List view

decouple quanitzers

Status: Open.

#787 In intel/auto-round;

· wenhuach21 opened on Sep 3, 2025

decouple calibration code

Status: Open.

#810 In intel/auto-round;

· wenhuach21 opened on Sep 10, 2025

refine model save folder name

Status: Open.

#828 In intel/auto-round;

· wenhuach21 opened on Sep 18, 2025

Enhance AutoScheme

Status: Open.

#910 In intel/auto-round;

· wenhuach21 opened on Oct 17, 2025

remove itrex format and update INC

Status: Open.

#808 In intel/auto-round;

· wenhuach21 opened on Sep 10, 2025

Support dynamic FP8

high priority

Status: Open.

#959 In intel/auto-round;

· wenhuach21 opened on Oct 29, 2025

support fp8 sdpa for tuning

high priority

Status: Open.

#938 In intel/auto-round;

· xin3he opened on Oct 23, 2025

MXFP4/MXFP8 evaluation

Status: Open.

#937 In intel/auto-round;

· yiliu30 opened on Oct 23, 2025

Weekly test Performance and Peak memory in case of regression

Status: Open.

#991 In intel/auto-round;

· xin3he opened on Nov 4, 2025

Add WAN 2.2-A14B support

Status: Open.

#928 In intel/auto-round;

· xin3he opened on Oct 22, 2025

Optimize RAM usage of AutoScheme

high priority

Status: Open.

#912 In intel/auto-round;

· wenhuach21 opened on Oct 17, 2025

device specification error

Status: Open.

#452 In intel/auto-round;

· n1ck-guo opened on Mar 4, 2025

torch compile causes gguf tuning large accuracy drop

Status: Open.

#702 In intel/auto-round;

· wenhuach21 opened on Jul 31, 2025

try to keep e_score_correction_bias dtype as FP32 for deepseek

high priority

Status: Open.

#845 In intel/auto-round;

· wenhuach21 opened on Sep 23, 2025

add support for mixed datatype packing and inference

Status: Open.

#948 In intel/auto-round;

· wenhuach21 opened on Oct 27, 2025

MiniMaxAI/MiniMax-M2 Q2ks

Status: Open.

#952 In intel/auto-round;

· groxaxo opened on Oct 28, 2025

Refactor all formats into classes and simplify the format-related checks

Status: Open.

#786 In intel/auto-round;

· wenhuach21 opened on Sep 3, 2025

SGlang cannot load AutoRound model

Status: Open.

#982 In intel/auto-round;

· haofanwang opened on Nov 3, 2025

torch_dtype is deprecated in Transformers

high priority

Status: Open.

#776 In intel/auto-round;

· wenhuach21 opened on Sep 1, 2025