Commits

Commits on Apr 10, 2026

[megatron] fix: Adjust the attention mask shape for VLM with Megatron on NPU (#5945)
ZLiao097
authored

Commits on Apr 8, 2026

[megatron] fix: Adjust the attention mask shape for VLM with Megatron on NPU (#5904)
ZLiao097
authored

Commits on Apr 2, 2026

[vllm] fix: Fix vLLM synchronization error caused by SGLang skipping resume optimize (#5866)
ZLiao097
authored

Commits on Mar 9, 2026

[ckpt] fix: Fix checkpoint engine backend unset error (#5473)
ZLiao097
authored

Commits on Feb 11, 2026

[fsdp, megatron] refactor: Refactor Fully Async Implementation via Engine Workers (#5269)
ZLiao097
authored

Commits on Feb 6, 2026

[reward] fix: Add assert to prevent reward NaN caused by overlong_cfg.len=0 (#5216)
ZLiao097
authored

Commits on Feb 2, 2026

[training_utils] fix: Resolved bugs and conflicts in the fully async caused by multiple PRs (#5100)
ZLiao097
authored

Commits on Dec 25, 2025

[fsdp] feat: Support zero2 optional feature for FSDP1 (#4659)
ZLiao097
authored

Commits on Nov 13, 2025

[fsdp] feat: add NPU fusion kernels for Qwen2 and Qwen2.5 dense model. (#3923)
ZLiao097
authored

Commits on Aug 29, 2025

[recipe] fix: Remove redundant parameters to resolve errors in the script caused by the latest Verl main branch. (#3252)
ZLiao097
authored

Commits on Aug 20, 2025

[recipe] feat: support qwen2.5-32B DAPO training script on ASCEND NPU (#3146)

ZLiao097
and
gemini-code-assist[bot]
authored