Commit Graph

  • f631dd6ff7 Templates: Switch to Jinja2 kingbri 2023-12-17 00:41:42 -05:00
  • 95fd0f075e Model: Fix no flash attention kingbri 2023-12-17 23:31:58 -05:00
  • ad8807a830 Model: Add support for num_experts_by_token kingbri 2023-12-17 18:03:01 -05:00
  • 70fbee3edd OAI: Fix model parameter placement kingbri 2023-12-17 14:36:28 -05:00
  • 1d0bdfa77c Model + OAI: Fix parameter parsing kingbri 2023-12-17 14:28:18 -05:00
  • 3e57125025 OAI: adding optional draft model properties for draft_rope alpha and scale (#28) Veden 2023-12-17 11:23:45 -08:00
  • 528d58f841 Requirements: Fix AMD kingbri 2023-12-17 00:45:43 -05:00
  • f196f1177d Requirements: Update exllamav2 to 0.0.11 kingbri 2023-12-16 19:32:46 -05:00
  • 1a331afe3a OAI: Add cache_mode parameter to model kingbri 2023-12-16 02:42:36 -05:00
  • ed868fd262 OAI: Remove unused parameters kingbri 2023-12-15 14:53:12 -05:00
  • 59729e2a4a Tests: Fix linting kingbri 2023-12-13 23:05:50 -05:00
  • 036ba2669c Auth: Migrate to Pydantic kingbri 2023-12-12 23:58:22 -05:00
  • eb8ccb9783 Tree: Fix linter issues kingbri 2023-12-12 23:48:13 -05:00
  • 083df7d585 Tree: Add generation logging support kingbri 2023-12-12 23:43:35 -05:00
  • b364de1005 Update README kingbri 2023-12-10 16:05:46 -05:00
  • db87efde4a OAI: Add ability to specify fastchat prompt template kingbri 2023-12-10 15:23:49 -05:00
  • 9f195af5ad Main: Fix function calls kingbri 2023-12-10 13:28:21 -05:00
  • fd9f3eac87 Model: Add params to current model endpoint kingbri 2023-12-10 00:40:56 -05:00
  • 0f4290f05c Model: Format kingbri 2023-12-09 22:38:06 -05:00
  • 5ae2a91c04 Tree: Use unwrap and coalesce for optional handling kingbri 2023-12-09 21:52:17 -05:00
  • 7380a3b79a Implement lora support (#24) DocShotgun 2023-12-08 20:36:40 -08:00
  • 161c9d2c19 Tests: Fix wheel test kingbri 2023-12-08 01:15:24 -05:00
  • fa1e99daf6 Model: Remove unused print statement kingbri 2023-12-07 21:13:52 -05:00
  • 47176a2a1e Requirements: Fix torch install kingbri 2023-12-06 19:04:35 -05:00
  • f8e9e22c43 API: Fix model load endpoint with draft kingbri 2023-12-06 18:05:55 -05:00
  • 6a71890d45 Model: Fix sampler bugs kingbri 2023-12-06 17:25:13 -05:00
  • 9f34af4906 Tests: Create kingbri 2023-12-06 00:53:42 -05:00
  • 21c25fd806 Update README kingbri 2023-12-06 00:21:52 -05:00
  • b83e1b704e Requirements: Split for configurations kingbri 2023-12-06 00:00:30 -05:00
  • 4c0e686e7d Model: Cleanup and fix fallbacks kingbri 2023-12-05 23:28:16 -05:00
  • 0ef2fe9b95 Merge pull request #23 from DocShotgun/main Brian Dashore 2023-12-05 22:24:53 -05:00
  • d8f7b93c54 Model: Fix fetching of draft args kingbri 2023-12-05 22:23:14 -05:00
  • 3f2fcbcc45 Add fallback to draft_rope_scale to 1.0 DocShotgun 2023-12-05 18:51:36 -08:00
  • 39f7a2aabd Expose draft_rope_scale DocShotgun 2023-12-05 12:59:32 -08:00
