Commit Graph

  • 4631c48273 Remove some todos. (#1042) Laurent Mazare 2023-10-05 22:42:20 +01:00
  • 716883e9b0 Add the clamping for stable-diffusion. (#1041) Laurent Mazare 2023-10-05 22:20:39 +01:00
  • 47c25a567b feat: [SAM] able to download the result as png (#1035) lichin-lin 2023-10-05 22:14:47 +01:00
  • 7f7d95e2c3 Add the round-to function. (#1039) Laurent Mazare 2023-10-05 20:28:09 +01:00
  • f47bd9bab5 Delete invalid comment (#1038) Juarez Bochi 2023-10-05 14:28:08 -04:00
  • 8f7973958c fix: fix index_select cuda kernel for src target dim different than ids dim when selecting dim > 0 (#1037) Gonzalo 2023-10-05 14:46:13 -03:00
  • f0c619a4af Use AsRef<str> for set_one. (#1033) Laurent Mazare 2023-10-05 06:05:44 +01:00
  • b86ac0c507 Quant t5: Add coedit model to wasm demo and readme (#1031) Juarez Bochi 2023-10-04 15:57:33 -04:00
  • 27e70a5093 Whisper quantized wasm (#1028) Radamés Ajna 2023-10-04 12:22:57 -07:00
  • c18a856e76 Add the rounding operators. (#1030) Laurent Mazare 2023-10-04 17:58:44 +01:00
  • 3349c89252 Add quantized t5 args for weight and config (#1029) Juarez Bochi 2023-10-04 12:02:49 -04:00
  • 11d3687cc6 Simd128 optimized q8k vecdot. (#1026) Laurent Mazare 2023-10-03 15:29:48 +01:00
  • dac73edb34 AVX optimized q8k vecdot. (#1024) Laurent Mazare 2023-10-03 12:10:58 +01:00
  • b4da19d1be Merge pull request #1023 from evgenyigumnov/simlified-book-polish Nicolas Patry 2023-10-03 12:29:41 +02:00
  • ff513314fc small misspeling and polish fix Evgeny Igumnov 2023-10-03 15:47:04 +06:00
  • 043cc25766 Fix for the index-select cuda setup. (#1022) Laurent Mazare 2023-10-03 10:21:46 +01:00
  • 7b06872f90 Merge pull request #926 from evgenyigumnov/book-trainin-simplified Nicolas Patry 2023-10-03 10:41:30 +02:00
  • 65825e7240 [SAM] Add undo button and background point mode (#1020) Radamés Ajna 2023-10-02 15:33:46 -07:00
  • 7670fe7d1f neon optimized q8k multiplication. (#1021) Laurent Mazare 2023-10-02 23:26:34 +01:00
  • cddfc3944c Add the q8k vec-dot multiplication. (#1019) Laurent Mazare 2023-10-02 21:53:34 +01:00
  • 089fc3b584 Improve the quantized whisper setup. (#1018) Laurent Mazare 2023-10-02 17:17:46 +01:00
  • e04c789230 Add a quantized variant of whisper (#1017) Laurent Mazare 2023-10-02 14:59:53 +01:00
  • 263a172202 Improve the testing of the optimized quantized vec-dot ops (#1016) Laurent Mazare 2023-10-02 09:50:43 +01:00
  • 638ccf9f46 Fix include code. Nicolas Patry 2023-10-02 10:22:44 +02:00
  • 0baf5a1e19 Fixed PR warnings. Nicolas Patry 2023-10-02 10:15:10 +02:00
  • 5130a7da32 Simd128 version of q6k vec-dot. (#1015) Laurent Mazare 2023-10-01 19:44:12 +01:00
  • 41143db1af [segment-anything] add multi point logic for demo site (#1002) lichin-lin 2023-10-01 18:25:22 +01:00
  • 096dee7073 Bump the version to 0.3.0. (#1014) Laurent Mazare 2023-10-01 13:51:57 +01:00
  • f6054e9d60 Fix the prompt for mistral when using instruct/interactive mode. (#1013) Laurent Mazare 2023-10-01 06:44:30 +01:00
  • 328167ec04 Integrate TheBloke quantized mistral weights. (#1012) Laurent Mazare 2023-09-30 22:39:42 +01:00
  • 4e55aaa51f Simd128 version of the q2k-q8k vecdot product. (#1011) Laurent Mazare 2023-09-30 20:12:41 +01:00
  • deee7612da Quantized version of mistral. (#1009) Laurent Mazare 2023-09-30 19:25:47 +02:00
