Hi, I am trying your FaceSwarp function, I have run your example normally, but when I run my own data, it always prompts that the size does not match, I tried 512, 768, 832, 1024 and other common values, but there is always a problem in dimensions 2 or 3 or 4, flowing is one example:
Using wan_video_vae from checkpoints/base_model/Wan2.1_VAE.pth.
No wan_video_image_encoder models available.
No wan_video_motion_controller models available.
No wan_video_vace models available.
Loading Stand-In weights from: checkpoints/Stand-In/Stand-In_wan2.1_T2V_14B_ver1.0.ckpt
Traceback (most recent call last):
File "//Stand-In/infer_face_swap.py", line 104, in
video = pipe(
^^^^^
File "/anaconda3/envs/standin/lib/python3.11/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "//Stand-In/pipelines/wan_video_face_swap.py", line 606, in call
inputs_shared, inputs_posi, inputs_nega = self.unit_runner(
^^^^^^^^^^^^^^^^^
File "//Stand-In/utils/init.py", line 366, in call
processor_outputs = unit.process(pipe, **processor_inputs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "Stand-In/pipelines/wan_video_face_swap.py", line 802, in process
latents = pipe.scheduler.add_noise(
^^^^^^^^^^^^^^^^^^^^^^^^^
File "//Stand-In/schedulers/flow_match.py", line 88, in add_noise
sample = (1 - sigma) * original_samples + sigma * noise
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~
RuntimeError: The size of tensor a (104) must match the size of tensor b (60) at non-singleton dimension 3
Hi, I am trying your FaceSwarp function, I have run your example normally, but when I run my own data, it always prompts that the size does not match, I tried 512, 768, 832, 1024 and other common values, but there is always a problem in dimensions 2 or 3 or 4, flowing is one example:
Using wan_video_vae from checkpoints/base_model/Wan2.1_VAE.pth.
No wan_video_image_encoder models available.
No wan_video_motion_controller models available.
No wan_video_vace models available.
Loading Stand-In weights from: checkpoints/Stand-In/Stand-In_wan2.1_T2V_14B_ver1.0.ckpt
Traceback (most recent call last):
File "//Stand-In/infer_face_swap.py", line 104, in
video = pipe(
^^^^^
File "/anaconda3/envs/standin/lib/python3.11/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "//Stand-In/pipelines/wan_video_face_swap.py", line 606, in call
inputs_shared, inputs_posi, inputs_nega = self.unit_runner(
^^^^^^^^^^^^^^^^^
File "//Stand-In/utils/init.py", line 366, in call
processor_outputs = unit.process(pipe, **processor_inputs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "Stand-In/pipelines/wan_video_face_swap.py", line 802, in process
latents = pipe.scheduler.add_noise(
^^^^^^^^^^^^^^^^^^^^^^^^^
File "//Stand-In/schedulers/flow_match.py", line 88, in add_noise
sample = (1 - sigma) * original_samples + sigma * noise
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~
RuntimeError: The size of tensor a (104) must match the size of tensor b (60) at non-singleton dimension 3