FM / layers

/usr/local/lib64/python3.6/site-packages/caffe2/python/layers/

/usr/local/lib64/python3.6/site-packages/caffe2/python/layers

Name	Size	Mode	Actions
__pycache__/	-	0755	rm
adaptive_weight.py	5687	0644	edit dl rm
add_bias.py	1396	0644	edit dl rm
arc_cosine_feature_map.py	7345	0644	edit dl rm
batch_huber_loss.py	3523	0644	edit dl rm
batch_lr_loss.py	11577	0644	edit dl rm
batch_mse_loss.py	2333	0644	edit dl rm
batch_normalization.py	3823	0644	edit dl rm
batch_sigmoid_cross_entropy_loss.py	1483	0644	edit dl rm
batch_softmax_loss.py	4580	0644	edit dl rm
blob_weighted_sum.py	2219	0644	edit dl rm
bpr_loss.py	1499	0644	edit dl rm
bucket_weighted.py	2355	0644	edit dl rm
build_index.py	1937	0644	edit dl rm
concat.py	4849	0644	edit dl rm
constant_weight.py	1208	0644	edit dl rm
conv.py	5050	0644	edit dl rm
dropout.py	1410	0644	edit dl rm
fc.py	9296	0644	edit dl rm
fc_without_bias.py	1954	0644	edit dl rm
fc_with_bootstrap.py	12788	0644	edit dl rm
feature_sparse_to_dense.py	14361	0644	edit dl rm
functional.py	4875	0644	edit dl rm
gather_record.py	3260	0644	edit dl rm
homotopy_weight.py	4306	0644	edit dl rm
label_smooth.py	3507	0644	edit dl rm
last_n_window_collector.py	2392	0644	edit dl rm
layers.py	17412	0644	edit dl rm
layer_normalization.py	4291	0644	edit dl rm
margin_rank_loss.py	1951	0644	edit dl rm
merge_id_lists.py	1500	0644	edit dl rm
pairwise_similarity.py	3549	0644	edit dl rm
position_weighted.py	2066	0644	edit dl rm
random_fourier_features.py	3187	0644	edit dl rm
reservoir_sampling.py	3013	0644	edit dl rm
sampling_train.py	2210	0644	edit dl rm
sampling_trainable_mixin.py	1366	0644	edit dl rm
select_record_by_context.py	2381	0644	edit dl rm
semi_random_features.py	5809	0644	edit dl rm
sparse_dropout_with_replacement.py	3943	0644	edit dl rm
sparse_feature_hash.py	4618	0644	edit dl rm
sparse_itemwise_dropout_with_replacement.py	3944	0644	edit dl rm
sparse_lookup.py	22170	0644	edit dl rm
split.py	2257	0644	edit dl rm
tags.py	4114	0644	edit dl rm
uniform_sampling.py	2779	0644	edit dl rm
__init__.py	943	0644	edit dl rm

Edit: /usr/local/lib64/python3.6/site-packages/caffe2/python/layers/batch_softmax_loss.py (4580B)

## @package batch_softmax_loss
# Module caffe2.python.layers.batch_softmax_loss

from caffe2.python import core, schema
from caffe2.python.layers.layers import ModelLayer
import numpy as np

class BatchSoftmaxLoss(ModelLayer):
    def __init__(
        self,
        model,
        input_record,
        name='batch_softmax_loss',
        label_smoothing_matrix=None,
        label_prob=False,
        scale=1.0,
        average_by_batch_size=False,
        **kwargs
    ):
        super(BatchSoftmaxLoss, self).__init__(
            model, name, input_record, **kwargs)

assert schema.is_schema_subset(
            schema.Struct(
                ('label', schema.Scalar()),
                ('prediction', schema.Scalar()),
            ),
            input_record
        )
        self.label_prob = label_prob
        self.scale = scale
        self.average_by_batch_size = average_by_batch_size

# label smoothing matrix: a K * K matrix where K is the label
        # cardinality; (i, j) element is the value of for label i
        # treated/smoothed as label j
        self.label_smoothing_matrix = label_smoothing_matrix
        if self.label_smoothing_matrix is not None:
            self.initialize_label_smoothing_constants()

self.output_schema = schema.Struct(
            (
                'softmax', schema.Scalar(
                    input_record.prediction.field_type(),
                    self.get_next_blob_reference('softmax')
                )
            ),
            (
                'loss', schema.Scalar(
                    np.float32, self.get_next_blob_reference('loss')
                )
            ),
        )

def initialize_label_smoothing_constants(self):
        assert self.label_smoothing_matrix is not None
        self.label_smoothing_matrix = np.array(
            self.label_smoothing_matrix).astype(np.float32)
        assert len(self.label_smoothing_matrix.shape) == 2
        label_dim = self.label_smoothing_matrix.shape[0]
        assert label_dim == self.label_smoothing_matrix.shape[1]

self.label_smoothing_matrix = self.model.add_global_constant(
            '%s_label_smoothing_matrix' % self.name,
            array=self.label_smoothing_matrix,
            dtype=np.dtype(np.float32),
        )
        self.label_dim = self.model.add_global_constant(
            '%s_label_dim' % self.name,
            array=label_dim,
            dtype=np.dtype(np.int64),
        )
        # default case: label is given NOT as target distribution
        # but when used in label smoothing, the label must be in probabilities
        self.label_prob = True

