marella/ctransformers
1
0
Fork 0
Code Issues 107 Pull requests 5 Projects Releases 30 Packages Wiki Activity Actions
30 releases 30 tags
  • v0.2.27 ed02cf4b93

    0.2.27 Stable

    marella released this 2023-09-10 11:13:26 -04:00 | 0 commits to main since this release

    Changes

    • Skip evaluating tokens that are evaluated in the past. This can significantly speed up prompt processing in chat applications that prepend previous messages to prompt.
    • Deprecate LLM.reset() method. Use high-level API instead.
    • Add support for batching and beam search to 🤗 model.
    • Remove universal binary option when building for AVX2, AVX on macOS.
  • v0.2.26 744c5e29dd

    0.2.26 Stable

    marella released this 2023-08-30 17:31:25 -04:00 | 7 commits to main since this release

    Changes

    • Add support for 🤗 Transformers
  • v0.2.25 5f02372a64

    0.2.25 Stable

    marella released this 2023-08-28 20:14:57 -04:00 | 9 commits to main since this release

    Changes

    • Add support for GGUF v2
    • Add CUDA support for Falcon GGUF models
    • Add ROCm support
    • Add low-level API for add_bos_token, bos_token_id
  • v0.2.24 00eefe1551

    0.2.24 Stable

    marella released this 2023-08-24 19:34:51 -04:00 | 15 commits to main since this release

    Changes

    • Add GGUF format support for Llama and Falcon models
    • Add support for Code Llama models
  • v0.2.23 bf56961408

    0.2.23 Stable

    marella released this 2023-08-20 15:16:31 -04:00 | 20 commits to main since this release

    Changes

    • Add mmap and mlock parameters for LLaMA and Falcon models
    • Add revision option for models on Hugging Face Hub
  • v0.2.22 aa38adef26

    0.2.22 Stable

    marella released this 2023-08-12 11:17:19 -04:00 | 24 commits to main since this release

    Changes

    • Add experimental CUDA support for StarCoder, StarChat models
    • Add gpt_bigcode as model type for StarCoder, StarChat models
    • Fix loading GPTQ models from a local path
  • v0.2.21 a2a5384ea0

    0.2.21 Stable

    marella released this 2023-08-07 14:56:47 -04:00 | 31 commits to main since this release

    Changes

    • Simplify CUDA installation by using precompiled runtime libraries from NVIDIA
  • v0.2.20 5be9eccc43

    0.2.20 Stable

    marella released this 2023-08-05 14:36:15 -04:00 | 34 commits to main since this release

    Changes

    • Add experimental CUDA support for MPT models
  • v0.2.19 561cd5d293

    0.2.19 Stable

    marella released this 2023-08-04 18:29:47 -04:00 | 38 commits to main since this release

    Changes

    • Add Metal support for LLaMA 2 70B models
    • Update llama.cpp
  • v0.2.18 bb4e134a8f

    0.2.18 Stable

    marella released this 2023-08-02 16:07:48 -04:00 | 42 commits to main since this release

    Changes

    • Add experimental support for GPTQ models using ExLlama