Commit graph

Select branches

Hide pull requests

main

Mono Color

• d06c34dea5
  Add an extension that makes chat replies longer (#3363) oobabooga 2023-07-31 13:34:41 -03:00
• e6be25ea11 Fix a regression oobabooga 2023-07-30 18:12:30 -07:00
• 5ca37765d3 Only replace {{user}} and {{char}} at generation time oobabooga 2023-07-30 11:42:30 -07:00
• 6e16af34fd Save uploaded characters as yaml oobabooga 2023-07-30 11:25:38 -07:00
• c25602eb65 Merge branch 'dev' oobabooga 2023-07-30 08:47:50 -07:00
• ca4188aabc Update the example extension oobabooga 2023-07-29 18:57:22 -07:00
• c4e14a757c
  Bump exllama module to 0.0.9 (#3338) jllllll 2023-07-29 20:16:23 -05:00
• ecd92d6a4e
  Remove unused variable from ROCm GPTQ install (#107) jllllll 2023-07-26 20:16:36 -05:00
• 1e3c950c7d
  Add AMD GPU support for Linux (#98) jllllll 2023-07-26 15:33:02 -05:00
• 4b37a2b397
  sd_api_pictures: Widen sliders for image size minimum and maximum (#3326) GuizzyQC 2023-07-26 12:49:46 -04:00
• d6314fd539 Change a comment oobabooga 2023-07-26 09:37:48 -07:00
• f24f87cfb0 Change a comment oobabooga 2023-07-26 09:37:48 -07:00
• de5de045e0 Set rms_norm_eps to 5e-6 for every llama-2 ggml model, not just 70b oobabooga 2023-07-26 08:23:24 -07:00
• 193c6be39c Add missing \n to llama-v2 template context oobabooga 2023-07-26 07:59:40 -07:00
• ec68d5211e Set rms_norm_eps to 5e-6 for every llama-2 ggml model, not just 70b oobabooga 2023-07-26 08:23:24 -07:00
• a9e10753df Add missing \n to llama-v2 template context oobabooga 2023-07-26 07:59:40 -07:00
• b780d520d2 Add a link to the gradio docs oobabooga 2023-07-26 07:49:22 -07:00
• b553c33dd0 Add a link to the gradio docs oobabooga 2023-07-26 07:49:22 -07:00
• d94ba6e68b Define visible_text before applying chat_input extensions oobabooga 2023-07-26 07:26:37 -07:00
• b31321c779 Define visible_text before applying chat_input extensions oobabooga 2023-07-26 07:26:37 -07:00
• b17893a58f Revert "Add tensor split support for llama.cpp (#3171)" oobabooga 2023-07-26 07:06:01 -07:00
• 517d40cffe Update Extensions.md oobabooga 2023-07-26 07:01:35 -07:00
• b11f63cb18 update extensions docs oobabooga 2023-07-26 07:00:33 -07:00
• 52e3b91f5e
  Fix broken gxx_linux-64 package. (#106) jllllll 2023-07-25 23:55:08 -05:00
• 4a24849715 Revert changes oobabooga 2023-07-25 21:09:32 -07:00
• 69f8b35bc9 Revert changes to README oobabooga 2023-07-25 20:49:00 -07:00
• ed80a2e7db Reorder llama.cpp params oobabooga 2023-07-25 20:45:20 -07:00
• 0e8782df03 Set instruction template when switching from default/notebook to chat oobabooga 2023-07-25 20:37:01 -07:00
• 28779cd959 Use dark theme by default oobabooga 2023-07-25 20:11:57 -07:00
• c2e0d46616 Add credits oobabooga 2023-07-25 15:49:04 -07:00
• 1b89c304ad Update README oobabooga 2023-07-25 15:46:12 -07:00
• d3abe7caa8 Update llama.cpp.md oobabooga 2023-07-25 15:33:16 -07:00
• 863d2f118f Update llama.cpp.md oobabooga 2023-07-25 15:31:05 -07:00
• 77d2e9f060 Remove flexgen 2 oobabooga 2023-07-25 15:18:25 -07:00
• 75c2dd38cf Remove flexgen support oobabooga 2023-07-25 15:15:29 -07:00
