set max length to 64

interfaces/cap.py

@@ -103,7 +103,7 @@ def predict(text, model_id, tokenizer_id):
     t2 = time.perf_counter()
     inputs = tokenizer(
         text,
-        max_length=256,
+        max_length=64,
         truncation=True,
         padding=True,
         return_tensors="pt"

utils.py

@@ -92,7 +92,7 @@ def download_hf_models():
                 return_tensors="pt",
                 padding=True,
                 truncation=True,
-                max_length=256
+                max_length=64
             )
 
             # JIT trace