  • e085b806e8 Merge pull request #22 from DocShotgun/main Brian Dashore 2023-12-05 01:22:33 -05:00
  • 67507105d0 Update colab, expose additional args DocShotgun 2023-12-04 22:20:46 -08:00
  • 37f8f3ef8b Merge pull request #20 from veryamazinglystupid/main Brian Dashore 2023-12-05 01:14:21 -05:00
  • 621e11b940 Update documentation kingbri 2023-12-05 00:33:43 -05:00
  • 8ba3bfa6b3 API: Fix load exception handling kingbri 2023-12-05 00:23:15 -05:00
  • 7c92968558 API: Fix mistaken debug statement kingbri 2023-12-04 18:07:12 -05:00
  • 5e54911cc8 API: Fix semaphore handling and chat completion errors kingbri 2023-12-04 15:51:25 -05:00
  • 30fc5b3d29 Merge branch 'main' of github.com:theroyallab/tabbyAPI kingbri 2023-12-03 22:55:51 -05:00
  • ed6c962aad API: Fix sequential requests kingbri 2023-12-03 22:54:34 -05:00
  • ad1a12a0f2 make colab better, fix libcudart errors veryamazinglystupid 2023-12-03 14:07:52 +05:30
  • 2a9e4ca051 Add Colab example DocShotgun 2023-12-02 23:19:53 -08:00
  • e740b53478 Requirements: Update Flash Attention 2 kingbri 2023-12-03 01:56:29 -05:00
  • c67c9f6d66 Model + Config: Remove low_mem option kingbri 2023-12-03 01:07:42 -05:00
  • 109e4223e0 Merge pull request #18 from DocShotgun/main Brian Dashore 2023-12-03 01:06:50 -05:00
  • 27fc0c0069 Model: Cleanup and compartmentalize auto rope functions kingbri 2023-12-03 01:05:09 -05:00
  • bd2c5d0d09 Force auto-alpha to 1.0 if config ctx == base ctx DocShotgun 2023-12-02 21:19:59 -08:00
  • 1c398b0be7 Add automatic NTK-aware alpha scaling to model DocShotgun 2023-12-02 21:02:29 -08:00
  • 61f6e51fdb OAI: Add separator style fallback kingbri 2023-12-01 23:30:19 -05:00
  • ae69b18583 API: Use FastAPI streaming instead of sse_starlette kingbri 2023-12-01 01:54:35 -05:00
  • 6493b1d2aa OAI: Add ability to send dummy models kingbri 2023-12-01 00:27:28 -05:00
  • aef411bed5 OAI: Fix chat completion streaming kingbri 2023-12-01 00:14:24 -05:00
  • c4d8c901e1 Merge pull request #13 from ziadloo/main Brian Dashore 2023-11-30 01:57:44 -05:00
  • 8a5ac5485b Model: Fix rounding kingbri 2023-11-30 01:55:46 -05:00
  • e703c716ee Merge branch 'main' of https://github.com/ziadloo/tabbyAPI into ziadloo-main kingbri 2023-11-30 01:01:48 -05:00
  • 56f9b1d1a8 API: Add generator error handling kingbri 2023-11-30 00:37:48 -05:00
  • 2bc3da0155 YAML: Force all files to open with utf8 kingbri 2023-11-29 22:04:29 -05:00
  • 3957316b79 Revert "API: Rename repetition_decay -> repetition_slope" kingbri 2023-11-29 21:54:51 -05:00
  • 94696543bc Model: Warn user if context > max_seq_len kingbri 2023-11-29 01:35:32 -05:00
  • cad144126f API: Rename repetition_decay -> repetition_slope kingbri 2023-11-29 01:13:05 -05:00
  • 5cbf7f13da OAI: Fix repetition range kingbri 2023-11-29 00:53:19 -05:00
  • b0c42d0f05 Leveraging local variables Mehran Ziadloo 2023-11-27 20:56:56 -08:00
  • ead503c75b Adding token usage support Mehran Ziadloo 2023-11-27 20:05:05 -08:00