def compute_smoothed_label(self, net):
        assert self.label_smoothing_matrix is not None
        label = self.input_record.label()
        original_label_type = self.input_record.label.field_type()
        if original_label_type.base != np.int64:
            int64_label = net.NextScopedBlob('int64_label')
            net.Cast([label], [int64_label], to=core.DataType.INT64)
        else:
            int64_label = label
        one_hot_label = net.NextScopedBlob('one_hot_label')
        smoothed_label = net.NextScopedBlob('smoothed_label')
        net.OneHot([int64_label, self.label_dim], [one_hot_label])
        net.MatMul([one_hot_label, self.label_smoothing_matrix], smoothed_label)
        return smoothed_label

def add_ops(self, net):
        label = self.input_record.label.field_blobs()
        if self.label_smoothing_matrix is not None:
            label = [self.compute_smoothed_label(net)]
        elif not self.label_prob:
            if self.input_record.label.field_types()[0].base != np.int32:
                label = [
                    net.Cast(label,
                             net.NextScopedBlob('int32_label'),
                             to=core.DataType.INT32)
                ]

softmax_input = self.input_record.prediction.field_blobs() + label

if 'weight' in self.input_record:
            weight_blob = self.input_record.weight()
            if self.input_record.weight.field_type().base != np.float32:
                weight_blob = net.Cast(
                    weight_blob,
                    weight_blob + '_float32',
                    to=core.DataType.FLOAT
                )

softmax_input += [weight_blob]

net.SoftmaxWithLoss(
            softmax_input,
            self.output_schema.field_blobs(),
            label_prob=self.label_prob,
            scale=self.scale,
            average_by_batch_size=self.average_by_batch_size,
        )

## @package batch_softmax_loss # Module caffe2.python.layers.batch_softmax_loss from caffe2.python import core, schema from caffe2.python.layers.layers import ModelLayer import numpy as np class BatchSoftmaxLoss(ModelLayer): def __init__( self, model, input_record, name='batch_softmax_loss', label_smoothing_matrix=None, label_prob=False, scale=1.0, average_by_batch_size=False, **kwargs ): super(BatchSoftmaxLoss, self).__init__( model, name, input_record, **kwargs) assert schema.is_schema_subset( schema.Struct( ('label', schema.Scalar()), ('prediction', schema.Scalar()), ), input_record ) self.label_prob = label_prob self.scale = scale self.average_by_batch_size = average_by_batch_size # label smoothing matrix: a K * K matrix where K is the label # cardinality; (i, j) element is the value of for label i # treated/smoothed as label j self.label_smoothing_matrix = label_smoothing_matrix if self.label_smoothing_matrix is not None: self.initialize_label_smoothing_constants() self.output_schema = schema.Struct( ( 'softmax', schema.Scalar( input_record.prediction.field_type(), self.get_next_blob_reference('softmax') ) ), ( 'loss', schema.Scalar( np.float32, self.get_next_blob_reference('loss') ) ), ) def initialize_label_smoothing_constants(self): assert self.label_smoothing_matrix is not None self.label_smoothing_matrix = np.array( self.label_smoothing_matrix).astype(np.float32) assert len(self.label_smoothing_matrix.shape) == 2 label_dim = self.label_smoothing_matrix.shape[0] assert label_dim == self.label_smoothing_matrix.shape[1] self.label_smoothing_matrix = self.model.add_global_constant( '%s_label_smoothing_matrix' % self.name, array=self.label_smoothing_matrix, dtype=np.dtype(np.float32), ) self.label_dim = self.model.add_global_constant( '%s_label_dim' % self.name, array=label_dim, dtype=np.dtype(np.int64), ) # default case: label is given NOT as target distribution # but when used in label smoothing, the label must be in probabilities self.label_prob = True def compute_smoothed_label(self, net): assert self.label_smoothing_matrix is not None label = self.input_record.label() original_label_type = self.input_record.label.field_type() if original_label_type.base != np.int64: int64_label = net.NextScopedBlob('int64_label') net.Cast([label], [int64_label], to=core.DataType.INT64) else: int64_label = label one_hot_label = net.NextScopedBlob('one_hot_label') smoothed_label = net.NextScopedBlob('smoothed_label') net.OneHot([int64_label, self.label_dim], [one_hot_label]) net.MatMul([one_hot_label, self.label_smoothing_matrix], smoothed_label) return smoothed_label def add_ops(self, net): label = self.input_record.label.field_blobs() if self.label_smoothing_matrix is not None: label = [self.compute_smoothed_label(net)] elif not self.label_prob: if self.input_record.label.field_types()[0].base != np.int32: label = [ net.Cast(label, net.NextScopedBlob('int32_label'), to=core.DataType.INT32) ] softmax_input = self.input_record.prediction.field_blobs() + label if 'weight' in self.input_record: weight_blob = self.input_record.weight() if self.input_record.weight.field_type().base != np.float32: weight_blob = net.Cast( weight_blob, weight_blob + '_float32', to=core.DataType.FLOAT ) softmax_input += [weight_blob] net.SoftmaxWithLoss( softmax_input, self.output_schema.field_blobs(), label_prob=self.label_prob, scale=self.scale, average_by_batch_size=self.average_by_batch_size, )