Search results

[Bug]: AWQ fails on Qwen3.5 MoE because default mapping matches all input_layernorm layers instead of a single smooth layer

bugSomething isn't workingSomething isn't working

Status: Open.

#2514 In vllm-project/llm-compressor;

· 1007281134yanyu-cloud opened on Mar 25, 2026

Support for AWQ quantization (W4A16) for Qwen3_5MoeForConditionalGeneration architecture

Status: Open.

#2513 In vllm-project/llm-compressor;

· Priya95715 opened on Mar 25, 2026

Has anyone performed inference on a GLM-5 model quantized with AWQ+INT4 using VLLM?

Status: Open.

#2512 In vllm-project/llm-compressor;

· artificialzjy opened on Mar 25, 2026

[Feature Request] QAT support: training-aware quantization for llm-compressor

enhancementNew feature or requestNew feature or request

Status: Open.

#2509 In vllm-project/llm-compressor;

· dzhengAP opened on Mar 24, 2026

[model_free_ptq] Runtime optimization: meta device shape validation, multi-GPU compression, reindexing elimination

enhancementNew feature or requestNew feature or request

Status: Open.

#2497 In vllm-project/llm-compressor;

· dzhengAP opened on Mar 20, 2026

[Feature Request/Question]fused QKV linear layers

enhancementNew feature or requestNew feature or request

Status: Open.

#2494 In vllm-project/llm-compressor;

· Arthur-wza opened on Mar 20, 2026

Consolidate Intermediate Offloading

enhancementNew feature or requestNew feature or request

good follow-up issueA good issue for users with some familiarity of the codebaseA good issue for users with some familiarity of the codebase

RefactorCode cleanup and/or improvements to existing featuresCode cleanup and/or improvements to existing features

Status: Open.

#2490 In vllm-project/llm-compressor;

· HDCharles opened on Mar 19, 2026

AWQ + KV cache quantization

enhancementNew feature or requestNew feature or request

Status: Open.

#2488 In vllm-project/llm-compressor;

· saranya-vp-15149 opened on Mar 19, 2026

[Good First Issue] Rename offload_model and dispatch_for_sequential to set_onload_device

compressed-tensorsRelates to compressed-tensorsRelates to compressed-tensors

enhancementNew feature or requestNew feature or request

good first issueA good first issue for users wanting to contributeA good first issue for users wanting to contribute

Status: Open.

#2483 In vllm-project/llm-compressor;

· kylesayrs opened on Mar 17, 2026

[Feature] Allow targeting multiples of sequential targets

enhancementNew feature or requestNew feature or request

good first issueA good first issue for users wanting to contributeA good first issue for users wanting to contribute

tracingIssues related to model tracingIssues related to model tracing

Status: Open.

#2481 In vllm-project/llm-compressor;

· kylesayrs opened on Mar 17, 2026

[Feature] Enable GPU offloading for reduced weight movement

enhancementNew feature or requestNew feature or request

good first issueA good first issue for users wanting to contributeA good first issue for users wanting to contribute

Status: Open.

#2480 In vllm-project/llm-compressor;

· kylesayrs opened on Mar 17, 2026

AWQ various enhancements and research

awqFor any issue / PR related to AWQ supportFor any issue / PR related to AWQ support

enhancementNew feature or requestNew feature or request

Status: Open.

#2479 In vllm-project/llm-compressor;

· HDCharles opened on Mar 17, 2026