  • 44e7f7b0ee Update README kingbri 2023-11-25 23:47:38 -05:00
  • 0914bc313f Merge pull request #12 from DocShotgun/main Brian Dashore 2023-11-25 00:29:47 -05:00
  • d929e0c826 API: Fix error points and exceptions kingbri 2023-11-25 00:27:02 -05:00
  • cffd20f580 Add start-up shell script for Linux DocShotgun 2023-11-23 19:03:52 -08:00
  • d47c39da54 API: Don't include draft directory in response kingbri 2023-11-23 00:07:56 -05:00
  • 13c9c09398 Update README kingbri 2023-11-22 00:20:21 -05:00
  • d25310e55d Requirements: Update Flash Attention 2 kingbri 2023-11-21 22:12:55 -05:00
  • 71b9a53336 API: Add temperature_last support kingbri 2023-11-21 21:20:59 -05:00
  • 3337fe6acc Warning if unsupported samplers are used turboderp 2023-11-21 18:35:22 +01:00
  • a54de11cf3 Add new samplers turboderp 2023-11-21 18:16:53 +01:00
  • c92ee24bb4 Tree: Add batch script kingbri 2023-11-20 01:48:06 -05:00
  • 2aa9c145be Auth: Fix an oops with headers kingbri 2023-11-20 00:04:10 -05:00
  • 39ea730be5 Auth: Allow admin keys to work with api key routes kingbri 2023-11-19 23:53:07 -05:00
  • 8ef730f016 Merge pull request #11 from veden/patch-1 turboderp 2023-11-20 04:23:34 +01:00
  • f960fac8ff Fix incorrect ratio calculation for draft model Veden 2023-11-19 13:12:53 -08:00
  • 4cddd0400c Model: Fix draft model loading kingbri 2023-11-19 02:04:02 -05:00
  • 698b0b1976 Update README kingbri 2023-11-19 01:19:31 -05:00
  • 581e1fc219 Sample config: Remove unused value kingbri 2023-11-19 01:16:03 -05:00
  • e0e93c103b Sample config: Uncomment all parameters kingbri 2023-11-19 01:10:57 -05:00
  • 63762654f0 Update README kingbri 2023-11-19 01:05:20 -05:00
  • e46676cb08 Merge pull request #9 from city-unit/main Brian Dashore 2023-11-19 00:53:24 -05:00
  • e4a8848445 Auth: Log API and admin key on startup kingbri 2023-11-19 00:52:39 -05:00
  • 31bc418795 Model: Add context in response output kingbri 2023-11-19 00:49:32 -05:00
  • 80c69939ae Remove unneeded stuffs city_unit 2023-11-19 00:34:54 -05:00
  • f47919b1d3 API: Add draft model support kingbri 2023-11-19 00:32:25 -05:00
  • 6b22dc0119 Rename, fschat support city_unit 2023-11-19 00:32:14 -05:00
  • 99cf0b6d7b Add basic docker support city_unit 2023-11-19 00:01:17 -05:00
  • 6b9af58cc1 Tree: Fix extraneous bugs and update T/s print kingbri 2023-11-18 22:34:40 -05:00
  • a51889bdb8 Requirements: Update Flash Attention kingbri 2023-11-18 21:59:34 -05:00
  • b2410a0436 Merge pull request #4 from waldfee/config_samples Brian Dashore 2023-11-18 13:16:23 -05:00
  • 27ebec3b35 Model: Add speculative decoding support via config kingbri 2023-11-18 01:38:54 -05:00
  • 2ad79cb9ea Model: Add tokens in responses kingbri 2023-11-17 23:33:48 -05:00
  • 7f18ea1d7c Tree: Remove SillyTavern shim docs kingbri 2023-11-17 22:03:46 -05:00
  • 6f2078cbe4 Update README kingbri 2023-11-17 22:02:21 -05:00