  • 06207332bc Streaming mode for reporting the generated tokens (#1007) Laurent Mazare 2023-09-30 16:04:11 +02:00
  • 4021272875 Use flash-attn for mistral. (#1004) Laurent Mazare 2023-09-30 13:15:10 +02:00
  • 87e3a4e175 Mistral: exit on eos token. (#1001) Laurent Mazare 2023-09-30 08:07:06 +02:00
  • 6203ced495 Add negative prompts to segment-anything. (#1000) Laurent Mazare 2023-09-30 07:17:42 +02:00
  • 34842fb234 [segment-anything] Print IOU values to help with debugging (#999) GeauxEric 2023-09-29 21:44:42 -07:00
  • d188d6a764 Fix the multiple points case for sam. (#998) Laurent Mazare 2023-09-29 22:39:43 +02:00
  • 0ac2db577b Add an entry about WSL slowness to the faq. (#997) Laurent Mazare 2023-09-29 18:04:52 +02:00
  • fc59bc31bf fix: add missing gpu fill_* (#996) Gonzalo 2023-09-29 11:49:30 -03:00
  • 03348e2e6f Update mistral README.md (#995) Laurent Mazare 2023-09-29 13:24:32 +02:00
  • 49fa184a35 Mistral readme (#994) Laurent Mazare 2023-09-29 12:50:50 +02:00
  • 6f17ef82be Mistral: print the generated text. (#992) Laurent Mazare 2023-09-29 11:56:11 +02:00
  • 01b92cd959 fixes slice_scatter dim type (#988) Gonzalo 2023-09-29 03:54:45 -03:00
  • 53510ce427 Use a silu activation in mistral. (#991) Laurent Mazare 2023-09-29 08:06:54 +02:00
  • 23b3576c47 Add the sliding window. (#986) Laurent Mazare 2023-09-28 18:26:33 +02:00
  • 716ab2ccdc Mistral gpu fix (#985) Laurent Mazare 2023-09-28 17:38:13 +02:00
  • ada8851a23 Add the mistral example. (#984) Laurent Mazare 2023-09-28 17:19:18 +02:00
  • c05a348e36 Add the Mistral 7b model (#983) Laurent Mazare 2023-09-28 14:29:41 +01:00
  • 25657804ef Simd128 q2k vecdot (#982) Laurent Mazare 2023-09-28 12:16:35 +01:00
  • 5e1c595e00 Optimize the index-select cuda kernel. (#976) Laurent Mazare 2023-09-28 09:05:29 +01:00
  • 8a49e01b9d Add the remaining quantized tests to the wasm suite. (#980) Laurent Mazare 2023-09-28 08:42:56 +01:00
  • 9cb110c44c Sketch a simd128 optimized q4k vecdot. (#977) Laurent Mazare 2023-09-27 20:19:38 +01:00
  • 667f01c173 Simd128 vec-dot for q4_0. (#974) Laurent Mazare 2023-09-27 14:15:30 +01:00
  • e59784e353 simd128 optimized q8_0 vecdot (#972) Laurent Mazare 2023-09-27 11:03:20 +01:00
  • 29bd6b2979 Phi 1.5 wasm module (#966) Radamés Ajna 2023-09-26 22:07:11 -07:00
  • 9571b200c9 fix firstToken, minor ui changes (#971) Radamés Ajna 2023-09-26 22:01:59 -07:00
  • ce0a4e3a85 Use the gelu-erf activation. (#969) Laurent Mazare 2023-09-26 22:30:21 +01:00
  • 4abc1ea34d Avoid some overflows on wasm32. (#968) Laurent Mazare 2023-09-26 11:15:38 +01:00
  • 2dd43d6cdd add eos token to phi example (#965) Radamés Ajna 2023-09-26 01:21:22 -07:00
  • 1fcac4afed Expose a function to clear the KV cache on mixformers. (#964) Laurent Mazare 2023-09-26 05:41:07 +01:00
  • a084f65f9a fix rep penalty min value (#963) Radamés Ajna 2023-09-25 21:23:50 -07:00
  • c798184c2b Configurable layer idx for the lstm layer. (#962) Laurent Mazare 2023-09-25 21:31:14 +01:00
  • c78a294323 Add some repeat penalty to the phi example. (#961) Laurent Mazare 2023-09-25 20:53:30 +01:00
  • a36d883254 Use a single flag for the point argument. (#958) Laurent Mazare 2023-09-25 12:53:24 +01:00
  • 7f2bbcf746 [segment-anything] Support multi-point as the prompt input (#945) GeauxEric 2023-09-25 04:14:10 -07:00
