Use stated vocab size

Author: akawrykow

convert-falcon-hf-to-gguf.py

@@ -131,7 +131,9 @@ def parse_args() -> argparse.Namespace:
 
 print("gguf: get gpt2 tokenizer vocab")
 
-vocab_size = len(tokenizer_json["model"]["vocab"])
+# The number of tokens in tokenizer.json can differ from the expected vocab size.
+# This causes downstream issues with mismatched tensor sizes when running the inference
+vocab_size = hparams["vocab_size"]
 
 # ref: https://github.com/cmp-nct/ggllm.cpp/blob/master/falcon_convert.py
 tokenizer = AutoTokenizer.from_pretrained(dir_model)