chenyushuo
chenyushuo
Firstly, thanks for your great work. I have followed the steps in README and do not modify any other parameters, but I cannot reproduce the result in your paper. I...
--- name: Add CPP Servicer about: Add CPP Servicer --- ## Description Rerewrite Servicer with CPP. ## Checklist Please check the following items before code is ready to be reviewed....
[BUG] Qwen3 MoE with FSDP2 meets `torch.utils.checkpoint.CheckpointError` when `offload_policy=True`
Hi, I encountered a problem while trying to run the following command: 我在尝试运行下面的命令时遇到了问题: ```bash export MODEL_PATH=/root/Qwen3-30B-A3B PYTHONUNBUFFERED=1 python3 -m verl.trainer.main_ppo \ data.train_files=$HOME/data/gsm8k/train.parquet \ data.val_files=$HOME/data/gsm8k/test.parquet \ data.train_batch_size=256 \ data.max_prompt_length=512 \ data.max_response_length=256...