheema 2025-01-22 22:00:13 +00:00
  • 98d6e986bd add back .circleci Alex Cheema 2025-01-22 21:58:46 +00:00
  • d80324fe20 disable test-m3-single-node Alex Cheema 2025-01-22 21:58:40 +00:00
  • 97f3bad38f fix peer_handle Alex Cheema 2025-01-22 21:07:49 +00:00
  • 461e4f37cb Merge remote-tracking branch 'origin/main' into runners2 Alex Cheema 2025-01-22 21:06:12 +00:00
  • 07ceb19f0a Merge pull request #614 from samiamjidkhan/main v0.0.12-alpha Alex Cheema 2025-01-22 14:59:54 +00:00
  • 27b4577f38 directory for images Sami Khan 2025-01-22 05:47:25 -05:00
  • a70943f8d2 base images for animation Sami Khan 2025-01-22 05:46:38 -05:00
  • 410d901505 Merge pull request #613 from samiamjidkhan/dmg-backend vv0.0.10-alpha v0.0.11-alpha v0.0.10-alpha Alex Cheema 2025-01-21 13:12:08 +00:00
  • 5c4ce5392c image and text mode fix Sami Khan 2025-01-21 04:33:54 -05:00
  • 819ec7626e Merge pull request #611 from exo-explore/fixbuildname v0.0.9-alpha Alex Cheema 2025-01-21 05:36:34 +00:00
  • ba5bb3e171 fix scripts/build_exo.py: com.exolabs.exo -> net.exolabs.exo fixbuildname Alex Cheema 2025-01-21 05:36:02 +00:00
  • f4bbcf4c8f Merge pull request #607 from tensorsofthewall/smol_fix Alex Cheema 2025-01-21 02:21:18 +00:00
  • 6b8cd0577e fix some issues with results Alex Cheema 2025-01-20 16:30:16 +00:00
  • 218c1e79d9 Merge branch 'main' into runners2 Alex Cheema 2025-01-20 16:12:55 +00:00
  • b9eccedc3d Formatting Sandesh Bharadwaj 2025-01-17 05:40:42 -05:00
  • 5f06aa2759 Replace netifaces (unmaintained,outdated) with scapy + add dependencies for previous fixes Sandesh Bharadwaj 2025-01-17 05:37:01 -05:00
  • 349b5344eb Minor fix for Shard typing Sandesh Bharadwaj 2025-01-16 14:36:46 -05:00
  • df3624d27a Add AMD GPU querying + Windows device capabilities Sandesh Bharadwaj 2025-01-14 20:37:02 -05:00
  • 6737e36e23 Fixed MLX import blocking native Windows execution of exo. (Not Final) Sandesh Bharadwaj 2025-01-14 20:35:21 -05:00
  • c260689a06 Merge pull request #602 from exo-explore/fixexodir v0.0.8-alpha v0.0.7-alpha v0.0.5-alpha Alex Cheema 2025-01-12 03:46:14 +00:00
  • fcc699a55f fix fixexodir Alex Cheema 2025-01-12 03:40:59 +00:00
  • e7b98f5ae5 fix unit tests Alex Cheema 2025-01-12 03:35:24 +00:00
  • ffe78f6d0b fix dummy test Alex Cheema 2025-01-12 03:30:06 +00:00
  • ce5041ee1b types Alex Cheema 2025-01-12 03:24:42 +00:00
  • 9b2c01c873 ensure dir exists Alex Cheema 2025-01-12 03:15:49 +00:00
  • 2aed3f3518 handle inference_state properly Alex Cheema 2025-01-12 03:13:17 +00:00
  • 2af5ee02e4 fix exo folder Alex Cheema 2025-01-12 03:10:11 +00:00
  • b5cbcbc7a2 Merge pull request #474 from pranav4501/stable-stable-diffusion-mlx Alex Cheema 2025-01-12 02:57:21 +00:00
  • 5f3d000a7b Merge branch 'main' into stable-stable-diffusion-mlx Alex Cheema 2025-01-12 02:56:34 +00:00
  • bd2e8e7a5a Merge pull request #598 from exo-explore/fixphitest Alex Cheema 2025-01-08 22:09:38 +00:00
  • 40696b21f7 typo in phi test fixphitest Alex Cheema 2025-01-08 22:09:04 +00:00