Update app.py

app.py

@@ -45,7 +45,7 @@ def infer(
     occurrence = {}
     state = None
     for i in range(int(token_count)):
-        out, state = model.forward(pipeline.encode(ctx) if i == 0 else [token], state)
+        out, state = model.forward(pipeline.encode(ctx)[:4096] if i == 0 else [token], state)
         for n in args.token_ban:
             out[n] = -float('inf')
         for n in occurrence: