What's Changed
- Bump exllamav2 to 0.0.13.1 by @oobabooga in #5463
- Update main.css to allow scrolling in code blocks by @tseven in #5495
- Bump hqq from 0.1.2.post1 to 0.1.3 by @dependabot in #5489
- Update numpy requirement from ==1.24.* to ==1.26.* by @dependabot in #5490
- Update peft requirement from ==0.7.* to ==0.8.* by @dependabot in #5446
- Merge dev branch by @oobabooga in #5496
- llamacpp_HF: do not use oobabooga/llama-tokenizer by @oobabooga in #5499
- Merge dev branch by @oobabooga in #5502
- Bump gradio to 4.19 by @DominikKowalczyk in #5419
- Remove "Maximum UI updates/second" parameter by @oobabooga in #5507
- Revert gradio to 3.50.2 by @oobabooga in #5513
- Add a "llamacpp_HF creator" menu by @oobabooga in #5519
- Downloader: start a new session for each file by @oobabooga in #5520
- Add a menu for customizing the instruction template for the model by @oobabooga in #5521
- Add --autosplit flag for ExLlamaV2 by @oobabooga in #5524
- Removed extra spaces from Mistral instruction template that were causing model to misbehave by @fschuh in #5517
- Merge dev branch by @oobabooga in #5530
- Merge dev branch by @oobabooga in #5534
New Contributors
- @tseven made their first contribution in #5495
- @DominikKowalczyk made their first contribution in #5419
- @fschuh made their first contribution in #5517
Full Changelog: snapshot-2024-02-11...snapshot-2024-02-18