Commits

Commits on Feb 11, 2026

[Bugfix] Enable attn quantization of Llama-4 by correctly permuting scales for rope (int8, fp8) (#34243)

eldarkurtic
and
Your Name
authored

Commits on Feb 6, 2026

[Docs] Update link to Benchmark CLI documentation (#33254)
eldarkurtic
authored

Commits on Jan 22, 2026

Add llmcompressor fp8 kv-cache quant (per-tensor and per-attn_head) (#30141)
eldarkurtic
authored

Commits on Nov 25, 2025

Attempt to fix GPU OOM in a spec-decoding test (#29419)
eldarkurtic
authored
Revert back to torch.equal over torch.allclose from #28819 (#29086)
eldarkurtic
authored

Commits on Nov 15, 2025

Add support for Eagle with separate lm-head and embed_tokens layers (#28549)
eldarkurtic
authored

Commits on Sep 22, 2025

Enable Eagle3 speculative decoding for GPT-OSS model (#25246)
eldarkurtic
authored

Commits on Aug 9, 2025

Fix loading of quantized BigCode models (#22463)
eldarkurtic
authored

Commits on Jun 25, 2025

static_scaled_fp8_quant should not run when scale.numel is not 1 (#20076)
eldarkurtic
authored

Commits on Feb 3, 2025

Properly check if all fused layers are in the list of targets (#12666)
eldarkurtic
authored

Commits on Feb 1, 2025

Fix target matching for fused layers with compressed-tensors (#12617)
eldarkurtic
authored