snapshot-2023-12-31

What's Changed

• Reset the model_name after unloading the model by Api by @Yiximail in #5051
• Synthia instruction templates by @StefanDanielSchwarz in #5041
• Support for sending images into OpenAI chat API by @kabachuha in #4827
• fix: function [get_reply_from_output_ids]: by @zhangningboo in #5045
• Bump AutoAWQ to 0.1.8 by @casper-hansen in #5061
• Merge dev branch by @oobabooga in #5073
• Bump llama-cpp-python to 0.2.25 by @oobabooga in #5077
• Merge dev branch by @oobabooga in #5078
• Improve chat styles by @TheLounger in #5044
• Merge dev branch by @oobabooga in #5100
• Specify utf-8 encoding for the model settings file open by @B611 in #5125
• Remove --sdp-attention, --xformers flags by @oobabooga in #5126
• Remove exllamav1 loaders by @oobabooga in #5128
• Remove RWKV loader by @oobabooga in #5130
• Superboogav2 Quick Fixes by @TheInvisibleMage in #5089
• Remove deprecated command-line flags by @oobabooga in #5131
• Merge dev branch by @oobabooga in #5132

New Contributors

• @zhangningboo made their first contribution in #5045
• @B611 made their first contribution in #5125
• @TheInvisibleMage made their first contribution in #5089

Full Changelog: snapshot-2023-12-24...snapshot-2023-12-31

snapshot-2023-12-24

What's Changed

• Update Training PRO by @FartyPants in #4972
• Add Orca-Vicuna instruction template by @netrunnereve in #4971
• Add detection for Eric Hartford's Dolphin models in models/config.yaml by @HiroseKoichi in #4966
• llama.cpp: read instruction template from GGUF metadata by @oobabooga in #4975
• Merge dev branch by @oobabooga in #4980
• Add HQQ quant loader for ooba by @waters222 in #4888
• Bump exllamav2 to 0.0.11 and add whl by @yhyu13 in #4973
• Update optimum requirement from ==1.15.* to ==1.16.* by @dependabot in #4986
• Merge dev branch by @oobabooga in #4988
• Bump llama-cpp-python to 0.2.24 by @oobabooga in #5001
• Merge dev branch by @oobabooga in #5002
• Add llama-cpp-python wheels with tensor cores support by @oobabooga in #5003
• Merge dev branch by @oobabooga in #5005
• Merge dev branch by @oobabooga in #5011
• Merge dev branch by @oobabooga in #5012
• let exllama v1 models load safetensor loras by @lun-4 in #4854
• Merge dev branch by @oobabooga in #5022
• Organize the CMD arguments by group by @oobabooga in #5027
• Merge dev branch by @oobabooga in #5039

New Contributors

• @HiroseKoichi made their first contribution in #4966
• @waters222 made their first contribution in #4888
• @lun-4 made their first contribution in #4854

Full Changelog: snapshot-2023-12-17...snapshot-2023-12-24

snapshot-2023-12-17

What's Changed

• Make new browser reloads recognize current model (fixes #4843) by @cyberfox in #4865
• Update callbacks.py by @Nehereus in #4892
• Update transformers requirement from ==4.35.* to ==4.36.* by @dependabot in #4882
• Update peft requirement from ==0.6.* to ==0.7.* by @dependabot in #4886
• Bump optimum from 1.14.0 to 1.15.0 by @dependabot in #4885
• Jinja templates for Instruct and Chat by @oobabooga in #4874
• Updated Docker Docs by @Penagwin in #4900
• Merge dev branch by @oobabooga in #4903
• Bypass coqui TTS EULA check by @missionfloyd in #4905
• Fixed invalid Jinja2 syntax in instruction templates by @illixion in #4911
• Merge dev branch by @oobabooga in #4912
• Allow symlinked folder within root directory by @noneabove1182 in #4863
• Fix deleting chat logs by @TheLounger in #4914
• Merge dev branch by @oobabooga in #4920
• Merge dev branch by @oobabooga in #4923
• Bump llama-cpp-python to 0.2.23 (NVIDIA & CPU-only, no AMD, no Metal) by @oobabooga in #4924
• Merge dev branch by @oobabooga in #4927
• Remove the elevenlabs extension by @oobabooga in #4928
• [OpenAI Extension] Add 'max_logits' parameter in logits endpoint by @kimjaewon96 in #4916
• [OpenAI Extension] Embeddings Endpoint by @Heimlock in #4895
• Bug fix: when generation fails, save the sent message by @oobabooga in #4915
• Bump llama-cpp-python to 0.2.23 (including Linux ROCm and MacOS >= 12) by @oobabooga in #4930
• Update IPEX to 2.1.10+xpu by @notsyncing in #4931
• Merge dev branch by @oobabooga in #4937
• Instruct style improvements by @oobabooga in #4951
• Better HF grammar implementation by @oobabooga in #4953
• Add --num_experts_per_token parameter (ExLlamav2) by @oobabooga in #4955
• Merge dev branch by @oobabooga in #4961

