我尝试在train里面增加load model的代码，以期能中断后重新开始train。
代码类似run_api中的load model。
但是在sess.run（）后会报错：
Fetch argument <tf.Operation 'Adam' type=NoOp> cannot be interpreted as a Tensor.

Tensorflow写的不多，求解惑

我运行成功 没有值 不知道怎么回事 谢谢！
[INFO] 模型加载成功！

Serving Flask app "predict_api" (lazy loading)
Environment: production
WARNING: This is a development server. Do not use it in a production deployment.
Use a production WSGI server instead.
Debug mode: off
Running on http://127.0.0.1:5000/ (Press CTRL+C to quit)
运行完这结果 怎么办 呜呜

你好 ，对这个很感兴趣，但是模型文件缺失，能否告知如何下载或生成
谢谢!@zepen

通过使用您的脚本进行预测发现不管是ssq还是dlt后区连续多期存在相同号码11

如题,大乐透的也可以搞一下

手动创建了data文件夹拉取历史数据,运行train_model的时候报这个错:

ValueError: A target array with shape (2616, 24) was passed for an output of shape (None, 0) while using as loss categorical_crossentropy. This loss expects targets to have the same shape as the output.

同样的数据，反复训练是不是没有什么用？还是说每增加一期数据训练一次？另外，这个代码后续还有时间更新维护吗？

System information

OS Platform and Distribution (e.g., Linux Ubuntu 16.04): MacBook Air Apple M1 chip, macOS Monterey version 12.2
Mobile device (e.g. iPhone 8, Pixel 2, Samsung Galaxy) if the issue happens on mobile device: N/A
TensorFlow installed from (source or binary): binary
TensorFlow version: 2.1.0
Python version: 3.6.12
Installed using virtualenv? pip? conda?: pip3

Describe the problem
when i run

python train_model.py
[1]    59236 illegal hardware instruction  python train_model.py

你好 请问你这代码为什么不显示预测结果？还是你这代码只是显示往期已经开奖的号码？

我尝试了几次，都有奇怪的报错。。

可以多出一些训练结果吗？还有config.py配置的规则是什么？

用这个版本tensorflow-cpu==2.2.0， 高版本的，tensorflow-addons不支持了

看loss的趋势感觉一直就是在波动，没有明显下降趋势，而且红球们很多次预测出的结果都一样，感觉最终结果可能不如概率？

File "F:/predict_Lottery_ticket-master/predict_api.py", line 40, in get_predict_result
p_data = np.array(data_list).reshape([-1, 3, 1]).astype(np.float32)
ValueError: cannot reshape array of size 1 into shape (3,1)

经过测试，0.10.0版本的tensorflow-addons并不支持2.6.1版本的tensorflow-cpu，并且，2.5以上版本的tensorflow-cpu貌似不支持CRF，2.4版本的tensorflow-cpu不能很好的进行序列化，因此，个人推荐以下依赖版本：
tensorflow-cpu==2.3.*
tensorflow-addons==0.13.*
requests
keras
beautifulsoup4
pandas
numpy
lxml
loguru
grpcio
protobuf

但是，这样还是会有save时序列化warning的小问题，测试后发觉不影响使用

这模型能有效果？收敛吗？

大佬能写个详细点的教程
报错报的我脑瓜子嗡嗡的。。。

如题，小菜对机器学习可以说是一无所知，能否对每个参数进行释义或者说明每个参数的用途

=> ERROR [5/5] RUN pip install --no-cache-dir -r requirements.txt 3.3s

[5/5] RUN pip install --no-cache-dir -r requirements.txt:
#9 0.765 Collecting Flask==1.1.2
#9 1.189 Downloading Flask-1.1.2-py2.py3-none-any.whl (94 kB)
#9 1.765 ERROR: Could not find a version that satisfies the requirement tensorflow-cpu==2.6.1 (from versions: none)
#9 1.765 ERROR: No matching distribution found for tensorflow-cpu==2.6.1

executor failed running [/bin/sh -c pip install --no-cache-dir -r requirements.txt]: exit code: 1

我使用的虚拟机，所有插件版本都是request.txt里的，除了numpy 是1.19.5以外（如果使用request里的1.17.2，在训练的时候会出现：Python Fatal Error: GC Object already Tracked 的报错。）. Python 是3.6.13.

取数据和训练都顺利完成（会提示没有nvidia驱动，不过也能完成训练）。

运行了run_api.py后，打开 127.0.0.1:5000 是 “Welcome to use!”，如果打开127.0.0.1:5000/predict_api/ ，则报错："GET /predict_api/ HTTP/1.1" 404 -; 如果打开127.0.0.1:5000/predict_api （没有"/")， 则报错：
ERROR in app: Exception on /predict_api [GET]
Traceback (most recent call last):
File "/home/xxxxxx/.local/lib/python3.6/site-packages/flask/app.py", line 2446, in wsgi_app
response = self.full_dispatch_request()
File "/home/xxxxxx/.local/lib/python3.6/site-packages/flask/app.py", line 1951, in full_dispatch_request
rv = self.handle_user_exception(e)
File "/home/xxxxxx/.local/lib/python3.6/site-packages/flask/app.py", line 1820, in handle_user_exception
reraise(exc_type, exc_value, tb)
File "/home/xxxxxx/.local/lib/python3.6/site-packages/flask/_compat.py", line 39, in reraise
raise value
File "/home/xxxxxx/.local/lib/python3.6/site-packages/flask/app.py", line 1949, in full_dispatch_request
rv = self.dispatch_request()
File "/home/xxxxxx/.local/lib/python3.6/site-packages/flask/app.py", line 1935, in dispatch_request
return self.view_functionsrule.endpoint
File "run_api.py", line 53, in get_predict_result
diff_number = windows_size - 1
SystemError: unknown opcode
"GET /predict_api HTTP/1.1" 500 -

是我哪里没配置对吗？请抽空解答下，谢谢。

predict_api.py
接口输出号码，和计算结果集里的不一样，被+1，这是个大坑，咳损失好几百万啊
return json.dumps({b_name: int(res) + 1

这个要怎么运行

`2022-01-05 19:00:23.411489: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libnvinfer.so.6'; dlerror: libnvinfer.so.6: cannot open shared object file: No such file or directory
2022-01-05 19:00:23.411637: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libnvinfer_plugin.so.6'; dlerror: libnvinfer_plugin.so.6: cannot open shared object file: No such file or directory
2022-01-05 19:00:23.411655: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:30] Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.
2022-01-05 19:00:29.139770: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory
2022-01-05 19:00:29.139906: E tensorflow/stream_executor/cuda/cuda_driver.cc:351] failed call to cuInit: UNKNOWN ERROR (303)
2022-01-05 19:00:29.139977: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:156] kernel driver does not appear to be running on this host (leon): /proc/driver/nvidia/version does not exist
2022-01-05 19:00:29.159883: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2022-01-05 19:00:29.247429: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 2904000000 Hz
2022-01-05 19:00:29.266784: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x6ec1b00 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2022-01-05 19:00:29.266858: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version
[INFO] 已加载红球模型！
[INFO] 已加载蓝球模型！

• Serving Flask app "run_api" (lazy loading)
• Environment: production
  WARNING: This is a development server. Do not use it in a production deployment.
  Use a production WSGI server instead.
• Debug mode: off
• Running on all addresses.
  WARNING: This is a development server. Do not use it in a production deployment.
• Running on http://192.168.6.200:5000/ (Press CTRL+C to quit)`

可以试试这个：
http://datachart.500.com/ssq/history/newinc/history.php?limit=1000&sort=0

中奖的 出来露个脸哟！

发生异常: TypeError
'NoneType' object is not subscriptable
File "D:\下载\predict_Lottery_ticket-master\get_train_data.py", line 19, in get_current_number
current_num = soup.find("div", class_="wrap_datachart").find("input", id="end")["value"]
File "D:\下载\predict_Lottery_ticket-master\get_train_data.py", line 64, in
print("[INFO] 最新一期期号：{}".format(get_current_number()))

模型相关参数

windows_size = 3
batch_size = 1
sequence_len = 6
red_n_class = 33
red_epochs = 1
red_embedding_size = 32
red_hidden_size = 32
red_layer_size = 1
blue_n_class = 16
blue_epochs = 1
blue_embedding_size = 32
blue_hidden_size = 32
blue_layer_size = 1

模型训练参数

red_learning_rate = 0.001
red_beta1 = 0.9
red_beta2 = 0.999
red_epsilon = 1e-08
blue_learning_rate = 0.001
blue_beta1 = 0.9
blue_beta2 = 0.999
blue_epsilon = 1e-08

