extract_qwen_vl.py

from transformers import AutoModelForCausalLM
import torch
from modelscope import (
    snapshot_download, AutoModelForCausalLM, AutoTokenizer, GenerationConfig
)
import torch
model_id = 'qwen/Qwen-VL-Chat'
revision = 'v1.0.3'

model_dir = snapshot_download(model_id, revision=revision)
model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, fp16=True).eval()

state_dict = model.state_dict()
save_dict = {}
for k,v in state_dict.items():
    if 'visual' in k:
        if 'transformer.visual.proj' not in k: # we don't need the proj layer
            save_dict[k.replace('transformer.visual.', '')] = v
torch.save(save_dict, './qwen_clip/pytorch_model.bin')