Max pit

1. There are two kinds of pooling, one is Max pooling, which means taking the maximum value in a fixed size sliding window, the other is mean pooling, which means taking the average value in a fixed size sliding window. As for the sliding mode of the window, it is the same as the convolution layer.

2. As an operation on features, pooling should depend on specific tasks. Some papers have mentioned that low-level networks are more responsive to abstract features, so with the continuous processing of features, high-level networks are more responsive to detailed features. As a feature processing operation, pooling should also consider what data is before pooling, what meaning data represents, and what is the correlation of data. Suppose to distinguish "a" and & quot; P"， The location correlation between the adjacent data in these two datasets is not large (from the perspective of NN, of course, it's another way to extract features from graphics). Then I think you can use the pooling of an average value as a feature extraction method. If some information in the data processed by a convolution kernel is very effective, for example, a large value may appear in it, and this value has a great influence on the result, then you can use Max pooling

3. Instead of everyone holding a letter waiting for the post office to pick it up, the maximum capacity of this mailbox is equivalent to the max pool size.

4. The maximum capacity of the buffer pool
just like the mail box of the post office, the letters to be processed can be put in it temporarily, instead of everyone holding a letter waiting for the post office to pick it up (to buffer and improve efficiency), then the maximum capacity of the mail box is equivalent to the max pool size.

5. First of all, I don't know what "JDBC" is in your statement. I understand it as a connection pool
jdbc.maxpoolsize... Should be a 32-bit int value. You can system.out int.maxvalue. This is just a theoretical connection
to make it clear, there are actually two connection pools, one is the connection pool on the application server side and the other is the connection pool on the database side. For example, if the connection pool on the EA side is 100 and the database side is 50, you still have only 50 available

and vice versa.

6. If you are connecting to sqlserver, you can configure the connection pool size in the connection string and specify it in the connection string. For example, maxpoolsize = 50, minpoolsize = 0, maxpoolsize is 100 by default, minpoolsize is 0 by default, for example: & quot; server=127.0.0.1; database=test1; UID=sa; Pwd=123456; MaxPoolSize=50"

7. 1. Connect timeout is time out; 2. Min pool size refers to the minimum program pool; 3. Pooling = true is to open the connection pool;

8. Do not add is according to the default value, in addition to the address pool, you can also add the connection timeout attribute

9. threadpool.setmaxthreads(2,2);