run with a error

[zjyb135246123]

File "/data/Phantom/gradio1.py", line 211, in inference
caption = model.generate({"image": image}, num_captions=1)[0]
File "/usr/local/lib/python3.10/site-packages/lavis/models/blip_models/blip_caption.py", line 188, in generate
decoder_out = self.text_decoder.generate_from_encoder(
File "/usr/local/lib/python3.10/site-packages/lavis/models/med.py", line 1360, in generate_from_encoder
outputs = self.generate(
File "/usr/local/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
File "/usr/local/lib/python3.10/site-packages/transformers/generation/utils.py", line 1797, in generate
return self.beam_search(
File "/usr/local/lib/python3.10/site-packages/transformers/generation/utils.py", line 3181, in beam_search
outputs = self(
File "/usr/local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/usr/local/lib/python3.10/site-packages/lavis/models/med.py", line 1210, in forward
outputs = self.bert(
File "/usr/local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/usr/local/lib/python3.10/site-packages/lavis/models/med.py", line 974, in forward
encoder_outputs = self.encoder(
File "/usr/local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/usr/local/lib/python3.10/site-packages/lavis/models/med.py", line 592, in forward
layer_outputs = layer_module(
File "/usr/local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/usr/local/lib/python3.10/site-packages/lavis/models/med.py", line 475, in forward
cross_attention_outputs = self.crossattention(
File "/usr/local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/usr/local/lib/python3.10/site-packages/lavis/models/med.py", line 346, in forward
self_outputs = self.self(
File "/usr/local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/usr/local/lib/python3.10/site-packages/lavis/models/med.py", line 219, in forward
attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2))
RuntimeError: The size of tensor a (3) must match the size of tensor b (9) at non-singleton dimension 0

[fangxia1997]

Maybe you can refer to this issue: salesforce/LAVIS#142
In fact, the PASD model does not have high requirements for prompts, so you can also try to skip this module and directly set the prompt to an empty string.