大佬，就我训练结束后运行predict_api文件后一直卡在下图这里。我之前没有使用过微服务，希望大佬给点指导，感谢。

WARNING:tensorflow:Layer lstm_2 will not use cuDNN kernel since it doesn't meet the cuDNN kernel criteria. It will use generic GPU kernel as fallback when running on GPU

Flask==1.1.2
tensorflow-cpu==2.6.2
tensorflow-addons==0.14.0
requests==2.22.0
beautifulsoup4==4.9.3
pandas==1.0.0
numpy==1.19.5
lxml==4.6.5
gunicorn==20.1.0
gevent==1.4.0
loguru==0.6.0
grpcio==1.46.0
protobuf==3.18.1

更新后按照这个配置能正常启动，麻烦楼主帮忙确认一下，是否有问题；
另外想请教下大神，当前的算法是基于之前出现的次数最少算出来的么。。模型有没有优化算法的想法或者计划。。

如题，使用默认参数，经过多次训练后发现红球，蓝球每次都会在固定的数字附近徘徊，每次训练后的结果趋于固定的结果，请指教模型参数怎样调试

问题1： bs4 只能安装 0.0.1版本

Could not fetch URL https://pypi.org/simple/bs4/: There was a problem confirming the ssl certificate: HTTPSConnectionPool(host='pypi.org', port=443): Max retries exceeded with url: /simple/bs4/ (Caused by SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:841)'),)) - skipping
ERROR: Could not find a version that satisfies the requirement bs4==4.9.3
ERROR: No matching distribution found for bs4==4.9.3
Could not fetch URL https://pypi.org/simple/pip/: There was a problem confirming the ssl certificate: HTTPSConnectionPool(host='pypi.org', port=443): Max retries exceeded with url: /simple/pip/ (Caused by SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:841)'),)) - skipping

问题2： 执行 python get_train_data.py

Traceback (most recent call last):
  File "get_train_data.py", line 64, in <module>
    print("[INFO] 最新一期期号：{}".format(get_current_number()))
  File "get_train_data.py", line 18, in get_current_number
    soup = BeautifulSoup(r.text, "lxml")
  File "E:\software\Python\lib\site-packages\bs4\__init__.py", line 246, in __init__
    % ",".join(features))
bs4.FeatureNotFound: Couldn't find a tree builder with the features you requested: lxml. Do you need to install a parser library?

问题3： 执行python run_api.py

Traceback (most recent call last):
  File "run_api.py", line 15, in <module>
    model_red_1 = load_model('model/lstm_model_红球号码_1.h5')
  File "E:\software\Python\lib\site-packages\tensorflow_core\python\keras\saving\save.py", line 149, in load_model
    loader_impl.parse_saved_model(filepath)
  File "E:\software\Python\lib\site-packages\tensorflow_core\python\saved_model\loader_impl.py", line 83, in parse_saved_model
    constants.SAVED_MODEL_FILENAME_PB))
OSError: SavedModel file does not exist at: model/lstm_model_红球号码_1.h5/{saved_model.pbtxt|saved_model.pb}

因为每期都有开奖，如果调用api来获取预测结果，是不是先要跑训练模型，每次都要先抓最新的数据下来

能否输出各个数字的概率数据，而不是单纯给一组号码。
比如红球 33 个，每个的概率在本期是多少，蓝球也是如此。
谢谢！

能换网站吗

我train的时候总是报WARNING:tensorflow:Layer lstm_1 will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU.结果就是，貌似大部分计算都去了cpu，gpu非常空闲。
我尝试按照stackoverflow的方法修改，但是没有成功，只能请教作者解决方法

{"red_1": 1, "red_2": 3, "red_3": 7, "red_4": 10, "red_5": 32, "red_6": 33, "blue": 12}

谢谢大神分享代码，已经能正常运行，请教：是否每期都需要运行get_train_data.py?

如题

其实彩票本身是一件随机事件，不过既然有数值，就有概率。我们可以排除一些概率最小的数字，建立一个排除列表，主要是红球，比如6个连续号码，这种概率是极低的。另外统计一下过往的连续号码最大位数和次数作为备选号码。还有一种就是重复开奖号码，如果重复开奖率极低，可以纳入排除列表。把范围缩小，概率自然就大一点。

预测过拟合，随机性不够。
后续的预测数字，重复数字太高了

在安装完依赖后，运行train model 报错