frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-20 21:14:14 +08:00
Code Issues Packages Projects Releases Wiki Activity

quantize_rnn_modules in ensemble_export (#20365)

Browse Source 
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20365

Enable quantization of `torch.nn.LSTM` module in the decoder of PyTorch natively exported beam search models.

Reviewed By: jmp84

Differential Revision: D15260631

fbshipit-source-id: bbdd3a30c2c2110986eb7aa7ff11ce1c9090ddf4
This commit is contained in:
James Cross
2019-05-10 15:30:28 -07:00
committed by Facebook Github Bot
parent a0c2829194
commit a4ae689636
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
torch/jit/quantized.py
View File

@ -378,7 +378,7 @@ class QuantizedLSTM(QuantizedRNNBase):
self.check_forward_args(input, hx, batch_sizes)
assert batch_sizes is None
result = _VF.quantized_lstm(input, hx, self._get_all_weights(), self.bias, self.num_layers,
self.dropout, self.training, self.bidirectional,
float(self.dropout), self.training, self.bidirectional,
self.batch_first)
output = result[0]