You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I follow the Introduction tutorial of Block(http://blocks.readthedocs.io/en/latest/tutorial.html) and train a model. But how to use it? I mean how to use the model finish the task describe in Introduction tutorial. ------------------------------------------------------------------------------- Training status: batch_interrupt_received: False epoch_interrupt_received: False epoch_started: False epochs_done: 1 iterations_done: 235 received_first_batch: True resumed_from: None training_started: True Log records from the iteration 235: test_cost_with_regularization: 0.657896590087 training_finish_requested: True training_finished: True
The text was updated successfully, but these errors were encountered:
Sorry for a late response, everyone is busy nowadays.
Blocks just wraps Theano, so you can compile a function just like in Theano and run it for a single example. Take a look at the Model class, it is basically a computational graph. You can use its inputs and outputs to compile a function.
To apply your model to a test set, use the monitoring extension. You can run it after training (after_training=True flag for the constructor).
I follow the Introduction tutorial of Block(http://blocks.readthedocs.io/en/latest/tutorial.html) and train a model. But how to use it? I mean how to use the model finish the task describe in Introduction tutorial.
-------------------------------------------------------------------------------
Training status:
batch_interrupt_received: False
epoch_interrupt_received: False
epoch_started: False
epochs_done: 1
iterations_done: 235
received_first_batch: True
resumed_from: None
training_started: True
Log records from the iteration 235:
test_cost_with_regularization: 0.657896590087
training_finish_requested: True
training_finished: True
The text was updated successfully, but these errors were encountered: