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

feat: Add function for generating vocab script and fix CLI opts

Browse source
This commit is contained in:
teleprint-me 2024-05-18 22:14:58 -04:00
parent 0479e9695f
commit bd32266c87
No known key found for this signature in database
GPG key ID: B0D11345E65C4D48
1 changed files with 20 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

25
gguf-py/scripts/gguf-gen-pre.py
View file

@ -128,8 +128,23 @@ def test_pre_tok(hf_voc_req: HFVocabRequest) -> None:
logger.info(f"Tests for {model["repo"]} written in {final}.*")
def generate_tokenizers(hf_voc_req: HFVocabRequest) -> None:
pass
def generate_vocab_script(hf_voc_req: HFVocabRequest) -> None:
# generate commands for creating vocab files
shscript = "#!/usr/bin/env bash\n\n"
for model in hf_voc_req.models:
# get the repo path
path = Path(f"{hf_voc_req.model_path}/{model["repo"]}")
# set the vocab path
vocab = path / f"ggml-vocab-{path.stem.lower()}.gguf"
# set the command line
tmpline = f"python3 convert-hf-to-gguf.py {path} --outfile {vocab} --vocab-only\n"
shscript += tmpline
logger.info(tmpline.strip())
with open("generate-vocab.sh", "w", encoding="utf-8") as f:
f.writelines(shscript)
logger.info(f"Wrote {len(shscript)} bytes to generate-vocab.sh")
def main():
@ -145,7 +160,7 @@ def main():
"-t", "--gen-tests", action="store_true", help="Generate the tokenizer tests. Default is False."
)
parser.add_argument(
"-g", "--gen-toks", action="store_true", help="Generate the gguf vocab files. Default is False."
"-s", "--gen-vocab-script", action="store_true", help="Generate the gguf vocab files. Default is False."
)
args = parser.parse_args()
@ -165,8 +180,8 @@ def main():
if args.gen_tests:
test_pre_tok(hf_vocab_req)
if args.gen_toks:
generate_tokenizers(hf_vocab_req)
if args.gen_vocab_script:
generate_vocab_script(hf_vocab_req)
if __name__ == '__main__':