AI_HUB utils package
Project description
AI_HUB
AI utils for developer. such as notice、send massage when model training is over.Bind WeChat Official Account(AI_HUB) 插入在代码里的小工具,可以在模型训练结束时通过公众号及时发送微信消息给自己,提高科研效率。 inferServer: server your ai model as a API and match the tianchi eval 简单的操作把你训练好的模型变为服务API,并且支持天池大赛的流评测。
INSTALL
pip install ai-hub
SAMPLE
NOTICE
from ai_hub import notice
#到AGIHub微信公众号获取个人openid如(oM8pVuBWl8Rw_vFz7rZNgeO4T8H8),需替换为自己的openid
nc = notice("oM8pVuBWl8Rw_vFz7rZNgeO4T8H8")
#借助AGIHub公众号发送消息给自己
nc.sendmsg("hi,AI_HUB.I am su")
inferServer
'''
依赖:pip install ai-hub #(version>=0.1.7)
测试用例:
model为y=2*x
请求数据为json:{"img":3}
-----------
post请求:
curl localhost:8080/tccapi -X POST -d '{"img":3}'
返回结果 6
'''
from ai_hub import inferServer
import json
class myInfer(inferServer):
def __init__(self, model):
super().__init__(model)
print("init_myInfer")
#数据前处理
def pre_process(self, data):
print("my_pre_process")
#json process
json_data = json.loads(data.decode('utf-8'))
img = json_data.get("img")
print("processed data: ", img)
return img
#数据后处理
def post_process(self, data):
print("post_process")
processed_data = data
return processed_data
#模型预测:默认执行self.model(preprocess_data),一般不用重写
#如需自定义,可覆盖重写
#def pridect(self, data):
# ret = self.model(data)
# return ret
if __name__ == "__main__":
mymodel = lambda x: x * 2
my_infer = myInfer(mymodel)
my_infer.run(debuge=True) #默认为("127.0.0.1", 80),可自定义端口,如用于天池大赛请默认即可,指定debuge=True可获得更多报错信息
TccProgressBar
from ai_hub import TccProgressBar
#定义progress,显示名为training,在竞赛平台TCC上显示该进度条(tccBar_show=false 不影响本地打印进度条)
progress = TccProgressBar(title="training", tccBar_show=True)
for j in progress(range(100)):
time.sleep(0.1)
TccTensorboard
from ai_hub import Logger
#Logger用法与tensorboard的logger包一致
info= {
'loss': loss.data[0],
'accuracy': accuracy.data[0]
}
for tag, value in info.items():
logger.scalar_summary(tag, value, step)
获取OPENID
1.扫描关注公众号AGIHub
2.发送“openid”给公众号 即可获得openid
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
ai_hub-0.5.3.tar.gz
(6.7 kB
view details)
Built Distribution
File details
Details for the file ai_hub-0.5.3.tar.gz
.
File metadata
- Download URL: ai_hub-0.5.3.tar.gz
- Upload date:
- Size: 6.7 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.2.0 pkginfo/1.5.0.1 requests/2.22.0 setuptools/50.3.2 requests-toolbelt/0.9.1 tqdm/4.36.1 CPython/3.7.4
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | d452c6e0747659491ddf3a58738650667e7cc06886219680ac8513167c56c32e |
|
MD5 | 0ce92c6891a185620a0e3150482cd1ea |
|
BLAKE2b-256 | a75cff87471a72519800e6346a79e545f573e70bfeee06bb5422eb50ce266ad5 |
File details
Details for the file ai_hub-0.5.3-py3-none-any.whl
.
File metadata
- Download URL: ai_hub-0.5.3-py3-none-any.whl
- Upload date:
- Size: 8.9 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.2.0 pkginfo/1.5.0.1 requests/2.22.0 setuptools/50.3.2 requests-toolbelt/0.9.1 tqdm/4.36.1 CPython/3.7.4
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | c14189d8febd5f76de7cbd4cf64014b044a95624cb351d96c013c4c927c5efc5 |
|
MD5 | dea87610a9913495592b24ff33b22b6e |
|
BLAKE2b-256 | db529be7019ea82b7b507c79ac645a6faea1089dec28eed8196fbba2a920f715 |