[CI] Reuse loaded config for cached tokenizer (#49509)

Andreas Karatzas committed 3 months ago

Add `sm_107` for Rubin (#49387)

Tyler Michael Smith committed 3 months ago

[CI][AMD] Deprecate DinD for MI355 tests (#49257)

Andreas Karatzas committed 3 months ago

[Kernel] ReplaySSM: cache SSM inputs for faster Mamba2 standard decode (#48018)

Johnny-Liou committed 3 months ago

[Bugfix] Accept RFC 2397 parameters in base64 data URLs (#48973)

Thomas Fahrner committed 3 months ago

[UX] Improve data-parallel launch validation (#49124)

Taneem Ibrahim committed 3 months ago

[UX] Reject incompatible nested runtime overrides (#49247)

Taneem Ibrahim committed 3 months ago

[ROCM] Fix AITER Fused AllReduce RMSNorm for Transformers Backend (#49673)

BadrBasowid committed 3 months ago

[Bugfix][Benchmarks] Restore --skip-tokenizer-init with custom dataset (#49180)

mgazz committed 3 months ago

Encoder cache extension hooks (#48218)

hotTea committed 3 months ago

[Bugfix] Skip linear bias in layerwise reload to avoid corruption (#49586)

li-jinpeng committed 3 months ago

[PD][NixlPush][Bugfix] Fix blocking handshake call on writer thread (#49221)

Nicolò Lucchesi committed 3 months ago

Remove Quantization test parallelism (#49693)

Kevin H. Luu committed 3 months ago

[Perf][GLM-5.2] Blackwell decode optimizations (#48597)

xiaozhoupy committed 3 months ago

Bump Transformers version to 5.14.1 (#49223)

Harry Mellor committed 3 months ago

[CI] Use explicit devices in IR tests (#49513)

Andreas Karatzas committed 3 months ago

[Bugfix] Fix humming kernel crash when layer.has_bias is None (#48769)

Kyle Sayers committed 3 months ago

[CompressedTensors] DeepSeek4 CT Quantization Support (#41276)

Kyle Sayers committed 3 months ago

[Bugfix] Detect mixed precision in packed KV cache specs (#49623)

Michael Goin committed 3 months ago

[Model] Support llm-compressor Inkling NVFP4 weights (#49258)

Michael Goin committed 3 months ago