mirror of
https://github.com/open-compass/opencompass.git
synced 2025-05-30 16:03:24 +08:00
add_qwen_api_qwq_32b
This commit is contained in:
wangjingchao
parent
4b09860f4a
commit
a80dcd4df7
@ -19,6 +19,7 @@ with read_base():
|
||||
|
||||
from opencompass.configs.models.openai.o1_mini_2024_09_12 import models as o1_mini_2024_09_12_model
|
||||
from opencompass.configs.models.qwq.lmdeploy_qwq_32b_preview import models as lmdeploy_qwq_32b_preview_model
|
||||
from opencompass.configs.models.qwq.qwq_32b import models as qwq_32b_model
|
||||
from opencompass.configs.models.deepseek.deepseek_r1_api_aliyun import models as deepseek_r1_api_aliyun_model
|
||||
from opencompass.configs.models.deepseek.deepseek_r1_distill_llama_8b_api_aliyun import models as deepseek_r1_distill_llama_8b_api_aliyun_model
|
||||
from opencompass.configs.models.deepseek.deepseek_r1_distill_qwen_7b_api_aliyun import models as deepseek_r1_distill_qwen_7b_api_aliyun_model
|
||||
@ -58,6 +59,7 @@ models = [
|
||||
*gpt_4o_mini_20240718_model,
|
||||
*gpt_4o_20241120_model,
|
||||
*o1_mini_2024_09_12_model,
|
||||
*qwq_32b_model,
|
||||
*deepseek_v3_api_aliyun_model,
|
||||
*deepseek_r1_api_aliyun_model,
|
||||
*deepseek_r1_distill_llama_8b_api_aliyun_model,
|
||||
|
||||
19
opencompass/configs/models/qwq/qwq_32b.py
Normal file
19
opencompass/configs/models/qwq/qwq_32b.py
Normal file
@ -0,0 +1,19 @@
|
||||
from opencompass.models import Qwen
|
||||
|
||||
api_meta_template = dict(round=[
|
||||
dict(role='HUMAN', api_role='HUMAN'),
|
||||
dict(role='BOT', api_role='BOT', generate=True),
|
||||
], )
|
||||
|
||||
models = [
|
||||
dict(
|
||||
abbr='QwQ-32B',
|
||||
type=Qwen,
|
||||
path='qwq-32b',
|
||||
key='ENV', # The key will be obtained from $OPENAI_API_KEY, but you can write down your key here as well
|
||||
meta_template=api_meta_template,
|
||||
query_per_second=1,
|
||||
max_out_len=2048,
|
||||
max_seq_len=4096,
|
||||
batch_size=8),
|
||||
]
|
||||
@ -1,3 +1,5 @@
|
||||
import os
|
||||
import random
|
||||
import time
|
||||
from concurrent.futures import ThreadPoolExecutor
|
||||
from typing import Dict, List, Optional, Union
|
||||
@ -43,7 +45,16 @@ class Qwen(BaseAPIModel):
|
||||
retry=retry,
|
||||
generation_kwargs=generation_kwargs)
|
||||
import dashscope
|
||||
dashscope.api_key = key
|
||||
if isinstance(key, str):
|
||||
if key == 'ENV':
|
||||
if 'DASHSCOPE_API_KEY' not in os.environ:
|
||||
raise ValueError('DASHSCOPE API key is not set.')
|
||||
self.keys = os.getenv('DASHSCOPE_API_KEY').split(',')
|
||||
else:
|
||||
self.keys = [key]
|
||||
else:
|
||||
self.keys = key
|
||||
self.path = path
|
||||
self.dashscope = dashscope
|
||||
|
||||
def generate(
|
||||
@ -131,7 +142,9 @@ class Qwen(BaseAPIModel):
|
||||
self.acquire()
|
||||
try:
|
||||
response = self.dashscope.Generation.call(
|
||||
api_key=random.choice(self.keys),
|
||||
model=self.path,
|
||||
stream=True,
|
||||
**data,
|
||||
)
|
||||
except Exception as err:
|
||||
@ -148,34 +161,51 @@ class Qwen(BaseAPIModel):
|
||||
# to slow down the request
|
||||
self.wait()
|
||||
continue
|
||||
#
|
||||
reasoning_content = "" # 定义完整思考过程
|
||||
answer_content = "" # 定义完整回复
|
||||
is_answering = False # 判断是否结束思考过程并开始回复
|
||||
for chunk in response:
|
||||
if (chunk.output.choices[0].message.content == "" and
|
||||
chunk.output.choices[0].message.reasoning_content == ""):
|
||||
pass
|
||||
else:
|
||||
if (chunk.output.choices[0].message.reasoning_content != "" and
|
||||
chunk.output.choices[0].message.content == ""):
|
||||
reasoning_content += chunk.output.choices[0].message.reasoning_content
|
||||
elif chunk.output.choices[0].message.content != "":
|
||||
if not is_answering:
|
||||
is_answering = True
|
||||
answer_content += chunk.output.choices[0].message.content
|
||||
reasoning_content = '<think>' + reasoning_content + '</think>'
|
||||
return reasoning_content + answer_content
|
||||
# if response.status_code == 200:
|
||||
# try:
|
||||
# msg = response.output.text
|
||||
# self.logger.debug(msg)
|
||||
# return msg
|
||||
# except KeyError:
|
||||
# print(response)
|
||||
# self.logger.error(str(response.status_code))
|
||||
# time.sleep(1)
|
||||
# continue
|
||||
# if response.status_code == 429:
|
||||
# print(response)
|
||||
# time.sleep(2)
|
||||
# continue
|
||||
# if response.status_code == 400:
|
||||
# print('=' * 128)
|
||||
# print(response)
|
||||
# msg = 'Output data may contain inappropriate content.'
|
||||
# return msg
|
||||
|
||||
if response.status_code == 200:
|
||||
try:
|
||||
msg = response.output.text
|
||||
self.logger.debug(msg)
|
||||
return msg
|
||||
except KeyError:
|
||||
print(response)
|
||||
self.logger.error(str(response.status_code))
|
||||
time.sleep(1)
|
||||
continue
|
||||
if response.status_code == 429:
|
||||
print(response)
|
||||
time.sleep(2)
|
||||
continue
|
||||
if response.status_code == 400:
|
||||
print('=' * 128)
|
||||
print(response)
|
||||
msg = 'Output data may contain inappropriate content.'
|
||||
return msg
|
||||
|
||||
if ('Range of input length should be ' in response.message
|
||||
or # input too long
|
||||
'Input data may contain inappropriate content.'
|
||||
in response.message): # bad input
|
||||
print(response.message)
|
||||
return ''
|
||||
print(response)
|
||||
max_num_retries += 1
|
||||
# if ('Range of input length should be ' in response.message
|
||||
# or # input too long
|
||||
# 'Input data may contain inappropriate content.'
|
||||
# in response.message): # bad input
|
||||
# print(response.message)
|
||||
# return ''
|
||||
# print(response)
|
||||
# max_num_retries += 1
|
||||
|
||||
raise RuntimeError(response.message)
|
||||
|
||||
Loading…
Reference in New Issue
Block a user