pydbm.loss package¶
Subpackages¶
Submodules¶
pydbm.loss.cross_entropy module¶
-
class
pydbm.loss.cross_entropy.
CrossEntropy
¶ Bases:
pydbm.loss.interface.computable_loss.ComputableLoss
Cross Entropy.
-
compute_delta
¶ Backward delta.
Parameters: - pred_arr – Predicted data.
- labeled_arr – Labeled data.
- delta_output – Delta.
Returns: Delta.
-
compute_loss
¶ Return of result from this Cost function.
Parameters: - pred_arr – Predicted data.
- labeled_arr – Labeled data.
- axis – Axis or axes along which the losses are computed. The default is to compute the losses of the flattened array.
Returns: Cost.
-
pydbm.loss.kl_divergence module¶
-
class
pydbm.loss.kl_divergence.
KLDivergence
¶ Bases:
pydbm.loss.interface.computable_loss.ComputableLoss
Kullback–Leibler Divergence (KLD).
-
compute_delta
¶ Backward delta.
Parameters: - pred_arr – Predicted data.
- labeled_arr – Labeled data.
- delta_output – Delta.
Returns: Delta.
-
compute_loss
¶ Return of result from this Cost function.
Parameters: - pred_arr – Predicted data.
- labeled_arr – Labeled data.
- axis – Axis or axes along which the losses are computed. The default is to compute the losses of the flattened array.
Returns: Cost.
-
pydbm.loss.mean_squared_error module¶
-
class
pydbm.loss.mean_squared_error.
MeanSquaredError
¶ Bases:
pydbm.loss.interface.computable_loss.ComputableLoss
The mean squared error (MSE).
References
- Pascanu, R., Mikolov, T., & Bengio, Y. (2012). Understanding the exploding gradient problem. CoRR, abs/1211.5063, 2.
- Pascanu, R., Mikolov, T., & Bengio, Y. (2013, February). On the difficulty of training recurrent neural networks. In International conference on machine learning (pp. 1310-1318).
-
compute_delta
¶ Backward delta.
Parameters: - pred_arr – Predicted data.
- labeled_arr – Labeled data.
- delta_output – Delta.
Returns: Delta.
-
compute_loss
¶ Return of result from this Cost function.
Parameters: - pred_arr – Predicted data.
- labeled_arr – Labeled data.
- axis – Axis or axes along which the losses are computed. The default is to compute the losses of the flattened array.
Returns: Cost.