Commits

Breadcrumbs

History for

onmaster

User selector

All users

All time

Commit History

Commits on Feb 3, 2026

• Version 4.7.0 (#2003)

  jordimas

  authored
  3d05d20

  Copy full SHA for 3d05d20

Commits on Jan 4, 2026

• T5Gemma conversion and inference (#1962)

  Show description for 4ebf2dc
  jordimas

  authored
  4ebf2dc

  Copy full SHA for 4ebf2dc

Commits on Dec 1, 2025

• Initial Qwen 3 support (#1943)

  jordimas

  authored
  10540dd

  Copy full SHA for 10540dd

Commits on Nov 29, 2025

• Gemma 3 support (text only) (#1936)

  jordimas

  authored
  e38043e

  Copy full SHA for e38043e

Commits on Apr 8, 2025

• Build wheels for Python 3.13 and drop support for 3.8 (#1858)

  Show description for c974ddc
  rinarakaki

  authored
  c974ddc

  Copy full SHA for c974ddc

Commits on Oct 10, 2024

• update doc AWQ quantization (#1795)

  minhthuc2502

  authored
  087d085

  Copy full SHA for 087d085

Commits on Jul 4, 2024

• Quantzation AWQ GEMM + GEMV (#1727)

  Show description for 39f48f2
  minhthuc2502

  authored
  39f48f2

  Copy full SHA for 39f48f2

Commits on Mar 11, 2024

• Bump version 4.1.0 (#1638)

  minhthuc2502

  authored
  27092e4

  Copy full SHA for 27092e4

Commits on Mar 5, 2024

• support tensor parallel (#1599)

  Show description for cec65f1
  minhthuc2502

  authored
  cec65f1

  Copy full SHA for cec65f1

Commits on Sep 15, 2023

• update install documentation (#1484)

  Show description for 0d1424d
  zxdvd

  authored
  0d1424d

  Copy full SHA for 0d1424d

Commits on Sep 7, 2023

• Update note in token streaming documentation

  guillaumekln

  committed
  feb6ed5

  Copy full SHA for feb6ed5

Commits on Aug 31, 2023

• Include XLM-RoBERTa in the list of supported models

  guillaumekln

  committed
  5c88c27

  Copy full SHA for 5c88c27

Commits on Aug 30, 2023

• Revise the documentation about multithreading and parallelism

  guillaumekln

  committed
  8d5f5c7

  Copy full SHA for 8d5f5c7

Commits on Aug 28, 2023

• Remove the default system prompt in the Llama 2 chat example (#1441)

  guillaumekln

  authored
  6a473ed

  Copy full SHA for 6a473ed

• Remove warning about the token type input not being implemented

  guillaumekln

  committed
  df46e21

  Copy full SHA for df46e21

Commits on Aug 4, 2023

• Implement DistilBERT converter (#1399)

  Show description for 6c0fcc6
  

  Natalia Segal

  and

  guillaumekln

  authored
  6c0fcc6

  Copy full SHA for 6c0fcc6

Commits on Aug 2, 2023

• Fix typo in decoding.md (#1397)

  Show description for b7b59f8
  eltociear

  authored
  b7b59f8

  Copy full SHA for b7b59f8

• By default, keep the same FP precision when converting models (#1389)

  Show description for b6daa04
  guillaumekln

  authored
  b6daa04

  Copy full SHA for b6daa04

• Update the documentation

  guillaumekln

  committed
  4c041c0

  Copy full SHA for 4c041c0

Commits on Jul 25, 2023

• Pass the system prompt in static_prompt in the Llama 2 example (#1364)

  guillaumekln

  authored
  16191ad

  Copy full SHA for 16191ad

Commits on Jul 20, 2023

• Stop the decoding when the tokens generator is closed (#1359)

  guillaumekln

  authored
  00f615f

  Copy full SHA for 00f615f

• Add an interactive example using Llama 2 (#1358)

  guillaumekln

  authored
  467fd48

  Copy full SHA for 467fd48

Commits on Jul 18, 2023

• Fix documentation in quantization.md

  guillaumekln

  committed
  9d8a1af

  Copy full SHA for 9d8a1af

Commits on Jul 17, 2023

• Support bfloat16 execution on the GPU (#1316)

  Show description for 7218d48
  guillaumekln

  authored
  7218d48

  Copy full SHA for 7218d48

Commits on Jun 15, 2023

• Update vocabulary file extension in the documentation

  guillaumekln

  committed
  5481035

  Copy full SHA for 5481035

Commits on Jun 6, 2023

• Initial support for encoder-only models (BERT) (#1256)

  Show description for c4b0fe9
  guillaumekln

  authored
  c4b0fe9

  Copy full SHA for c4b0fe9

• Cache the model state for static prompts (#1271)

  Show description for 5818e7a
  guillaumekln

  authored
  5818e7a

  Copy full SHA for 5818e7a

Commits on Jun 5, 2023

• Fix the required CUDA version for the Python package (#1269)

  guillaumekln

  authored
  532f34e

  Copy full SHA for 532f34e

• Add conversion code for Falcon models (#1260)

  Show description for 4108c0a
  guillaumekln

  authored
  4108c0a

  Copy full SHA for 4108c0a

Commits on Jun 2, 2023

• Remove unused line in the C++ quickstart code snippet

  guillaumekln

  committed
  7b1984c

  Copy full SHA for 7b1984c

• add docs for CodeGen GPT (#1258)

  michaelfeil

  authored
  3977a80

  Copy full SHA for 3977a80

Commits on May 26, 2023

• Integrate the HF LLaMa converter (#1242)

  guillaumekln

  authored
  8ae299b

  Copy full SHA for 8ae299b

Commits on May 22, 2023

• Support the MPT model from MosaicML (#1228)

  guillaumekln

  authored
  9ce09fe

  Copy full SHA for 9ce09fe

Commits on Apr 25, 2023

• Fallback to a custom threading implementation when OpenMP is not used (#1188)

  Show description for 4bdab79
  guillaumekln

  authored
  4bdab79

  Copy full SHA for 4bdab79

Commits on Apr 24, 2023

• Add token streaming example in the documentation (#1189)

  guillaumekln

  authored
  790677a

  Copy full SHA for 790677a