Update app.py

app.py

@@ -51,7 +51,7 @@ def model_generate_ans(img=None,img_audio=None,val_q=None):
     with torch.no_grad():
     
         # image
-        if image:
+        if img:
             image_processed  = processor(images=img, return_tensors="pt").to(device)
             clip_val_outputs = clip_model(**image_processed).last_hidden_state[:,1:,:]
             val_image_embeds = projection(clip_val_outputs)
@@ -76,7 +76,7 @@ def model_generate_ans(img=None,img_audio=None,val_q=None):
             val_q_embeds    = merged_model.model.embed_tokens(val_q_tokenised).unsqueeze(0)
 
         val_combined_embeds = []
-        if image:
+        if img:
             #val_combined_embeds = torch.cat([val_combined_embeds, val_image_embeds, img_token_embeds], dim=1)
             val_combined_embeds.append(val_image_embeds)
             val_combined_embeds.append(img_token_embeds)