chunzhi/chatgpt_academic
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

修复文心一言最大文本长度限制带来的问题

Browse Source
This commit is contained in:
binary-husky 2023-08-25 13:09:08 +08:00
parent 67fff17917
commit 6cc7d4ed69
3 changed files with 22 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
request_llm/bridge_all.py
View File

@ -175,7 +175,7 @@ model_info = {
"fn_with_ui": qianfan_ui, "fn_with_ui": qianfan_ui,
"fn_without_ui": qianfan_noui, "fn_without_ui": qianfan_noui,
"endpoint": None, "endpoint": None,
"max_token": 4096, "max_token": 2000,
"tokenizer": tokenizer_gpt35, "tokenizer": tokenizer_gpt35,
"token_cnt": get_token_num_gpt35, "token_cnt": get_token_num_gpt35,
}, },

1
request_llm/bridge_chatgpt.py
View File

@ -216,7 +216,6 @@ def handle_error(inputs, llm_kwargs, chatbot, history, chunk_decoded, error_msg)
history = clip_history(inputs=inputs, history=history, tokenizer=model_info[llm_kwargs['llm_model']]['tokenizer'], history = clip_history(inputs=inputs, history=history, tokenizer=model_info[llm_kwargs['llm_model']]['tokenizer'],
max_token_limit=(model_info[llm_kwargs['llm_model']]['max_token'])) # history至少释放二分之一 max_token_limit=(model_info[llm_kwargs['llm_model']]['max_token'])) # history至少释放二分之一
chatbot[-1] = (chatbot[-1][0], "[Local Message] Reduce the length. 本次输入过长, 或历史数据过长. 历史缓存数据已部分释放, 您可以请再次尝试. (若再次失败则更可能是因为输入过长.)") chatbot[-1] = (chatbot[-1][0], "[Local Message] Reduce the length. 本次输入过长, 或历史数据过长. 历史缓存数据已部分释放, 您可以请再次尝试. (若再次失败则更可能是因为输入过长.)")
# history = [] # 清除历史
elif "does not exist" in error_msg: elif "does not exist" in error_msg:
chatbot[-1] = (chatbot[-1][0], f"[Local Message] Model {llm_kwargs['llm_model']} does not exist. 模型不存在, 或者您没有获得体验资格.") chatbot[-1] = (chatbot[-1][0], f"[Local Message] Model {llm_kwargs['llm_model']} does not exist. 模型不存在, 或者您没有获得体验资格.")
elif "Incorrect API key" in error_msg: elif "Incorrect API key" in error_msg:

28
request_llm/bridge_qianfan.py
View File

@ -1,9 +1,9 @@
import time, requests, json import time, requests, json
from toolbox import update_ui, get_conf
from multiprocessing import Process, Pipe from multiprocessing import Process, Pipe
from functools import wraps from functools import wraps
from datetime import datetime, timedelta from datetime import datetime, timedelta
from toolbox import get_conf, update_ui, is_any_api_key, select_api_key, what_keys, clip_history, trimmed_format_exc, get_conf
model_name = '千帆大模型平台' model_name = '千帆大模型平台'
timeout_bot_msg = '[Local Message] Request timeout. Network error.' timeout_bot_msg = '[Local Message] Request timeout. Network error.'
@ -101,6 +101,7 @@ def generate_from_baidu_qianfan(inputs, llm_kwargs, history, system_prompt):
response = requests.request("POST", url, headers=headers, data=payload, stream=True) response = requests.request("POST", url, headers=headers, data=payload, stream=True)
buffer = "" buffer = ""
for line in response.iter_lines(): for line in response.iter_lines():
if len(line) == 0: continue
try: try:
dec = line.decode().lstrip('data:') dec = line.decode().lstrip('data:')
dec = json.loads(dec) dec = json.loads(dec)
@ -108,7 +109,10 @@ def generate_from_baidu_qianfan(inputs, llm_kwargs, history, system_prompt):
buffer += incoming buffer += incoming
yield buffer yield buffer
except: except:
if 'error_code' in dec: raise RuntimeError(dec['error_msg']) if ('error_code' in dec) and ("max length" in dec['error_msg']):
raise ConnectionAbortedError(dec['error_msg'])
elif ('error_code' in dec):
raise RuntimeError(dec['error_msg'])
def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False): def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False):
@ -137,11 +141,21 @@ def predict(inputs, llm_kwargs, plugin_kwargs, chatbot, history=[], system_promp
from core_functional import handle_core_functionality from core_functional import handle_core_functionality
inputs, history = handle_core_functionality(additional_fn, inputs, history, chatbot) inputs, history = handle_core_functionality(additional_fn, inputs, history, chatbot)
# 开始接收回复 yield from update_ui(chatbot=chatbot, history=history)
for response in generate_from_baidu_qianfan(inputs, llm_kwargs, history, system_prompt): # 开始接收回复
chatbot[-1] = (inputs, response) try:
yield from update_ui(chatbot=chatbot, history=history) for response in generate_from_baidu_qianfan(inputs, llm_kwargs, history, system_prompt):
chatbot[-1] = (inputs, response)
yield from update_ui(chatbot=chatbot, history=history)
except ConnectionAbortedError as e:
from .bridge_all import model_info
if len(history) >= 2: history[-1] = ""; history[-2] = "" # 清除当前溢出的输入：history[-2] 是本次输入, history[-1] 是本次输出
history = clip_history(inputs=inputs, history=history, tokenizer=model_info[llm_kwargs['llm_model']]['tokenizer'],
max_token_limit=(model_info[llm_kwargs['llm_model']]['max_token'])) # history至少释放二分之一
chatbot[-1] = (chatbot[-1][0], "[Local Message] Reduce the length. 本次输入过长, 或历史数据过长. 历史缓存数据已部分释放, 您可以请再次尝试. (若再次失败则更可能是因为输入过长.)")
yield from update_ui(chatbot=chatbot, history=history, msg="异常") # 刷新界面
return
# 总结输出 # 总结输出
response = f"[Local Message]: {model_name}响应异常 ..." response = f"[Local Message]: {model_name}响应异常 ..."
if response == f"[Local Message]: 等待{model_name}响应中 ...": if response == f"[Local Message]: 等待{model_name}响应中 ...":