• 5134d5b1c6 Update README oobabooga 2023-07-25 15:13:07 -07:00
• 85b3a26e25
  Ignore values which are not string in training.py (#3287) Foxtr0t1337 2023-07-26 06:00:25 +08:00
• 031fe7225e
  Add tensor split support for llama.cpp (#3171) Shouyi 2023-07-26 07:59:26 +10:00
• f653546484
  README updates and improvements (#3198) Eve 2023-07-25 17:58:13 -04:00
• b09e4f10fd
  Fix typo in README.md (#3286) Ikko Eltociear Ashimine 2023-07-26 06:56:25 +09:00
• 7bc408b472 Change rms_norm_eps to 5e-6 for llama-2-70b ggml oobabooga 2023-07-25 14:54:57 -07:00
• ef8637e32d
  Add extension example, replace input_hijack with chat_input_modifier (#3307) oobabooga 2023-07-25 18:49:56 -03:00
• 08c622df2e Autodetect rms_norm_eps and n_gqa for llama-2-70b oobabooga 2023-07-24 15:26:29 -07:00
• a07d070b6c
  Add llama-2-70b GGML support (#3285) oobabooga 2023-07-24 16:37:03 -03:00
• 6f4830b4d3 Bump peft commit oobabooga 2023-07-24 09:49:57 -07:00
• 90a4ab631c
  extensions/openai: Fixes for: embeddings, tokens, better errors. +Docs update, +Images, +logit_bias/logprobs, +more. (#3122) matatonic 2023-07-24 10:28:12 -04:00
• 1141987a0d
  Add checks for ROCm and unsupported architectures to llama_cpp_cuda loading (#3225) jllllll 2023-07-24 09:25:36 -05:00
• 74fc5dd873
  Add user-agent to download-model.py requests (#3243) iongpt 2023-07-24 16:19:13 +02:00
• b2d5433409
  Fix typo in deepspeed_parameters.py (#3222) Ikko Eltociear Ashimine 2023-07-24 23:17:28 +09:00
• eb105b0495
  Bump llama-cpp-python to 0.1.74 (#3257) jllllll 2023-07-24 09:15:42 -05:00
• 152cf1e8ef
  Bump bitsandbytes to 0.41.0 (#3258) jllllll 2023-07-24 09:06:18 -05:00
• 8d31d20c9a
  Bump exllama module to 0.0.8 (#3256) jllllll 2023-07-24 09:05:54 -05:00
• cc2ed46d44
  Make chat the default again oobabooga 2023-07-20 18:55:09 -03:00
• fcb215fed5
  Add check for compute support for GPTQ-for-LLaMa (#104) jllllll 2023-07-20 09:11:00 -05:00
• 63ece46213 Merge branch 'main' into dev oobabooga 2023-07-20 07:06:41 -07:00
• 6415cc68a2 Remove obsolete information from README oobabooga 2023-07-19 21:20:40 -07:00
• 4b19b74e6c Add CUDA wheels for llama-cpp-python by jllllll oobabooga 2023-07-19 19:31:19 -07:00
• 05f4cc63c8 Merge branch 'main' into dev oobabooga 2023-07-19 19:22:34 -07:00
• 4df3f72753
  Fix GPTQ fail message not being shown on update (#103) jllllll 2023-07-19 20:25:09 -05:00
• 87926d033d
  Bump exllama module to 0.0.7 (#3211) jllllll 2023-07-19 20:24:47 -05:00
• 913e060348 Change the default preset to Divine Intellect oobabooga 2023-07-19 08:24:37 -07:00
• 0d7f43225f Merge branch 'dev' oobabooga 2023-07-19 07:20:13 -07:00
• 08c23b62c7 Bump llama-cpp-python and transformers oobabooga 2023-07-19 07:19:12 -07:00
• 5447e75191 Merge branch 'dev' oobabooga 2023-07-18 15:36:26 -07:00
• 8ec225f245 Add EOS/BOS tokens to Llama-2 template oobabooga 2023-07-18 15:35:27 -07:00
• 3ef49397bb
  Merge pull request #3195 from oobabooga/dev oobabooga 2023-07-18 17:33:11 -03:00
• 070a886278 Revert "Prevent lists from flickering in chat mode while streaming" oobabooga 2023-07-18 13:23:29 -07:00
