latte模型的pos_embed没有被正确初始化

[WeYeYeYeYe]

在运行latte模型时，出现如下报错，检查 pos_embed的确没有被正确初始化，返回为none Traceback (most recent call last): File "/xDiT/./examples/latte_example.py", line 78, in main() File "/xDiT/./examples/latte_example.py", line 39, in main output = pipe( File "/miniconda3/envs/xdit/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context return func(*args, kwargs) File "/xDiT/xfuser/model_executor/pipelines/base_pipeline.py", line 133, in data_parallel_fn return func(self, *args, *kwargs) File "/xDiT/xfuser/model_executor/pipelines/base_pipeline.py", line 149, in check_naive_forward_fn return func(self, args, kwargs) File "/xDiT/xfuser/model_executor/pipelines/pipeline_latte.py", line 298, in call noise_pred = self.transformer( File "/miniconda3/envs/xdit/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1553, in _wrapped_call_impl return self._call_impl(*args, *kwargs) File "/miniconda3/envs/xdit/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1562, in _call_impl return forward_call(args, **kwargs) File "/xDiT/xfuser/model_executor/models/transformers/latte_transformer_3d.py", line 111, in forward hidden_states = self.pos_embed( TypeError: 'NoneType' object is not callable

[dannyxiaocn]

我没太遇到这个问题，你试着回退一下diffusers版本看一下？

[feifeibear]

Could you please report the diffusers version you used?

[feifeibear]

我们可能需要用最新的diffusers版本去测试一下latte。 diffusers变化很快，可能会影响latte实现？