TTS模型,加载模型报错,说是pytorch版本导致 #709
Labels
No Label
bug
duplicate
enhancement
help wanted
invalid
question
wontfix
No Milestone
No project
No Assignees
2 Participants
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: HswOAuth/llm_course#709
Loading…
Reference in New Issue
Block a user
No description provided.
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
with torch.no_grad():
output = model(**inputs).waveform
执行报错ModuleNotFoundError: No module named 'torch.distributed.fsdp'
torch.distributed.fsdp 说是 PyTorch 1.11+ 版本引入的 完全分片数据并行,老师为什么执行pytorch==1.10.0版本可以成功
模型代码可能有条件分支,例如:
if torch.__version__ >= "1.11":
from torch.distributed.fsdp import FullyShardedDataParallel
else:
# 回退到其他并行策略(如 DDP)
from torch.nn.parallel import DistributedDataParallel
建议直接升级到 PyTorch 1.11.0 或更高版本