06c9cefed73cfd43f9453c616e1b9d3ef63f58cf,fairseq/models/speech_to_text/s2t_transformer.py,S2TTransformerEncoder,forward,#S2TTransformerEncoder#Any#Any#,297

Before Change

            self.layer_norm = None

    def forward(self, src_tokens, src_lengths):
        x, input_lengths = self.subsample(src_tokens, src_lengths)
        x = self.embed_scale * x

        encoder_padding_mask = lengths_to_padding_mask(input_lengths)
        positions = self.embed_positions(encoder_padding_mask).transpose(0, 1)
        x += positions
        x = self.dropout_module(x)

        for layer in self.transformer_layers:
            x = layer(x, encoder_padding_mask)

        if self.layer_norm is not None:
            x = self.layer_norm(x)

        return {
            "encoder_out": [x],  // T x B x C
            "encoder_padding_mask": [encoder_padding_mask] if encoder_padding_mask.any() else [],  // B x T
            "encoder_embedding": [],  // B x T x C
            "encoder_states": [],  // List[T x B x C]
            "src_tokens": [],
            "src_lengths": [],
        }

    def reorder_encoder_out(self, encoder_out, new_order):
        new_encoder_out = (
            [] if len(encoder_out["encoder_out"]) == 0

After Change

    def forward(self, src_tokens, src_lengths):
        if self.num_updates < self.encoder_freezing_updates:
            with torch.no_grad():
                x = self._forward(src_tokens, src_lengths)
        else:
            x = self._forward(src_tokens, src_lengths)
        return x

    def reorder_encoder_out(self, encoder_out, new_order):
        new_encoder_out = (
            [] if len(encoder_out["encoder_out"]) == 0

Instances