| from peft import PeftModel |
| import torch |
| from transformers import BitsAndBytesConfig, Qwen2VLForConditionalGeneration, AutoProcessor, AutoConfig, Qwen2_5_VLForConditionalGeneration |
| import warnings |
| import os |
| import json |
|
|
| def disable_torch_init(): |
| """ |
| Disable the redundant torch default initialization to accelerate model creation. |
| """ |
| setattr(torch.nn.Linear, "reset_parameters", lambda self: None) |
| setattr(torch.nn.LayerNorm, "reset_parameters", lambda self: None) |
|
|
| |
| def load_pretrained_model(model_path, model_base, model_name, load_8bit=False, load_4bit=False, |
| device_map="auto", device="cuda", use_flash_attn=False, **kwargs): |
| kwargs = {"device_map": device_map} |
| |
| if device != "cuda": |
| kwargs['device_map'] = {"":device} |
| |
| if load_8bit: |
| kwargs['load_in_8bit'] = True |
| elif load_4bit: |
| kwargs['quantization_config'] = BitsAndBytesConfig( |
| load_in_4bit=True, |
| bnb_4bit_compute_dtype=torch.float16, |
| bnb_4bit_use_double_quant=True, |
| bnb_4bit_quant_type='nf4' |
| ) |
| else: |
| kwargs['torch_dtype'] = torch.float16 |
|
|
| if use_flash_attn: |
| kwargs['_attn_implementation'] = 'flash_attention_2' |
|
|
| if 'lora' in model_name.lower() and model_base is None: |
| warnings.warn('There is `lora` in model name but no `model_base` is provided. If you are loading a LoRA model, please provide the `model_base` argument.') |
| if 'lora' in model_name.lower() and model_base is not None: |
| lora_cfg_pretrained = AutoConfig.from_pretrained(model_path) |
| if hasattr(lora_cfg_pretrained, 'quantization_config'): |
| del lora_cfg_pretrained.quantization_config |
| processor = AutoProcessor.from_pretrained(model_base) |
| print('Loading Qwen2-VL from base model...') |
| if "Qwen2.5" in model_base: |
| model = Qwen2_5_VLForConditionalGeneration.from_pretrained(model_base, low_cpu_mem_usage=True, config=lora_cfg_pretrained, **kwargs) |
| else: |
| model = Qwen2VLForConditionalGeneration.from_pretrained(model_base, low_cpu_mem_usage=True, config=lora_cfg_pretrained, **kwargs) |
| token_num, tokem_dim = model.lm_head.out_features, model.lm_head.in_features |
| if model.lm_head.weight.shape[0] != token_num: |
| model.lm_head.weight = torch.nn.Parameter(torch.empty(token_num, tokem_dim, device=model.device, dtype=model.dtype)) |
| model.model.embed_tokens.weight = torch.nn.Parameter(torch.empty(token_num, tokem_dim, device=model.device, dtype=model.dtype)) |
|
|
| print('Loading additional Qwen2-VL weights...') |
| non_lora_trainables = torch.load(os.path.join(model_path, 'non_lora_state_dict.bin'), map_location='cpu') |
| non_lora_trainables = {(k[11:] if k.startswith('base_model.') else k): v for k, v in non_lora_trainables.items()} |
| if any(k.startswith('model.model.') for k in non_lora_trainables): |
| non_lora_trainables = {(k[6:] if k.startswith('model.') else k): v for k, v in non_lora_trainables.items()} |
| model.load_state_dict(non_lora_trainables, strict=False) |
| |
| print('Loading LoRA weights...') |
| model = PeftModel.from_pretrained(model, model_path) |
|
|
| print('Merging LoRA weights...') |
| model = model.merge_and_unload() |
|
|
| print('Model Loaded!!!') |
|
|
| else: |
| with open(os.path.join(model_path, 'config.json'), 'r') as f: |
| config = json.load(f) |
|
|
| if "Qwen2.5" in config["_name_or_path"]: |
| processor = AutoProcessor.from_pretrained(model_path) |
| model = Qwen2_5_VLForConditionalGeneration.from_pretrained(model_path, low_cpu_mem_usage=True, **kwargs) |
|
|
| else: |
| processor = AutoProcessor.from_pretrained(model_path) |
| model = Qwen2VLForConditionalGeneration.from_pretrained(model_path, low_cpu_mem_usage=True, **kwargs) |
|
|
| return processor, model |
|
|
|
|
| def get_model_name_from_path(model_path): |
| model_path = model_path.strip("/") |
| model_paths = model_path.split("/") |
| if model_paths[-1].startswith('checkpoint-'): |
| return model_paths[-2] + "_" + model_paths[-1] |
| else: |
| return model_paths[-1] |