• a2918176ea Update LLaMA-v2-model.md (thanks Panchovix) oobabooga 2023-07-18 13:21:18 -07:00
• e0631e309f
  Create instruction template for Llama-v2 (#3194) oobabooga 2023-07-18 17:19:18 -03:00
• 603c596616 Add LLaMA-v2 conversion instructions oobabooga 2023-07-18 10:29:56 -07:00
• c535f14e5f
  Bump bitsandbytes Windows wheel to 0.40.2 (#3186) jllllll 2023-07-18 09:39:43 -05:00
• d7a14174a2
  Remove auto-loading when only one model is available (#3187) jllllll 2023-07-18 09:39:08 -05:00
• a69955377a
  [GGML] Support for customizable RoPE (#3083) randoentity 2023-07-18 03:32:37 +02:00
• 89e0d15cf5
  Use 'torch.backends.mps.is_available' to check if mps is supported (#3164) appe233 2023-07-18 08:27:18 +08:00
• 234c58ccd1
  Bump bitsandbytes from 0.40.1.post1 to 0.40.2 (#3178) dependabot[bot] 2023-07-17 21:24:51 -03:00
• 49a5389bd3
  Bump accelerate from 0.20.3 to 0.21.0 oobabooga 2023-07-17 21:23:59 -03:00
• 8c1c2e0fae Increase max_new_tokens upper limit oobabooga 2023-07-17 17:08:22 -07:00
• 5e5d926d2b Prevent lists from flickering in chat mode while streaming oobabooga 2023-07-17 17:00:49 -07:00
• 02a5fe6aa2
  Bump accelerate from 0.20.3 to 0.21.0 dependabot[bot] 2023-07-17 20:18:31 +00:00
• 60a3e70242 Update LLaMA links and info oobabooga 2023-07-17 12:51:01 -07:00
• f83fdb9270 Don't reset LoRA menu when loading a model oobabooga 2023-07-17 12:50:25 -07:00
• 4ce766414b Bump AutoGPTQ version oobabooga 2023-07-17 10:02:12 -07:00
• b1a6ea68dd Disable "autoload the model" by default oobabooga 2023-07-17 07:40:56 -07:00
• 656b457795 Add Airoboros-v1.2 template oobabooga 2023-07-17 07:27:42 -07:00
• a199f21799 Optimize llamacpp_hf a bit oobabooga 2023-07-16 20:49:48 -07:00
• 9f08038864
  Merge pull request #3163 from oobabooga/dev oobabooga 2023-07-16 02:43:18 -03:00
• 6a3edb0542 Clean up llamacpp_hf.py oobabooga 2023-07-15 22:40:55 -07:00
• 2de0cedce3 Fix reload screen color oobabooga 2023-07-15 22:39:39 -07:00
• 13449aa44d Decrease download timeout oobabooga 2023-07-15 22:30:08 -07:00
• 27a84b4e04 Make AutoGPTQ the default again oobabooga 2023-07-15 22:29:23 -07:00
• 5e3f7e00a9
  Create llamacpp_HF loader (#3062) oobabooga 2023-07-16 02:21:13 -03:00
• 7c4d4fc7d3
  Increase alpha value limit for NTK RoPE scaling for exllama/exllama_HF (#3149) Panchovix 2023-07-16 00:56:04 -04:00
• 780a2f2e16
  Bump llama cpp version (#3160) ofirkris 2023-07-16 07:54:56 +03:00
• ed3ffd212d
  Bump bitsandbytes to 0.40.1.post1 (#3156) jllllll 2023-07-15 23:53:32 -05:00
• 94dfcec237
  Make it possible to evaluate exllama perplexity (#3138) oobabooga 2023-07-16 01:52:55 -03:00
• 11a8fd1eb9
  Add cuBLAS llama-cpp-python wheel installation (#102) jllllll 2023-07-15 23:31:33 -05:00
• b284f2407d Make ExLlama_HF the new default for GPTQ oobabooga 2023-07-14 14:03:56 -07:00
• 32f12b8bbf
  Bump bitsandbytes Windows wheel to 0.40.0.post4 (#3135) jllllll 2023-07-13 15:32:37 -05:00
• 9800745db9
  Color tokens by probability and/or perplexity (#3078) SeanScripts 2023-07-13 15:30:22 -05:00
• 146e8b2a6c Merge remote-tracking branch 'refs/remotes/origin/dev' into dev oobabooga 2023-07-13 13:23:38 -07:00