L20卡型支持吗？跑了一下报错flash_apply_rotary_emb为NoneType，求大佬们看一下

[12-25 23:03:53|INFO|turbodiffusion/inference/wan2.1_t2v_infer.py:88:<module>] Generating with prompt: A stylish woman walks down a Tokyo street filled with warm glowing neon and animated city signage. She wears a black leather jacket, a long red dress, and black boots, and carries a black purse. She wears sunglasses and red lipstick. She walks confidently and casually. The street is damp and reflective, creating a mirror effect of the colorful lights. Many pedestrians walk about. Sampling: 0%| | 0/4 [00:00<?, ?it/s] Traceback (most recent call last): File "/app/TurboDiffusion/turbodiffusion/inference/wan2.1_t2v_infer.py", line 131, in <module> v_pred = net(x_B_C_T_H_W=x.to(**tensor_kwargs), timesteps_B_T=(t_cur.float() * ones * 1000).to(**tensor_kwargs), **condition).to( ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/envs/turbodiffusion/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/envs/turbodiffusion/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/app/TurboDiffusion/turbodiffusion/rcm/networks/wan2pt1.py", line 728, in forward x_B_L_D = block(x_B_L_D, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/envs/turbodiffusion/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/envs/turbodiffusion/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/envs/turbodiffusion/lib/python3.12/site-packages/torch/distributed/algorithms/_checkpoint/checkpoint_wrapper.py", line 171, in forward return self.checkpoint_fn( # type: ignore[misc] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/envs/turbodiffusion/lib/python3.12/site-packages/torch/_compile.py", line 53, in inner return disable_fn(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/envs/turbodiffusion/lib/python3.12/site-packages/torch/_dynamo/eval_frame.py", line 1044, in _fn return fn(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/envs/turbodiffusion/lib/python3.12/site-packages/torch/utils/checkpoint.py", line 503, in checkpoint ret = function(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/envs/turbodiffusion/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/envs/turbodiffusion/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/app/TurboDiffusion/turbodiffusion/rcm/networks/wan2pt1.py", line 456, in forward y = self.self_attn((self.norm1(x).float() * (1 + e[1]) + e[0]).type_as(x), seq_lens, video_size, freqs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/envs/turbodiffusion/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/root/miniconda3/envs/turbodiffusion/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/app/TurboDiffusion/turbodiffusion/rcm/networks/wan2pt1.py", line 320, in forward x = self.attn_op(rope_apply(q, video_size, freqs), rope_apply(k, video_size, freqs), v) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/app/TurboDiffusion/turbodiffusion/rcm/networks/wan2pt1.py", line 226, in rope_apply rotated = flash_apply_rotary_emb(x.to(torch.float32), cos, sin, interleaved=True, inplace=False) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ TypeError: 'NoneType' object is not callable

Start with turbodiffusion/inference/wan2.1_t2v_infer.py and trace the failing call through turbodiffusion/rcm/networks/wan2pt1.py, especially rope_apply and the self-attention path. Reproduce the L20 run, inspect how flash_apply_rotary_emb is initialized, and establish whether the reported configuration is supported; done means the failure is fixed or compatibility requirements are clearly documented.