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

tidy up

Browse source
This commit is contained in:
Christian Zhou-Zheng 2024-06-09 11:20:14 -04:00
parent 69d6e7a8e9
commit 0779f2f74f
2 changed files with 9 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

7
convert-hf-to-gguf.py
View file

@ -2891,13 +2891,14 @@ def main() -> None:
model_instance.gguf_writer.add_quantization_version(gguf.GGML_QUANT_VERSION)
if args.vocab_only:
logger.info(f"Exporting model vocab to '{model_instance.fname_out}'")
logger.info(f"Exporting model vocab...")
model_instance.write_vocab()
logger.info(f"Model vocab successfully exported.")
else:
logger.info(f"Exporting model to '{model_instance.fname_out}'")
logger.info(f"Exporting model...")
model_instance.write()
logger.info(f"Model successfully exported.")
logger.info(f"Model successfully exported to '{model_instance.fname_out}'")
if __name__ == '__main__':

9
gguf-py/gguf/gguf_writer.py
View file

@ -95,7 +95,7 @@ class GGUFWriter:
fout: list[BufferedWriter | None] | None
path: os.PathLike[str] | str | None
temp_file: tempfile.SpooledTemporaryFile[bytes] | None
tensors: list[dict[str, TensorInfo]]
tensors: list[dict[str, TensorInfo | np.ndarray[Any, Any]]]
kv_data: list[dict[str, GGUFValue]]
state: WriterState
_simple_value_packing = {
@ -182,7 +182,7 @@ class GGUFWriter:
def print_plan(self) -> None:
logger.info("Writing the following files:")
for i in range(len(self.fout)):
logger.info(f" {self.fout[i].name}: n_tensors = {len(self.tensors[i])}, total_size = {GGUFWriter.format_n_bytes_to_str(GGUFWriter.get_tensors_total_size(self.tensors[i].values()))}")
logger.info(f"{self.fout[i].name}: n_tensors = {len(self.tensors[i])}, total_size = {GGUFWriter.format_n_bytes_to_str(GGUFWriter.get_tensors_total_size(self.tensors[i].values()))}")
if self.split_arguments.dry_run:
logger.info("Dry run, not writing files")
@ -327,8 +327,9 @@ class GGUFWriter:
if self.state is not WriterState.NO_FILE:
raise ValueError(f'Expected output file to be not yet opened, got {self.state}')
if name in self.tensors:
raise ValueError(f'Duplicated tensor name {name!r}')
for i in range(len(self.tensors)):
if name in self.tensors[i]:
raise ValueError(f'Duplicated tensor name {name!r}')
if raw_dtype is None:
if tensor_dtype == np.float16: