Update app.py

app.py

@@ -15,8 +15,6 @@ def predict_step(image_paths, model):
     i_image = Image.open(image_path)
     if i_image.mode != "RGB":
       i_image = i_image.convert(mode="RGB")
-    
-    #i_image.resize((640, 480))
 
     images.append(i_image)
 
@@ -69,9 +67,9 @@ def predict_step_pixel(dataset_pixel_values, model):
 """
 def load_image2txt_model(image_model_name):
     model = VisionEncoderDecoderModel.from_pretrained(image_model_name)
-    feature_extractor = AutoFeatureExtractor.from_pretrained("microsoft/swin-large-patch4-window7-224")
+    feature_extractor = AutoFeatureExtractor.from_pretrained("microsoft/swin-large-patch4-window7-224", use_auth_token=auth_token)
     
-    tokenizer = GPT2Tokenizer.from_pretrained("gpt2")
+    tokenizer = GPT2Tokenizer.from_pretrained("gpt2", use_auth_token=auth_token)
     tokenizer.pad_token = tokenizer.eos_token
 
     model = model.to(device)
@@ -81,7 +79,6 @@ def inference_image_pipe(image_input):
     image_model_name = "./checkpoint-21000"
 
     tokenizer, feature_extractor, image_model = load_image2txt_model(image_model_name)
-    #with autocast('cpu'):
     text = predict_step_single_image(image_input, tokenizer, feature_extractor, image_model)[0]
     return text
 
@@ -89,7 +86,7 @@ with gr.Interface(fn=inference_image_pipe,
              inputs=gr.Image(shape=(256, 256)),
              outputs="text",
              examples=["3212210S4492629-1.png", "3216497S4499373-1.png"]) as demo:
-    gr.Markdown("POC V0 - XRay Automatic Medical Report")
+    gr.Markdown("POC XRaySwinGen - Automatic Medical Report")
     
     
 if __name__ == "__main__":