Update app.py

app.py

@@ -73,12 +73,15 @@ def generate_answers(img=None, aud = None, q = None, max_tokens = 30):
     inputs_embeddings.append(end_iq_embeds)
     # Combine embeddings
     combined_embeds  = torch.cat(inputs_embeddings, dim=1)
+    print("----------",combined_embeds.shape)
 
     for pos in range(max_tokens - 1):
         model_output_logits = phi2_model.forward(inputs_embeds = combined_embeds)['logits']
+        print("-=-=-=-", model_output_logits.shape)
         predicted_word_token_logits = model_output_logits[:, -1, :].unsqueeze(1)
         predicted_word_token = torch.argmax(predicted_word_token_logits, dim = -1)
         predicted_caption[:, pos] = predicted_word_token.view(1,-1).to('cpu')
+        print(predicted_caption)
         next_token_embeds = phi2_model.model.embed_tokens(predicted_word_token)
         combined_embeds   = torch.cat([combined_embeds, next_token_embeds], dim=1)
     predicted_captions_decoded = tokenizer.batch_decode(predicted_caption,ignore_index = 50256)[0]
@@ -110,4 +113,4 @@ with gr.Blocks() as demo:
     
 if __name__ == "__main__":
     
-    demo.launch(share=True)
+    demo.launch(share=True, debug=True)