Commit graph

Select branches

Hide pull requests

main

Mono Color

• 42f816312d Merge remote-tracking branch 'refs/remotes/origin/dev' into dev oobabooga 2023-11-02 11:09:26 -07:00
• 77abd9b69b Add no_flash_attn option oobabooga 2023-11-02 08:19:42 -07:00
• a56ef2a942
  make torch.load a bit safer (#4448) Julien Chaumond 2023-11-02 18:07:08 +01:00
• deba039c03
  (fix): OpenOrca-Platypus2 models should use correct instruction_template and custom_stopping_strings (#4435) deevis 2023-10-31 22:51:00 -06:00
• aaf726dbfb
  Updating the shared settings object when loading a model (#4425) Mehran Ziadloo 2023-10-31 21:29:57 -07:00
• 9bd0724d85 Change frequency/presence penalty ranges oobabooga 2023-10-31 20:57:56 -07:00
• 6b7fa45cc3
  Update exllamav2 version (#4417) Orang 2023-11-01 05:12:14 +07:00
• 41e159e88f
  Bump AutoAWQ to v0.1.5 (#4410) Casper 2023-10-31 23:11:22 +01:00
• 0707ed7677
  updated wiki link (#4415) Meheret 2023-11-01 01:09:05 +03:00
• 262f8ae5bb Use default gr.Dataframe for evaluation table oobabooga 2023-10-27 06:49:14 -07:00
• f481ce3dd8
  Adding platform_system to autoawq (#4390) James Braza 2023-10-26 21:02:28 -07:00
• af98587580
  Update accelerate requirement from ==0.23.* to ==0.24.* (#4400) dependabot[bot] 2023-10-27 00:46:16 -03:00
• 839a87bac8 Fix is_ccl_available & is_xpu_available imports oobabooga 2023-10-26 20:26:25 -07:00
• 778a010df8
  Intel Gpu support initialization (#4340) Abhilash Majumder 2023-10-27 08:09:51 +05:30
• 317e2c857e
  sd_api_pictures: fix Gradio warning message regarding custom value (#4391) GuizzyQC 2023-10-26 22:03:21 -04:00
• 92b2f57095 Minor metadata bug fix (second attempt) oobabooga 2023-10-26 18:57:32 -07:00
• 2d97897a25 Don't install flash-attention on windows + cuda 11 oobabooga 2023-10-25 11:21:18 -07:00
• 0ced78fdfa
  Replace hashlib.sha256 with hashlib.file_digest so we don't need to load entire files into ram before hashing them. (#4383) LightningDragon 2023-10-25 09:15:34 -06:00
• 72f6fc6923
  Rename additive_repetition_penalty to presence_penalty, add frequency_penalty (#4376) tdrussell 2023-10-25 10:10:28 -05:00
• ef1489cd4d Remove unused parameter in AutoAWQ oobabooga 2023-10-23 20:45:43 -07:00
• 1edf321362 Lint oobabooga 2023-10-23 13:09:03 -07:00
• 280ae720d7 Organize oobabooga 2023-10-23 13:07:17 -07:00
• 49e5eecce4 Merge remote-tracking branch 'refs/remotes/origin/main' oobabooga 2023-10-23 12:54:05 -07:00
• 82c11be067 Update 04 - Model Tab.md oobabooga 2023-10-23 12:49:07 -07:00
• 306d764ff6 Minor metadata bug fix oobabooga 2023-10-23 12:46:24 -07:00
• 4bc411332f
  Fix broken links (#4367) adrianfiedler 2023-10-23 19:09:57 +02:00
• 92691ee626 Disable trust_remote_code by default oobabooga 2023-10-23 09:57:44 -07:00
• 4440f87722
  Add additive_repetition_penalty sampler setting. (#3627) tdrussell 2023-10-23 00:28:07 -05:00
• 6086768309 Bump gradio to 3.50.* oobabooga 2023-10-22 21:20:41 -07:00
• b8183148cf
  Update 04 ‐ Model Tab.md oobabooga 2023-10-22 17:15:55 -03:00
