File size: 2,428 Bytes
c3f3b0b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
import argparse
from transformers import AutoTokenizer, AutoModelForCausalLM, StoppingCriteria
import torch
import os
import json
from tqdm import tqdm
import shortuuid

from cumo.conversation import default_conversation
from cumo.utils import disable_torch_init


@torch.inference_mode()
def eval_model(model_name, questions_file, answers_file):
    # Model
    disable_torch_init()
    model_name = os.path.expanduser(model_name)
    tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=False)
    model = AutoModelForCausalLM.from_pretrained(model_name,
        torch_dtype=torch.float16).cuda()


    ques_file = open(os.path.expanduser(questions_file), "r")
    ans_file = open(os.path.expanduser(answers_file), "w")
    for i, line in enumerate(tqdm(ques_file)):
        idx = json.loads(line)["question_id"]
        qs = json.loads(line)["text"]
        cat = json.loads(line)["category"]
        conv = default_conversation.copy()
        conv.append_message(conv.roles[0], qs)
        prompt = conv.get_prompt()
        inputs = tokenizer([prompt])
        input_ids = torch.as_tensor(inputs.input_ids).cuda()
        output_ids = model.generate(
            input_ids,
            do_sample=True,
            use_cache=True,
            temperature=0.7,
            max_new_tokens=1024,)
        outputs = tokenizer.batch_decode(output_ids, skip_special_tokens=True)[0]
        try:
            index = outputs.index(conv.sep, len(prompt))
        except ValueError:
            outputs += conv.sep
            index = outputs.index(conv.sep, len(prompt))

        outputs = outputs[len(prompt) + len(conv.roles[1]) + 2:index].strip()
        ans_id = shortuuid.uuid()
        ans_file.write(json.dumps({"question_id": idx,
                                   "text": outputs,
                                   "answer_id": ans_id,
                                   "model_id": model_name,
                                   "metadata": {}}) + "\n")
        ans_file.flush()
    ans_file.close()

if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument("--model-name", type=str, default="facebook/opt-350m")
    parser.add_argument("--question-file", type=str, default="tables/question.jsonl")
    parser.add_argument("--answers-file", type=str, default="answer.jsonl")
    args = parser.parse_args()

    eval_model(args.model_name, args.question_file, args.answers_file)