New Contributors

• @Nehereus made their first contribution in #4892
• @Penagwin made their first contribution in #4900
• @illixion made their first contribution in #4911
• @noneabove1182 made their first contribution in #4863
• @kimjaewon96 made their first contribution in #4916
• @Heimlock made their first contribution in #4895

Full Changelog: snapshot-2023-12-10...snapshot-2023-12-17

snapshot-2023-12-10

What's Changed

• Gallery improvements by @TheLounger in #4789
• Fix typo in README.md by @eltociear in #4793
• Bump safetensors from 0.4.0 to 0.4.1 by @dependabot in #4750
• Import accelerate very early to make Intel GPU happy by @notsyncing in #4704
• Merge dev branch by @oobabooga in #4807
• Update accelerate requirement from ==0.24.* to ==0.25.* by @dependabot in #4810
• Bump exllamav2 from 0.0.8 to 0.0.10 & Fix code change by @yhyu13 in #4782
• Merge dev branch by @oobabooga in #4811
• Optimize HF text generation by @oobabooga in #4814
• Merge dev branch by @oobabooga in #4815
• Clean-up Ctrl+C Shutdown by @erew123 in #4802
• Add QuIP# support by @oobabooga in #4803
• Update IPEX install URL by @notsyncing in #4825
• Parameters: change max_new_tokens & repetition_penalty_range defaults by @oobabooga in #4842
• Fix partial unicode characters issue by @Yiximail in #4837
• Merge dev branch by @oobabooga in #4849
• Merge dev branch by @oobabooga in #4851

New Contributors

• @TheLounger made their first contribution in #4789
• @notsyncing made their first contribution in #4704
• @erew123 made their first contribution in #4802

Full Changelog: snapshot-2023-12-03...snapshot-2023-12-10

snapshot-2023-12-03

What's Changed

• add openhermes mistral support by @netrunnereve in #4730
• Fix detection of stopping strings by @tsukanov-as in #4728
• resonable cli args for docker container by @xr4dsh in #4727
• Add direnv artifacts to gitignore by @den-is in #4737
• Fixed naming for sentence-transformers library by @manukashyap in #4764
• feature/docker_improvements by @Callum17 in #4768
• Merge dev branch by @oobabooga in #4773
• Merge dev branch by @oobabooga in #4777

New Contributors

• @tsukanov-as made their first contribution in #4728
• @xr4dsh made their first contribution in #4727
• @den-is made their first contribution in #4737
• @manukashyap made their first contribution in #4764
• @Callum17 made their first contribution in #4768

Full Changelog: snapshot-2023-11-26...snapshot-2023-12-03

snapshot-2023-11-26

What's Changed

• Bugfix/aenglema/issue#4665 by @drew9781 in #4670
• Fix "Illegal instruction" bug in llama.cpp CPU only version by @oobabooga in #4677
• add XTTSv2 by @kanttouchthis in #4673
• Merge dev branch by @oobabooga in #4683
• Merge dev branch by @oobabooga in #4686
• Detect Orca 2 template by @oobabooga in #4697
• Bump to flash-attention 2.3.4 + switch to Github Actions wheels on Windows by @oobabooga in #4700
• Bump llama-cpp-python to 0.2.19 & add min_p and typical_p parameters to llama.cpp loader by @oobabooga in #4701
• Merge dev branch by @oobabooga in #4702

New Contributors

• @drew9781 made their first contribution in #4670
• @kanttouchthis made their first contribution in #4673

Full Changelog: snapshot-2023-11-19...snapshot-2023-11-26

snapshot-2023-11-19

What's Changed

• Fix bug with /internal/model/load by @GuizzyQC in #4549
• Merge dev branch by @oobabooga in #4579
• Merge dev branch by @oobabooga in #4606
• Merge dev branch by @oobabooga in #4608
• update AutoGPTQ to higher version for lora applying error fixing by @rogozinushka in #4604
• Fix CPU memory limit error (issue #3763) by @null-dev in #4597
• Use standard hyphens in filenames by @naptastic in #4576
• Bump AutoAWQ to 0.1.7 by @casper-hansen in #4620
• Bump llama-cpp-python to 0.2.18 by @oobabooga in #4611
• Merge dev branch by @oobabooga in #4627
• Merge dev branch by @oobabooga in #4628
• Merge dev branch by @oobabooga in #4632
• Bump llama-cpp-python to 0.2.18 (2nd attempt) by @oobabooga in #4637
• New feature: "random preset" button by @oobabooga in #4647
• Use system message in chat instruct by @jordanbtucker in #4648
• Openai embedding fix to support jina-embeddings-v2 by @wizd in #4642
• Add --admin-key flag for API by @oobabooga in #4649
• Add /v1/internal/logits endpoint by @oobabooga in #4650
• Add --nowebui flag for pure API mode by @oobabooga in #4651
• Add /v1/internal/lora endpoints by @oobabooga in #4652
• Use shared chat-instruct_command with api by @jordanbtucker in #4653
• New feature: enlarge character pictures on click by @oobabooga in #4654
• Merge dev branch by @oobabooga in #4660
• Merge dev branch by @oobabooga in #4662
• Merge dev branch by @oobabooga in #4664

