ByteDance-Seed / ByteDance-Seed/cudaLLM

assert tensor_dict1.batch_size == tensor_dict2.batch_size error

Open
#1 2 comments 0 reactions 1 assignee Claimed by @preminstrel View on GitHub
good first issue
Dominant language
Python
Stars
155
Forks
7
PR merge metrics
No merged PRs in 30d

Description

执行 sh scripts/rl.sh
File "*****/cudaLLM_rl/main_ppo.py", line 561, in run
trainer.fit()
File "*****/cudaLLM_rl/ray_trainer.py", line 892, in fit
batch = batch.union(gen_batch_output)
File "*****/verl/verl/protocol.py", line 620, in union
self.batch = union_tensor_dict(self.batch, other.batch)
File "*****/verl/verl/protocol.py", line 107, in union_tensor_dict
assert tensor_dict1.batch_size == tensor_dict2.batch_size, (
AssertionError: Two tensor dict must have identical batch size. Got torch.Size([2048]) and torch.Size([256])

唯一的改动就是把n_nodes改成了1,这是什么原因

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.