Training
Here we will train our model with multiple GPU's or CPU's device.
Overview:
Training a model simply means learning (determining) good values for all the weights and the bias from labeled examples.
Deep learning models are built using neural networks. A neural network takes in inputs, which are then processed in hidden layers using weights that are adjusted during training. Then the model spits out a prediction.
Deep Learning Studio supports multiple GPU's devices which accelerates the training.
In the training tab, you can change the default “Run Name”.
Also, you can “Save Weights” in three ways :
“End of epoch”
“Best Accuracy”
“Lowest Loss"
Select CPU or GPU whichever you have by clicking on the drop-down button of “Select Devices”.
You can choose multiple devices (“CPU’s” or “GPUs”) at the same time.
You can also start new training by “Loading the previous weight”.
While training is going on you will see the graph of Accuracy and Loss.
After completion of every batch, you will see the changes in Accuracy and Loss graph.
While training is in progress, you can stop the training in between if you wish.
Note : Early stopping will stop the model from training before the number of epochs is reached if the model stops improving.
suppose you are stop the training after 3 epoch. This means that after 3 epochs in a row in which the model doesn’t improve, training will stop.
On the training screen you can able to see the :
“Processing Speed”
“Epoch Speed”
"Loss"
"Accuracy"
“Load Memory” for CPU and GPU
“Free Memory” for CPU and GPU
“Elapsed Time” for training
“Epochs” .
Last updated