Mirrors/text-generation-webui
1
0
Fork 0
mirror of https://github.com/oobabooga/text-generation-webui.git synced 2024-11-22 08:07:56 +01:00
Code Issues Packages Projects Releases Wiki Activity

Revert walrus operator for params['max_memory'] (#5878)

Browse Source
This commit is contained in:
Colin 2024-04-24 00:09:14 -04:00 committed by GitHub
parent c725d97368
commit f3c9103e04
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
modules/models.py
View File

@ -179,7 +179,7 @@ def huggingface_loader(model_name):
# DeepSpeed ZeRO-3
elif shared.args.deepspeed:
model = LoaderClass.from_pretrained(path_to_model, torch_dtype=params['torch_dtype'], trust_remote_code=params['trust_remote_code'])
model = LoaderClass.from_pretrained(path_to_model, torch_dtype=params['torch_dtype'], trust_remote_code=params.get('trust_remote_code'))
model = deepspeed.initialize(model=model, config_params=ds_config, model_parameters=None, optimizer=None, lr_scheduler=None)[0]
model.module.eval() # Inference
logger.info(f'DeepSpeed ZeRO-3 is enabled: {is_deepspeed_zero3_enabled()}')
@ -215,15 +215,15 @@ def huggingface_loader(model_name):
else:
params['quantization_config'] = BitsAndBytesConfig(load_in_8bit=True)
if params['max_memory'] is not None:
if params.get('max_memory') is not None:
with init_empty_weights():
model = LoaderClass.from_config(config, trust_remote_code=params['trust_remote_code'])
model = LoaderClass.from_config(config, trust_remote_code=params.get('trust_remote_code'))
model.tie_weights()
params['device_map'] = infer_auto_device_map(
model,
dtype=torch.int8,
max_memory=params['max_memory'],
max_memory=params.get('max_memory'),
no_split_module_classes=model._no_split_modules
)