|
|
|
|
|
import json |
|
import pprint |
|
|
|
|
|
def read_json(file_path): |
|
with open(file_path, 'r', encoding='utf-8') as file: |
|
data = json.load(file) |
|
return data |
|
|
|
def write_json(file_path, data): |
|
with open(file_path, 'w', encoding='utf-8') as file: |
|
json.dump(data, file, ensure_ascii=False, indent=4) |
|
|
|
data = read_json("DataSet/train_samples_all_tuning.json") |
|
|
|
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
|
|
model_name = "Model/QwQ-32B-Preview" |
|
model = AutoModelForCausalLM.from_pretrained( |
|
model_name, |
|
torch_dtype="auto", |
|
device_map="auto" |
|
) |
|
tokenizer = AutoTokenizer.from_pretrained(model_name) |
|
|
|
|
|
|
|
def chat_QwQ(prompt): |
|
|
|
messages = [ |
|
{"role": "system", "content": "You are a helpful and harmless assistant. You are Qwen developed by Alibaba. You should think step-by-step."}, |
|
{"role": "user", "content": prompt} |
|
] |
|
|
|
text = tokenizer.apply_chat_template( |
|
messages, |
|
tokenize=False, |
|
add_generation_prompt=True |
|
) |
|
|
|
model_inputs = tokenizer([text], return_tensors="pt").to(model.device) |
|
|
|
generated_ids = model.generate( |
|
**model_inputs, |
|
max_new_tokens=512 |
|
) |
|
|
|
generated_ids = [ |
|
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids) |
|
] |
|
|
|
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0] |
|
|
|
return response |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
for i in data: |
|
sent1 = i['conversations'][0]['value'] |
|
sent2 = i['conversations'][1]['value'] |
|
sentence = sent1 + sent2 |
|
prompt = "This is a question-answering datapoint based on image information. Determine whether the answer can be judged without relying on the image. If it can, this is considered bad data; if it requires the image, it is considered good data. Rate this datapoint on a scale from 1 (bad) to 5 (good). ###### " + sentence |
|
answer = chat_QwQ(prompt) |
|
|
|
|
|
|
|
|
|
|
|
|
|
pprint.pprint(prompt) |
|
pprint.pprint(answer) |
|
break |