Update app.py

app.py

@@ -38,31 +38,43 @@ import torch
 
 #     f1_metric.set(f1)
 
-vitgpt_processor = ViTImageProcessor.from_pretrained("model")
-vitgpt_model = VisionEncoderDecoderModel.from_pretrained("model")
-vitgpt_tokenizer = AutoTokenizer.from_pretrained("model", return_tensors="pt")
+feature_extractor = ViTImageProcessor.from_pretrained("model")
+cap_model = VisionEncoderDecoderModel.from_pretrained("model")
+tokenizer = AutoTokenizer.from_pretrained("model")
 
 device = "cuda" if torch.cuda.is_available() else "cpu"
 
 vitgpt_model.to(device)
 
 def generate_caption(processor, model, image, tokenizer=None):
-    inputs = processor(images=image, return_tensors="pt").to(device)
+  max_length = 16
+  num_beams = 4
+  gen_kwargs = {"max_length": max_length, "num_beams": num_beams}
+
+  pixel_values = feature_extractor(images=image, return_tensors="pt").pixel_values
+  pixel_values = pixel_values.to(device)
+
+  output_ids = model.generate(pixel_values, **gen_kwargs)
+
+  preds = tokenizer.batch_decode(output_ids, skip_special_tokens=True)
+  preds = [pred.strip() for pred in preds]
+  return preds
+    # inputs = processor(images=image, return_tensors="pt").to(device)
     
-    generated_ids = model.generate(pixel_values=inputs.pixel_values, max_length=50)
+    # generated_ids = model.generate(pixel_values=inputs.pixel_values, max_length=50)
 
-    if tokenizer is not None:
-        generated_caption = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
-    else:
-        generated_caption = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
+    # if tokenizer is not None:
+    #     generated_caption = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
+    # else:
+    #     generated_caption = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
    
-    return generated_caption
+    # return generated_caption
 
-def predict_event(input):
+def predict_event(image):
     
 
 
-    caption_vitgpt = generate_caption(vitgpt_processor, vitgpt_model, image, vitgpt_tokenizer)
+    caption_vitgpt = generate_caption(feature_extractor, cap_model, image, tokenizer)
 
     return caption_vitgpt