Search results

Qwen3_5_MoeQModel is a VL (image-text-to-text) model but only quantizes the language tower — does text-only calibration hurt multimodal task accuracy?

Status: Open.

#2928 In ModelCloud/GPTQModel;

· jiahong1314 opened on Jun 24, 2026

[GUIDE] a good guide is needed.

Status: Open.

#2442 In ModelCloud/GPTQModel;

· MaltoseFlower opened on Mar 5, 2026

[BUG] Rotation parameter not saved and not used in inference ?

bugSomething isn't workingSomething isn't working

Status: Open.

#2354 In ModelCloud/GPTQModel;

· 12345txy opened on Jan 12, 2026

Moe-vram for QQQ

performance

Status: Open.

#2115 In ModelCloud/GPTQModel;

· avtc opened on Oct 25, 2025

How to quantize both vision encoder and llm together?

Status: Open.

#1998 In ModelCloud/GPTQModel;

· XULU42 opened on Oct 9, 2025

Support the lm_head of QWEN2.5_OMNI？

Status: Open.

#1841 In ModelCloud/GPTQModel;

· LiMa-cas opened on Sep 22, 2025

TextSummarizationTask module not found

Status: Open.

#1659 In ModelCloud/GPTQModel;

· TianL123 opened on Jul 17, 2025

[BUG] Gemma 3 4b quantization outputs multi-modal config despite it being text only

bugSomething isn't workingSomething isn't working

Status: Open.

#1655 In ModelCloud/GPTQModel;

· ethan-tonic opened on Jul 9, 2025

Quantized QQQ models encountered configuration field exceptions and inference garbled text issues when deployed in vLLM 0.9.1.

Status: Open.

#1654 In ModelCloud/GPTQModel;

· ShiningMaker opened on Jul 9, 2025

[BUG] GPTQ-quantized OVIS 1B model yields poor performance & misaligned outputs in vLLM-0.9.1

bugSomething isn't workingSomething isn't working

Status: Open.

#1653 In ModelCloud/GPTQModel;

· AstonyJ opened on Jul 9, 2025

Feature: Add FSDP Peft/Lora finetuning sample

Status: Open.

#1648 In ModelCloud/GPTQModel;

· Qubitium opened on Jul 3, 2025

Kernels can not be built when I use pip install -v gptqmodel --no-build-isolation

Status: Open.

#1639 In ModelCloud/GPTQModel;

· BUGBOY101 opened on Jul 1, 2025