Training Neural Network
Once the Deep Neural Network is configured, it is ready to be trained to learn features from data-set.
Figure 3.10: Training Neural Network Interface.
- Choose the Training Parameters used for training the Neural Network.
- The training parameters are the parameters required for training algorithms to work and perform well. Certain parameter settings are only available for certain Training Algorithm. The recommended training parameters are set based on given data-set, but advanced users can also tweak to achieve better training performance. (Detail description of each parameter is described in the sub-section)
- Set up your Stopping Criteria
- The stopping criteria contains parameters that control when to end/stop training procedure.
- Once you determined the Training Algorithm and Stopping Criteria, press START TRAINING button.
- Training will begin immediately, and you may monitor the progress from the graph and the Loss and Evaluation values.
- A great feature of DLHUB is that it utilizes GPU for training (if compatible GPU is detected), this will drastically improve the training speed.
- Training will stop due to the following conditions:
- Training goal is reached. The training goal is measured by cost function value for given Neural Network' weights and given training data inputs. If this value < training goal specified in stopping criteria (0.0001 for example), the training goal is reached.
- Training PASS flag will indicate as GREEN. Training Completed flag will indicate as BLUE.
- Max Fail Count is reached.
- Epoch will increment until the Training Epoch has reached. If cost function has not reach the Training Goal, it will increment the Fail Count, reset the Current Epoch value and restart training.
- This process will repeat until the Training Goal is reached, or when the Max Fail count has been reached.
- only Training Completed flag will indicate as BLUE, Training PASS will remain off.
- You may stop the training any time by pressing the STOP button.
- you can proceed by pressing NEXT or reconfigure your parameters and train again.
- Press NEXT when you are done with training and ready to evaluate your trained network.
During the training process, the training data-set is used to adjust the Neural Network weights to optimize the cost/loss function. The value of this Loss function is called a performance index or performance of a training data-set. The validation set is used to prevent over-fitting/over-trained issue. The over-fitting problem happens with the Neural Network is trained to work well with training data-set but fail to predict outputs of the new data-set. The technique that uses a validation data-set to prevent the over-fitting issue is called the early stopping technique.