Update app.py

app.py

@@ -1,26 +1,45 @@
 import gradio as gr
 import onnxruntime as rt
 from transformers import AutoTokenizer
-import torch, json
+import torch
+import json
 
+# Load tokenizer and genre types
 tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
 
 with open("genre_types_encoded.json", "r") as fp:
-  encode_genre_types = json.load(fp)
+    encode_genre_types = json.load(fp)
 
 genres = list(encode_genre_types.keys())
 
+# Load ONNX model
 inf_session = rt.InferenceSession('movie-genre-classifier-quantized.onnx')
 input_name = inf_session.get_inputs()[0].name
 output_name = inf_session.get_outputs()[0].name
 
 def classify_movie_genre(summary):
-  input_ids = tokenizer(summary)['input_ids'][:512]
-  logits = inf_session.run([output_name], {input_name: [input_ids]})[0]
-  logits = torch.FloatTensor(logits)
-  probs = torch.sigmoid(logits)[0]
-  return dict(zip(genres, map(float, probs))) 
+    # Tokenize input
+    tokens = tokenizer(summary, padding='max_length', truncation=True, return_tensors="pt")
+    input_ids = tokens['input_ids'][0].tolist()[:512]
+    
+    # Debug prints
+    print("Input summary:", summary)
+    print("Tokenized input:", input_ids)
+    
+    # Run inference
+    logits = inf_session.run([output_name], {input_name: [input_ids]})[0]
+    
+    # Convert logits to probabilities
+    logits = torch.FloatTensor(logits)
+    probs = torch.sigmoid(logits)[0]
+    
+    # Debug prints
+    print("Logits:", logits)
+    print("Probabilities:", probs)
+    
+    return dict(zip(genres, map(float, probs))) 
 
+# Setup Gradio interface
 label = gr.Label(num_top_classes=5)
 iface = gr.Interface(fn=classify_movie_genre, inputs="text", outputs=label)
 iface.launch(inline=False)