MCPcopy Create free account
hub / github.com/TaoRuijie/TalkNet-ASD / forward

Method forward

model/visualEncoder.py:119–130  ·  view source on GitHub ↗
(self, inputBatch)

Source from the content-addressed store, hash-verified

117
118
119 def forward(self, inputBatch):
120 inputBatch = inputBatch.transpose(0, 1).transpose(1, 2)
121 batchsize = inputBatch.shape[0]
122 batch = self.frontend3D(inputBatch)
123
124 batch = batch.transpose(1, 2)
125 batch = batch.reshape(batch.shape[0]*batch.shape[1], batch.shape[2], batch.shape[3], batch.shape[4])
126 outputBatch = self.resnet(batch)
127 outputBatch = outputBatch.reshape(batchsize, -1, 512)
128 outputBatch = outputBatch.transpose(1 ,2)
129 outputBatch = outputBatch.transpose(1, 2).transpose(0, 1)
130 return outputBatch
131
132class DSConv1d(nn.Module):
133 def __init__(self):

Callers

nothing calls this directly

Calls

no outgoing calls

Tested by

no test coverage detected