patrickjohncyh

/

fashion-clip

like 185

Hello,
why after inference, text_model_output.hidden_states and text_model_output.attention are None while text_model_output.last_hidden_state and text_model_output.pooler_output are properly populated?
The same goes for vision_model_output