vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

use vocab size from config.json

Browse source
This commit is contained in:
akawrykow 2023-08-29 17:12:33 -07:00
parent 22c3522d78
commit ad45bb37f4
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
convert-falcon-hf-to-gguf.py
View file

@ -148,7 +148,7 @@ if Path(dir_model + "/tokenizer.json").is_file():
print("gguf: get gpt2 tokenizer vocab") print("gguf: get gpt2 tokenizer vocab")
vocab_size = len(tokenizer_json["model"]["vocab"]) vocab_size = hparams["vocab_size"]
# ref: https://github.com/cmp-nct/ggllm.cpp/blob/master/falcon_convert.py # ref: https://github.com/cmp-nct/ggllm.cpp/blob/master/falcon_convert.py
tokenizer = AutoTokenizer.from_pretrained(dir_model) tokenizer = AutoTokenizer.from_pretrained(dir_model)