dorkai's picture
Upload 293 files
6a4546d
import logging
from pathlib import Path
import torch
from peft import PeftModel
import modules.shared as shared
def add_lora_to_model(lora_names):
prior_set = set(shared.lora_names)
added_set = set(lora_names) - prior_set
removed_set = prior_set - set(lora_names)
shared.lora_names = list(lora_names)
# If no LoRA needs to be added or removed, exit
if len(added_set) == 0 and len(removed_set) == 0:
return
# Add a LoRA when another LoRA is already present
if len(removed_set) == 0 and len(prior_set) > 0:
logging.info(f"Adding the LoRA(s) named {added_set} to the model...")
for lora in added_set:
shared.model.load_adapter(Path(f"{shared.args.lora_dir}/{lora}"), lora)
return
# If any LoRA needs to be removed, start over
if len(removed_set) > 0:
shared.model.disable_adapter()
shared.model = shared.model.base_model.model
if len(lora_names) > 0:
logging.info("Applying the following LoRAs to {}: {}".format(shared.model_name, ', '.join(lora_names)))
params = {}
if not shared.args.cpu:
params['dtype'] = shared.model.dtype
if hasattr(shared.model, "hf_device_map"):
params['device_map'] = {"base_model.model." + k: v for k, v in shared.model.hf_device_map.items()}
elif shared.args.load_in_8bit:
params['device_map'] = {'': 0}
shared.model = PeftModel.from_pretrained(shared.model, Path(f"{shared.args.lora_dir}/{lora_names[0]}"), **params)
for lora in lora_names[1:]:
shared.model.load_adapter(Path(f"{shared.args.lora_dir}/{lora}"), lora)
if not shared.args.load_in_8bit and not shared.args.cpu:
shared.model.half()
if not hasattr(shared.model, "hf_device_map"):
if torch.has_mps:
device = torch.device('mps')
shared.model = shared.model.to(device)
else:
shared.model = shared.model.cuda()