66584dea87782aed5509e4269a9f015002e1f5c1,pytext/torchscript/module.py,PyTextEmbeddingModuleWithDense,forward,#PyTextEmbeddingModuleWithDense#,1232

Before Change

            languages=squeeze_1d(None),
        )
        // call model
        dense_feat = self.normalizer.normalize(dense_feat)
        dense_tensor = torch.tensor(dense_feat, dtype=torch.float)
        if self.tensorizer.device != "":
            dense_tensor = dense_tensor.to(self.tensorizer.device)

        sentence_embedding = self._forward(inputs, dense_tensor)
        if self.concat_dense:
            return torch.cat([sentence_embedding, dense_tensor], 1)
        else:
            return sentence_embedding

    @torch.jit.script_method

After Change

        result = self.forward_impl(texts[:max_batch], dense_feat[:max_batch])

        if input_len > max_batch:
            texts = texts[max_batch:]
            dense_feat = dense_feat[max_batch:]

            while len(texts) > 0:
                result_extension = self.forward_impl(
                    texts[:max_batch], dense_feat[:max_batch]
                )
                // the result of forward is either a torch.Tensor or a List[Any]
                if isinstance(result, torch.Tensor):
                    result = torch.cat([result, result_extension], dim=0)
                else:
                    result.extend(result_extension)

                texts = texts[max_batch:]
                dense_feat = dense_feat[max_batch:]

        return result

    @torch.jit.script_method
    def make_prediction(
        self,

Instances