Commits

Branch selector

 main

User selector

marella

Datepicker

All time

Commit History

Commits on Sep 10, 2023

• 0.2.27

  marella

  committed
  ed02cf4

  Copy full SHA for ed02cf4

• Remove universal binary option when building for AVX2, AVX on macOS

  marella

  committed
  271c565

  Copy full SHA for 271c565

• Improve error messages for Metal

  marella

  committed
  12dc4ce

  Copy full SHA for 12dc4ce

• Add support for batching and beam search to 🤗 model

  marella

  committed
  98dd42a

  Copy full SHA for 98dd42a

• Skip evaluating tokens that are evaluated in the past

  Show description for f9ec5ca
  marella

  committed
  f9ec5ca

  Copy full SHA for f9ec5ca

Commits on Sep 9, 2023

• Store past tokens in Python instead of C++

  marella

  committed
  59771f7

  Copy full SHA for 59771f7

Commits on Aug 30, 2023

• 0.2.26

  marella

  committed
  744c5e2

  Copy full SHA for 744c5e2

• Add support for 🤗 Transformers

  marella

  committed
  4967122

  Copy full SHA for 4967122

Commits on Aug 29, 2023

• 0.2.25

  marella

  committed
  5f02372

  Copy full SHA for 5f02372

Commits on Aug 28, 2023

• Add ROCm support

  marella

  committed
  163709e

  Copy full SHA for 163709e

• Fix detokenize()

  marella

  committed
  a71db2c

  Copy full SHA for a71db2c

• Update llama.cpp

  marella

  committed
  964c251

  Copy full SHA for 964c251

• Add low-level API for add_bos_token, bos_token_id

  marella

  committed
  4f1c61c

  Copy full SHA for 4f1c61c

Commits on Aug 24, 2023

• 0.2.24

  marella

  committed
  00eefe1

  Copy full SHA for 00eefe1

• Update README

  marella

  committed
  2d11491

  Copy full SHA for 2d11491

• Fix Falcon GGML

  marella

  committed
  80726b5

  Copy full SHA for 80726b5

• Add GGUF format support for Llama and Falcon models

  marella

  committed
  c5bc93f

  Copy full SHA for c5bc93f

Commits on Aug 20, 2023

• 0.2.23

  marella

  committed
  bf56961

  Copy full SHA for bf56961

• Add revision option for models on Hugging Face Hub

  marella

  committed
  d28c5c2

  Copy full SHA for d28c5c2

Commits on Aug 15, 2023

• Add mmap and mlock parameters for LLaMA and Falcon models

  marella

  committed
  bef7eef

  Copy full SHA for bef7eef

Commits on Aug 12, 2023

• 0.2.22

  marella

  committed
  aa38ade

  Copy full SHA for aa38ade

• Update docs

  marella

  committed
  4c7a1a5

  Copy full SHA for 4c7a1a5

• Add experimental CUDA support for StarCoder, StarChat models

  Show description for 7fc3407
  marella

  committed
  7fc3407

  Copy full SHA for 7fc3407

Commits on Aug 9, 2023

• Disable debug logging for Metal

  marella

  committed
  bb4a5ce

  Copy full SHA for bb4a5ce

Commits on Aug 7, 2023

• 0.2.21

  marella

  committed
  a2a5384

  Copy full SHA for a2a5384

• Update README

  marella

  committed
  0540dce

  Copy full SHA for 0540dce

• Use precompiled CUDA runtime libraries from NVIDIA

  marella

  committed
  2bf4fa7

  Copy full SHA for 2bf4fa7

Commits on Aug 5, 2023

• 0.2.20

  marella

  committed
  5be9ecc

  Copy full SHA for 5be9ecc

• Update docs

  marella

  committed
  c8c9dcf

  Copy full SHA for c8c9dcf

• Add experimental CUDA support for MPT models

  marella

  committed
  1477a4b

  Copy full SHA for 1477a4b

Commits on Aug 4, 2023

• 0.2.19

  marella

  committed
  561cd5d

  Copy full SHA for 561cd5d

• Add GitHub Action to free disk space on Ubuntu for CUDA

  marella

  committed
  64717d5

  Copy full SHA for 64717d5

• Update llama.cpp

  marella

  committed
  4201ae0

  Copy full SHA for 4201ae0

Commits on Aug 2, 2023

• 0.2.18

  marella

  committed
  bb4e134

  Copy full SHA for bb4e134

• Add experimental support for GPTQ models using ExLlama

  marella

  committed
  315ff40

  Copy full SHA for 315ff40

Pagination

Previous

Next