Commits

History for

Commits on Feb 2, 2026

Introduce AMD GPU support with ROCm HIP (#1989)
sssshhhhhh
authored

Commits on Jan 19, 2026

Add activation and dilation to conv1d (#1979)
sssshhhhhh
authored

Commits on Jan 14, 2026

Commits on Jan 11, 2026

Add causal flag to fa2 (#1976)
sssshhhhhh
authored

Commits on Jan 10, 2026

Fix CUDA bf16 median filter (#1972)
sssshhhhhh
authored

Commits on Jan 6, 2026

Version 4.6.3: change log and version bump (#1967)
jordimas
authored

Commits on Jan 4, 2026

T5Gemma conversion and inference (#1962)
jordimas
authored

Commits on Jan 3, 2026

Update to CUDA 12.8 (#1939)
Purfview
authored

Commits on Dec 22, 2025

Fix issue #939 - tokens batch_type may exceed max_batch_size (#1948)
anterart
authored

Commits on Dec 18, 2025

Add GPU implementation for median filter (#1917)
a2d8a4v
authored

Commits on Nov 29, 2025

Gemma 3 support (text only) (#1936)
jordimas
authored

Commits on Mar 28, 2025

Support for returning the hidden vector in the Wav2Vec2 and Wav2Vec2Bert models (#1867)

homink
and
hkwon
authored

Commits on Nov 5, 2024

Backward compatibility for the Wav2Vec2 ASR model (#1810)

homink
and
hkwon
authored

Commits on Sep 13, 2024

Wav2Vec2Bert ASR Inference Support (#1778)

homink
and
hkwon
authored

Commits on Sep 6, 2024

support minimum gemma 2 (#1772)
minhthuc2502
authored

Commits on Aug 26, 2024

Add log probs for all tokens (#1755)
minhthuc2502
authored

Commits on Aug 19, 2024

Wav2Vec2 upgrade with Conv1D options (#1758)

homink
and
hkwon
authored

Commits on Aug 13, 2024

feat: grouped conv1d (#1749)
ebraraktas
authored

Commits on Aug 12, 2024

fix: implement llama3 RoPE scaling type and fix converter (#1751)
ebraraktas
authored

Commits on Jul 4, 2024

Quantzation AWQ GEMM + GEMV (#1727)
minhthuc2502
authored

Commits on Jun 26, 2024

Small improvement in flash attention (#1732)
minhthuc2502
authored

Commits on May 17, 2024

support phi-3 128k (#1700)
minhthuc2502
authored

Commits on Apr 9, 2024

Flash attention (#1651)
minhthuc2502
authored

Commits on Mar 25, 2024

perf: conv1d quantization (#1601)
ebraraktas
authored

Commits on Mar 5, 2024

support tensor parallel (#1599)
minhthuc2502
authored

Commits on Feb 28, 2024

Support Gemma model (#1631)

minhthuc2502
and
thucpham
authored

Commits on Feb 9, 2024

perf: implement conv1d with im2col and GEMM (#1597)
ebraraktas
authored

Commits on Dec 26, 2023

Support of new option offset for ignoring token score of special tokens (#1592)

ASenart
and
asenellart
authored

Commits on Nov 17, 2023

Sliding window + chunking input for mistral model (#1524)

minhthuc2502
and
thucpham
authored

Commits on Nov 8, 2023

whisper-large-v3 compatibility (#1530)
funboarder13920
authored

Commits on Nov 7, 2023

Minimal support for Mistral 7B model. (#1528)
vince62s
authored

Commits on Nov 3, 2023

Support Transformers in the Wav2Vec2 Encoder for the ASR Inference (#1520)

homink
and
hkwon
authored

Commits on Sep 11, 2023

Use a macro to factorize declaration of some float specializations (#1463)
guillaumekln
authored