• cea7fc2435 Update html_instruct_style.css oobabooga 2023-10-22 12:28:23 -07:00
• df90d03e0b Replace --mul_mat_q with --no_mul_mat_q oobabooga 2023-10-22 12:22:06 -07:00
• d0c3b407b3
  transformers loader: multi-LoRAs support (#3120) Googulator 2023-10-22 21:06:22 +02:00
• 4405513ca5
  Option to select/target additional linear modules/layers in LORA training (#4178) omo 2023-10-22 11:57:19 -07:00
• 7a3f885ea8
  Update 03 ‐ Parameters Tab.md oobabooga 2023-10-22 14:52:23 -03:00
• 63688004dc Add default cmd flags to colab oobabooga 2023-10-22 09:56:43 -07:00
• 613feca23b Make colab functional for llama.cpp oobabooga 2023-10-22 09:06:20 -07:00
• 994502d41b Colab fixes oobabooga 2023-10-22 08:57:16 -07:00
• c544f5cc51
  Support LLaVA v1.5 7B (#4348) Jiashu Xu 2023-10-22 11:49:04 -04:00
• 05741821a5 Minor colab changes oobabooga 2023-10-22 08:44:05 -07:00
• 6a61158adf
  Training PRO a month worth of updates (#4345) FartyPants (FP HAM) 2023-10-22 11:38:09 -04:00
• c18504f369
  USE_CUDA118 from ENV remains null one_click.py + cuda-toolkit (#4352) mongolu 2023-10-22 18:37:24 +03:00
• cd45635f53 tqdm improvement for colab oobabooga 2023-10-21 22:00:29 -07:00
• ae79c510cc Merge remote-tracking branch 'refs/remotes/origin/main' oobabooga 2023-10-21 21:46:15 -07:00
• 2d1b3332e4 Ignore warnings on Colab oobabooga 2023-10-21 21:45:25 -07:00
• caf6db07ad
  Update README.md oobabooga 2023-10-22 01:22:17 -03:00
• 1a34927314 Make API URLs more visible oobabooga 2023-10-21 21:11:07 -07:00
• 09f807af83 Use ExLlama_HF for GPTQ models by default oobabooga 2023-10-21 20:45:38 -07:00
• 619093483e Add Colab notebook oobabooga 2023-10-21 20:27:52 -07:00
• 506d05aede Organize command-line arguments oobabooga 2023-10-21 18:52:59 -07:00
• b1f33b55fd
  Update 01 ‐ Chat Tab.md oobabooga 2023-10-21 20:17:56 -03:00
• ac6d5d50b7
  Update README.md oobabooga 2023-10-21 20:03:43 -03:00
• 6efb990b60
  Add a proper documentation (#3885) oobabooga 2023-10-21 19:15:54 -03:00
• 5a5bc135e9
  Docker: Remove explicit CUDA 11.8 Reference (#4343) Adam White 2023-10-21 13:09:34 -05:00
• b98fbe0afc Add download link oobabooga 2023-10-20 23:58:05 -07:00
• fbac6d21ca Add missing exception oobabooga 2023-10-20 23:53:24 -07:00
• 3345da2ea4
  Add flash-attention 2 for windows (#4235) Brian Dashore 2023-10-21 02:46:23 -04:00
• 258d046218 More robust way of initializing empty .git folder oobabooga 2023-10-20 23:13:09 -07:00
• 1d5a015ce7
  Enable special token support for exllamav2 (#4314) Johan 2023-10-21 06:54:06 +02:00
• 8f6405d2fa
  Python 3.11, 3.9, 3.8 support (#4233) mjbogusz 2023-10-21 02:13:33 +02:00
• 9be74fb57c Change 2 margins oobabooga 2023-10-20 14:04:14 -07:00
• e208128d68 Lint the CSS files oobabooga 2023-10-20 13:02:18 -07:00
• dedbdb46c2 Chat CSS improvements oobabooga 2023-10-20 12:30:01 -07:00
• 32984ea2f0
  Support LLaVA v1.5 (#4305) Haotian Liu 2023-10-20 00:28:14 -05:00
• bb71272903 Detect WizardCoder-Python-34B & Phind-CodeLlama-34B oobabooga 2023-10-19 14:35:56 -07:00
