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

reduce diffs with master

Browse source
This commit is contained in:
Christian Zhou-Zheng 2024-06-05 13:49:20 -04:00
parent bb5ee02096
commit 5ad397d610
2 changed files with 5 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

5
convert-hf-to-gguf.py
View file

@ -77,10 +77,8 @@ class Model:
self.lazy = not eager self.lazy = not eager
self.part_names = Model.get_model_part_names(self.dir_model, ".safetensors") self.part_names = Model.get_model_part_names(self.dir_model, ".safetensors")
self.is_safetensors = len(self.part_names) > 0 self.is_safetensors = len(self.part_names) > 0
if not self.is_safetensors: if not self.is_safetensors:
self.part_names = Model.get_model_part_names(self.dir_model, ".bin") self.part_names = Model.get_model_part_names(self.dir_model, ".bin")
self.hparams = Model.load_hparams(self.dir_model) self.hparams = Model.load_hparams(self.dir_model)
self.block_count = self.find_hparam(["n_layers", "num_hidden_layers", "n_layer"]) self.block_count = self.find_hparam(["n_layers", "num_hidden_layers", "n_layer"])
self.tensor_map = gguf.get_tensor_name_map(self.model_arch, self.block_count) self.tensor_map = gguf.get_tensor_name_map(self.model_arch, self.block_count)
@ -331,7 +329,7 @@ class Model:
self.write_tensors() self.write_tensors()
self.gguf_writer.write_header_to_file() self.gguf_writer.write_header_to_file()
self.gguf_writer.write_kv_data_to_file() self.gguf_writer.write_kv_data_to_file()
self.gguf_writer.write_tensors_to_file() self.gguf_writer.write_tensors_to_file(progress=True)
self.gguf_writer.close() self.gguf_writer.close()
def write_vocab(self): def write_vocab(self):
@ -1568,6 +1566,7 @@ class MiniCPMModel(Model):
return [(self.map_tensor_name(name), data_torch)] return [(self.map_tensor_name(name), data_torch)]
@Model.register("QWenLMHeadModel") @Model.register("QWenLMHeadModel")
class QwenModel(Model): class QwenModel(Model):
model_arch = gguf.MODEL_ARCH.QWEN model_arch = gguf.MODEL_ARCH.QWEN

5
gguf-py/gguf/gguf_manager.py
View file

@ -125,6 +125,7 @@ class SplitStrategy(deque):
num /= 1024.0 num /= 1024.0
return f"{num:.1f}T - over 1TB, --split recommended" return f"{num:.1f}T - over 1TB, --split recommended"
class GGUFManager(GGUFWriter): class GGUFManager(GGUFWriter):
kv_data: KVTempData kv_data: KVTempData
tensors: list[TensorTempData] tensors: list[TensorTempData]
@ -204,7 +205,7 @@ class GGUFManager(GGUFWriter):
self.state = WriterState.KV_DATA self.state = WriterState.KV_DATA
def write_tensors_to_file(self) -> None: def write_tensors_to_file(self, progress: bool = False) -> None:
if self.split_arguments.dry_run: if self.split_arguments.dry_run:
return return
@ -227,7 +228,7 @@ class GGUFManager(GGUFWriter):
# need to write everything down here # need to write everything down here
writer.write_header_to_file() writer.write_header_to_file()
writer.write_kv_data_to_file() writer.write_kv_data_to_file()
writer.write_tensors_to_file(progress=True) writer.write_tensors_to_file(progress=progress)
del tensors del tensors
self.state = WriterState.TI_DATA self.state = WriterState.TI_DATA