Update app.py

app.py

@@ -4,8 +4,15 @@ import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
 MODEL_ID = "devoppro/FastLLM"
+TOKENIZER_FALLBACK_ID = "Qwen/Qwen2.5-0.5B"  # FastLLM reuses the Qwen2.5 BPE vocab
+
+try:
+    tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True)
+except Exception as e:
+    print(f"Could not load tokenizer from {MODEL_ID} ({e}); "
+          f"falling back to {TOKENIZER_FALLBACK_ID}")
+    tokenizer = AutoTokenizer.from_pretrained(TOKENIZER_FALLBACK_ID)
 
-tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True)
 model = AutoModelForCausalLM.from_pretrained(
     MODEL_ID,
     torch_dtype=torch.float16,