List view

adjust 3 bits hyperparameters at auto-round-best

Status: Merged (completed).

intel/auto-roundnumber 1081

#1081 In intel/auto-round;

· by wenhuach21 was closed on Dec 2, 2025

Fix MXFP/NVFP + FP8 Attn/KV

Status: Merged (completed).

intel/auto-roundnumber 1086

#1086 In intel/auto-round;

· by yiliu30 was closed on Dec 3, 2025

Support alg_ext on windows

Status: Merged (completed).

intel/auto-roundnumber 1082

#1082 In intel/auto-round;

· by chensuyue was closed on Dec 2, 2025

[vLLM Ext]Fix MXFP4 Quant

Status: Merged (completed).

intel/auto-roundnumber 1088

#1088 In intel/auto-round;

· by yiliu30 was closed on Dec 3, 2025

Add MXFP8 MOE/Linear and MXFP4 Linear

Status: Merged (completed).

intel/auto-roundnumber 1034

#1034 In intel/auto-round;

· by yiliu30 was closed on Nov 28, 2025

remove numpy restriction for gptq kernel

Status: Merged (completed).

intel/auto-roundnumber 1084

#1084 In intel/auto-round;

· by wenhuach21 was closed on Dec 3, 2025

Refactor input normalization by replaying inputs for consistent preprocessing

llmc

Status: Merged (completed).

intel/auto-roundnumber 1094

#1094 In intel/auto-round;

· by yiliu30 was closed on Dec 5, 2025

add system checker in backend

Status: Merged (completed).

intel/auto-roundnumber 1097

#1097 In intel/auto-round;

· by wenhuach21 was closed on Dec 5, 2025

fix gguf acc and oom bug when iters > 0

Status: Merged (completed).

intel/auto-roundnumber 1098

#1098 In intel/auto-round;

· by n1ck-guo was closed on Dec 5, 2025

lm_head is not converted to QuantLinear with MXFP4/8

bugSomething isn't workingSomething isn't working

high priority

Status: Closed (completed).

#1040 In intel/auto-round;

· by xin3he was closed on Nov 28, 2025

Why random sampler during training?

Status: Closed (completed).

#1076 In intel/auto-round;

· by HsuWanTing was closed on Dec 3, 2025

MXFP4/MXFP8 evaluation

Status: Closed (completed).

#937 In intel/auto-round;

· by yiliu30 was closed on Dec 5, 2025