Add INT4 compressed-tensors + LoRA support#1

Merged

sheikheddy merged 22 commits intomainfrom

feat/int4-compressed-tensors-lora-support

Nov 17, 2025

Commits on Nov 16, 2025

feat: Add INT4 compressed-tensors + LoRA support

sheikheddy
and
claude
committed
fix: Add LoRA compatibility to compressed-tensors MoE methods

sheikheddy
and
claude
committed
fix: Correct embedding dimension logic in LoRA dummy creation

sheikheddy
and
claude
committed
fix: Add LoRA compatibility to W4A4 and W4A8 MoE methods
sheikheddy
committed
Adding a benchmark for batch invariance (#28161)

bwasti
and
gemini-code-assist[bot]
authored
[Benchmark] Fix client seed synchronization in multi-turn benchmark (#28512)
ai-jz
authored
[Model] Allow users to control skip reading cache per request. (#28194)
noooop
authored
[V1] Support MP Executor for multi node distributed inference (#23691)

luccafong
and
njhill
authored
Fixed gpt-oss _load_weights_other() parameter position bug (#28715)

River12
and
Dezhan Tu
authored
[Bugfix] Fix host and port join for ipv6 in bench serve (#28679)

scottzh8
and
Scott Zhang
authored
Fix gpt oss weight loading with EP + bf16 (#28765)
ashors1
authored
[Doc]: fix typos in various files (#28811)
didier-durand
authored
fix comment typo (#28802)
andyxning
authored
[Model][QwenVL] Optimize `Qwen2_5_VisionAttention` q,k preparation (#28769)

lgeiger
and
Isotr0py
authored
Feature: Support Relu2 in FusedMoE fp8 cutlass path (#27261)
amirkl94
authored
[BugFix] Fix async scheduling + chunked prefill + preemption (#28787)
njhill
authored

Commits on Nov 17, 2025