-
0.2.7 Stable
released this
2023-06-11 11:28:31 -04:00 | 90 commits to main since this releaseChanges
- Add support for StarChat special tokens
- Add
context_lengthparameter support for MPT - Add low-level API to get
eos_token_id,vocab_size,context_length
-
0.2.5 Stable
released this
2023-06-02 09:18:51 -04:00 | 97 commits to main since this releaseChanges
- Add
local_files_onlyoption - Fix utf-8 decode error
- Add
-
0.2.4 Stable
released this
2023-05-31 15:03:38 -04:00 | 100 commits to main since this releaseChanges
- Use smallest model file in a Hugging Face Hub model repo if
model_fileis not specified
- Use smallest model file in a Hugging Face Hub model repo if
-
0.2.3 Stable
released this
2023-05-30 19:10:50 -04:00 | 103 commits to main since this releaseChanges
- Add cuBLAS support for LLaMA models
- Simplify installing and building from source
-
0.2.2 Stable
released this
2023-05-28 07:59:25 -04:00 | 112 commits to main since this releaseChanges
- Add low-level API for
logits - Add embeddings API for LLaMA models
- Add
context_lengthparameter for LLaMA models
- Add low-level API for
-
0.2.1 Stable
released this
2023-05-25 12:30:44 -04:00 | 117 commits to main since this releaseChanges
- Fix
threadsparameter
- Fix
-
0.2.0 Stable
released this
2023-05-20 21:59:18 -04:00 | 121 commits to main since this releaseChanges
- Add support for LLaMA, MPT models
- Change
repetition_penaltydefault value from1.0to1.1
Breaking Changes
- Update GGML library which has breaking changes to quantization formats. Old models have to be re-quantized.
Some of the latest quantized models are available here.
-
0.1.2 Stable
released this
2023-05-19 17:43:55 -04:00 | 130 commits to main since this releaseChanges
- Add
streamoption
- Add
-
0.1.1 Stable
released this
2023-05-18 16:22:33 -04:00 | 133 commits to main since this releaseChanges
- Add
stopoption
- Add