  • dc47224ab9 Override the default cudnn heuristics. (#957) Laurent Mazare 2023-09-25 10:31:53 +01:00
  • 1ce7fe2543 Add more examples to the phi readme. (#956) Laurent Mazare 2023-09-24 18:19:05 +01:00
  • 402ddcfcb4 Add the missing kernel. (#955) Laurent Mazare 2023-09-24 17:21:37 +01:00
  • f5069dd354 Use the repo for the quantized phi model. (#954) Laurent Mazare 2023-09-24 16:30:26 +01:00
  • 0007ae9c11 Add the quantized mixformer model. (#953) Laurent Mazare 2023-09-24 15:03:48 +01:00
  • e15862cfdb Shared the quantized var-builder code. (#952) Laurent Mazare 2023-09-24 12:55:07 +01:00
  • 4aeb449017 Depreate the VarBuilder::from_safetensors function. (#951) Laurent Mazare 2023-09-24 11:18:17 +01:00
  • bcb0ed8f1c Self-contained safetensors for the multiprocess llama example. (#950) Laurent Mazare 2023-09-24 06:54:49 +01:00
  • 7edd755756 Pass directly the buffer ownership. (#949) Laurent Mazare 2023-09-24 06:34:44 +01:00
  • e32c89d90c Add the buffered safetensor wrapper. (#948) Laurent Mazare 2023-09-23 22:57:42 +01:00
  • bb3471ea31 Adapt more examples to the updated safetensor api. (#947) Laurent Mazare 2023-09-23 21:26:03 +01:00
  • 890d069092 Self-contained safetensor wrappers (#946) Laurent Mazare 2023-09-23 20:39:52 +01:00
  • 5dbe46b389 Add tracing. (#943) Laurent Mazare 2023-09-23 16:55:46 +01:00
  • ccf352f3d1 Use yoke to provide a self-referential container for mmaped safetenso… (#939) Laurent Mazare 2023-09-23 15:43:11 +01:00
  • 402d207f0f VarMap setter functions (#938) Laurent Mazare 2023-09-23 10:27:51 +01:00
  • 7582937a32 Add the causal mask in mixformer. (#937) Laurent Mazare 2023-09-23 09:50:26 +01:00
  • b54acfa3d0 Tracing for the phi model (#936) Laurent Mazare 2023-09-23 09:19:34 +01:00
  • cda1786eed smaller t5 models quantized (#934) Radamés Ajna 2023-09-22 14:31:23 -07:00
  • 912a3d63b0 Use the proper block size for quantizing models. (#933) Laurent Mazare 2023-09-22 21:36:56 +01:00
  • 3ef328c53d Mention the new phi model in the readme. (#932) Laurent Mazare 2023-09-22 21:24:51 +01:00
  • 0c8e983514 update link to t5 (#931) Radamés Ajna 2023-09-22 12:30:01 -07:00
  • df6f5240ba Complete the mixformer implementation. (#930) Laurent Mazare 2023-09-22 20:03:16 +01:00
  • a46b1b4657 Mixformer (#929) Laurent Mazare 2023-09-22 16:17:14 +01:00
  • 19e52e5007 T5 Wasm (#918) Radamés Ajna 2023-09-22 07:31:10 -07:00
  • 8601537e31 Add slice-scatter. (#927) Laurent Mazare 2023-09-22 12:18:16 +01:00
  • 4ac6039a42 Merge branch 'main' into book-trainin-simplified Evgeny Igumnov 2023-09-22 11:01:23 +06:00
  • 52a60ca3ad https://github.com/huggingface/candle/issues/637 Evgeny Igumnov 2023-09-22 10:57:11 +06:00
  • a96878f235 cuda cast i64 (#925) Gonzalo 2023-09-21 15:52:39 -03:00
  • aa8ec06fd2 Add the t5-xxl version. (#924) Laurent Mazare 2023-09-21 14:48:13 +01:00
  • b43ca493f6 Add more quantized flan t5 variants (#923) Laurent Mazare 2023-09-21 13:23:30 +01:00
  • 3b557765e8 T5 quantized example (#922) Laurent Mazare 2023-09-21 12:33:15 +01:00
  • 2619c4307f Add a quantized version of the t5 model. (#921) Laurent Mazare 2023-09-21 11:13:39 +01:00
  • c89b82b2d4 Add a clear cache function to the t5 model. (#919) Laurent Mazare 2023-09-21 09:01:06 +01:00
  • 7b26e513f1 Add the erf function. (#917) Laurent Mazare 2023-09-21 06:19:10 +01:00