File size: 7,605 Bytes
a12c07f |
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 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 |
import argparse
import os
import json
import random
import re
import torch
import numpy as np
from tqdm import tqdm
import shortuuid
import sys
import transformers
from torch.utils.data import Dataset, DataLoader
from PIL import Image
import math
from .gpt4v import TaskSpec, ParsedAnswer, Question
from .exceptions import GPTOutputParseException, GPTMaxTriesExceededException
import threading
from typing import List, Tuple, Union
from loguru import logger
from copy import deepcopy
import time
import os
seed = 42
torch.manual_seed(seed)
np.random.seed(seed)
random.seed(seed)
torch.backends.cudnn.benchmark = False
class LlamaModel(object):
def __init__(self, task:TaskSpec,
model:str = "meta-llama/Meta-Llama-3.1-8B-Instruct"):
self.task:TaskSpec = task
self.model_id = model
num_gpus = torch.cuda.device_count()
# Assign device based on the available GPUs
if num_gpus == 2:
self.device_map = "cuda:1"
if num_gpus == 1:
self.device_map = "cuda:0"
self.pipeline = transformers.pipeline(
"text-generation",
model=self.model_id,
model_kwargs={"attn_implementation":"flash_attention_2", "torch_dtype": torch.float16},
device_map=self.device_map, # Map the model to GPU 2 (index 1)
)
# self.pipeline = transformers.pipeline(
# "text-generation",
# model=self.model_id,
# model_kwargs={"torch_dtype": torch.bfloat16},
# device_map="auto", # Map the model to GPU 2 (index 1)
# )
def ask(self, payload:dict, n_choices=1, temperature=0.7) -> Tuple[List[dict], List[dict]]:
"""
args:
payload: json dictionary, prepared by `prepare_payload`
"""
def llama_thread(self, idx, payload, results, temperature):
# creation of payload
mod_payload = deepcopy(payload)
messages = payload['messages']
max_tokens = payload['max_tokens']
try:
with torch.autocast(device_type='cuda'):
output_text = self.pipeline(
messages,
max_new_tokens=max_tokens,
)
except Exception as e:
raise e
# print('outputs: ', output_text)
message = output_text[0]["generated_text"][-1]
results[idx] = {"metadata": output_text, "message": message}
return
assert n_choices >= 1
results = [None] * n_choices
if n_choices > 1:
llama_jobs = [threading.Thread(target=llama_thread,
args=(self, idx, payload, results, temperature))
for idx in range(n_choices)]
for job in llama_jobs:
job.start()
for job in llama_jobs:
job.join()
else:
llama_thread(self, 0, payload, results, temperature)
messages:List[dict] = [ res["message"] for res in results]
metadata:List[dict] = [ res["metadata"] for res in results]
return messages, metadata
@staticmethod
def prepare_payload(question:Question,
max_tokens=1000,
verbose:bool=False,
prepend:Union[dict, None]=None,
**kwargs
) -> dict:
image_dic = None
text = ''
dic_list = question.get_json()
for dic in question.get_json():
# The case of text
if dic['type'] == 'text':
text += dic['text']
# The case of vision input
elif dic['type'] == 'image_url':
image_dic = dic['image']
payload = {
"messages": [
{"role": "user", "content": text},
],
"max_tokens": max_tokens,
}
return payload
def rough_guess(self, question:Question, max_tokens=1000,
max_tries=1, query_id:int=0,
verbose=False, temperature=1,
**kwargs):
p = self.prepare_payload(question, max_tokens = max_tokens, verbose=verbose, prepend=None)
ok = False
reattempt = 0
while not ok:
response, meta_data = self.ask(p, temperature=temperature)
response = response[0]
# logger.info(f'response: {response}')
try:
parsed_response = self.task.answer_type.parser(response["content"])
except GPTOutputParseException as e:
# logger.warning(f"The following was not parseable:\n\n{response}\n\nBecause\n\n{e}")
# if not os.path.exists('errors/'):
# # Create the directory if it doesn't exist
# os.makedirs('errors/')
# error_saved = f'errors/{time.strftime("%Y-%m-%d-%H-%M-%S")}.json'
# with open(error_saved, "w") as f:
# f.write(p_ans.code)
# logger.warning(f"The following was not parseable. Saved in {error_saved}.")
reattempt += 1
if reattempt > max_tries:
logger.error(f"max tries ({max_tries}) exceeded.")
raise GPTMaxTriesExceededException
logger.warning(f"Reattempt #{reattempt} querying LLM")
continue
ok = True
return parsed_response, response, meta_data, p
def many_rough_guesses(self, num_threads:int,
question:Question, max_tokens=1000,
verbose=False, max_tries=1, temperature=1
) -> List[Tuple[ParsedAnswer, str, dict, dict]]:
"""
Args:
num_threads : number of independent threads.
all other arguments are same as those of `rough_guess()`
Returns
List of elements, each element is a tuple following the
return signature of `rough_guess()`
"""
p = self.prepare_payload(question, max_tokens = max_tokens, verbose=verbose, prepend=None)
# TODO
n_choices = num_threads
# TODO: wrap in robust-ask method, repeatedly asks until parseable output.
ok = False
reattempt = 0
while not ok:
response, meta_data = self.ask(p, n_choices=n_choices, temperature=temperature)
try:
parsed_response = [self.task.answer_type.parser(r["content"]) for r in response]
except GPTOutputParseException as e:
# logger.warning(f"The following was not parseable:\n\n{response}\n\nBecause\n\n{e}")
# TODO provide the parse error message into GPT for the next round to be parsable
reattempt += 1
if reattempt > max_tries:
logger.error(f"max tries ({max_tries}) exceeded.")
raise GPTMaxTriesExceededException
logger.warning(f"Reattempt #{reattempt} querying LLM")
continue
ok = True
return parsed_response, response, meta_data, p
def run_once(self, question:Question, max_tokens=1000, temperature=1, **kwargs):
q = self.task.first_question(question)
p_ans, ans, meta, p = self.rough_guess(q, max_tokens=max_tokens, temperature=temperature, **kwargs)
return p_ans, ans, meta, p
|