New Contributors

• @rogozinushka made their first contribution in #4604
• @null-dev made their first contribution in #4597
• @naptastic made their first contribution in #4576
• @jordanbtucker made their first contribution in #4648

Full Changelog: snapshot-2023-11-12...snapshot-2023-11-19

snapshot-2023-11-12

What's Changed

• [Fix] fix openai embedding_model loading as str by @yhyu13 in #4147
• Make OpenAI API the default API by @oobabooga in #4430
• Merge dev branch by @oobabooga in #4488
• Add /v1/internal/stop-generation to OpenAI API by @oobabooga in #4498
• Update peft requirement from ==0.5.* to ==0.6.* by @dependabot in #4494
• Bump optimum from 1.13.1 to 1.14.0 by @dependabot in #4492
• Update 12 - OpenAI API.md by @mocheng in #4501
• Separate context and system message in instruction formats by @oobabooga in #4499
• fix openai frequency_penalty type in requests by @hronoas in #4512
• Include trust remote code usage in openai api's embedder by @MrMojoR in #4513
• Merge dev branch by @oobabooga in #4522
• Merge dev branch by @oobabooga in #4532
• Fix deprecated API by @ashleykleynhans in #4539
• Merge dev branch by @oobabooga in #4541

New Contributors

• @yhyu13 made their first contribution in #4147
• @mocheng made their first contribution in #4501
• @MrMojoR made their first contribution in #4513
• @ashleykleynhans made their first contribution in #4539

Full Changelog: snapshot-2023-11-05...snapshot-2023-11-12

snapshot-2023-11-05

What's Changed

• updated wiki link by @senadev42 in #4415
• Bump AutoAWQ to v0.1.5 by @casper-hansen in #4410
• Bump exllamav2 version to 0.0.7 by @Soefati in #4417
• Bugfix: Updating the shared settings object when loading a model by @ziadloo in #4425
• [Fix] OpenOrca-Platypus2 models should use correct instruction_template when matching against models/config.yaml by @deevis in #4435
• make torch.load a bit safer by @julien-c in #4448
• transformers: Add a flag to force load from safetensors by @julien-c in #4450
• Implement Min P as a sampler option in HF loaders by @kalomaze in #4449
• Add temperature_last parameter by @oobabooga in #4472
• Bump AWQ to 0.1.6 by @casper-hansen in #4470
• fixed two links in the ui by @wvanderp in #4452
• add use_flash_attention_2 to param for Model loader Transformers by @fenglui in #4373
• Bump transformers to 4.35.* by @Soefati in #4474
• Merge dev branch by @oobabooga in #4475
• Merge dev branch by @oobabooga in #4476
• Fix openai extension not working because of absent new defaults by @kabachuha in #4477

New Contributors

• @senadev42 made their first contribution in #4415
• @Soefati made their first contribution in #4417
• @ziadloo made their first contribution in #4425
• @deevis made their first contribution in #4435
• @julien-c made their first contribution in #4448
• @wvanderp made their first contribution in #4452
• @fenglui made their first contribution in #4373

Full Changelog: snapshot-2023-10-29...snapshot-2023-11-05

snapshot-2023-10-29

What's Changed

• Add additive_repetition_penalty sampler setting. by @tdrussell in #3627
• Fix training.py tutorial url by @adrianfiedler in #4367
• Rename additive_repetition_penalty to presence_penalty, add frequency_penalty by @tdrussell in #4376
• Replace hashlib.sha256 with hashlib.file_digest, so we don't need to load entire files into ram before hashing them. by @LightningDragon in #4383
• Fix Gradio warning message regarding custom value by @GuizzyQC in #4391
• Intel Gpu support initialization by @abhilash1910 in #4340
• Update accelerate requirement from ==0.23.* to ==0.24.* by @dependabot in #4400
• Adding platform_system to autoawq by @jamesbraza in #4390

New Contributors

• @adrianfiedler made their first contribution in #4367
• @LightningDragon made their first contribution in #4383
• @abhilash1910 made their first contribution in #4340

Full Changelog: snapshot-2023-10-22...snapshot-2023-10-29