• eda7126b25 Organize the .gitignore oobabooga 2023-10-19 14:33:44 -07:00
• ae8cd449ae
  ExLlamav2_HF: Convert logits to FP32 (#4310) turboderp 2023-10-19 04:16:05 +02:00
• c0ffb77fd8
  More silero languages (#3950) missionfloyd 2023-10-16 14:12:32 -06:00
• db7ecdd274
  openai: fix empty models list on query present in url (#4139) hronoas 2023-10-16 23:02:47 +03:00
• f17f7a6913 Increase the evaluation table height oobabooga 2023-10-16 12:55:35 -07:00
• 8ea554bc19 Check for torch.xpu.is_available() oobabooga 2023-10-16 12:53:40 -07:00
• 188d20e9e5 Reduce the evaluation table height oobabooga 2023-10-16 10:53:42 -07:00
• 2d44adbb76 Clear the torch cache while evaluating oobabooga 2023-10-16 10:52:50 -07:00
• 388d1864a6 Merge remote-tracking branch 'refs/remotes/origin/main' oobabooga 2023-10-15 21:58:16 -07:00
• 71cac7a1b2 Increase the height of the evaluation table oobabooga 2023-10-15 21:56:40 -07:00
• e14bde4946 Minor improvements to evaluation logs oobabooga 2023-10-15 20:51:43 -07:00
• b88b2b74a6 Experimental Intel Arc transformers support (untested) oobabooga 2023-10-15 20:51:11 -07:00
• d331501ebc
  Fix for using Torch with CUDA 11.8 (#4298) Sam 2023-10-16 08:27:19 +10:00
• 3bb4046fad
  Update auto-release.yml oobabooga 2023-10-15 17:27:16 -03:00
• 45fa803943
  Create auto-release.yml oobabooga 2023-10-15 17:25:29 -03:00
• 2706394bfe
  Relax numpy version requirements (#4291) Johan 2023-10-15 17:05:06 +02:00
• 8cce1f1126
  Exllamav2 lora support (#4229) Forkoz 2023-10-14 19:12:41 +00:00
• 1f5a2c5597
  Use Pytorch 2.1 exllama wheels (#4285) jllllll 2023-10-14 13:27:59 -05:00
• cd1cad1b47 Bump exllamav2 oobabooga 2023-10-14 11:23:07 -07:00
• 6e2dec82f1
  add chatml support + mistral-openorca (#4275) Eve 2023-10-13 14:49:17 +00:00
• ed66ca3cdf
  Add HTTPS support to APIs (openai and default) (#4270) Jesus Alvarez 2023-10-12 21:31:13 -07:00
• 43be1be598 Manually install CUDA runtime libraries oobabooga 2023-10-12 21:02:44 -07:00
• faf5c4dd58 Fix code blocks in instruct mode oobabooga 2023-10-11 12:18:46 -07:00
• 773c17faec Fix a warning oobabooga 2023-10-10 20:53:38 -07:00
• f63361568c Fix safetensors kwarg usage in AutoAWQ oobabooga 2023-10-10 19:03:09 -07:00
• 39f16ff83d Fix default/notebook tabs css oobabooga 2023-10-10 18:45:12 -07:00
• fae8062d39
  Bump to latest gradio (3.47) (#4258) oobabooga 2023-10-10 22:20:49 -03:00
• 2b75d725e6
  Initial support for LLaVA-LLaMA-2. (#3377) Haotian Liu 2023-10-10 16:40:52 -05:00
• 9fab9a1ca6 Minor fix oobabooga 2023-10-10 14:08:11 -07:00
• a49cc69a4a Ignore rope_freq_base if value is 10000 oobabooga 2023-10-10 13:57:40 -07:00
• 3a9d90c3a1 Download models with 4 threads by default oobabooga 2023-10-10 13:52:10 -07:00
• 520cbb2ab1
  Bump safetensors from 0.3.2 to 0.4.0 (#4249) dependabot[bot] 2023-10-10 17:41:09 -03:00
• 35695e18c7
  Remove import. (#4247) Forkoz 2023-10-09 21:06:11 +00:00
• 2e471071af
  Update llama_attn_hijack.py (#4231) Forkoz 2023-10-08 18:16:48 +00:00
• 2e8b5f7c80
  Update ROCm command oobabooga 2023-10-08 10:12:13 -03:00