Comments (2)
I wanted to PR this change but I need to sign some agreement to do this and I am not sure about the details there so I'll just post the changes since they are small:
in qrnn.py add the following class:
class BiDirQRNNLayer(nn.Module):
def __init__(self, input_size, hidden_size=None, save_prev_x=False, zoneout=0, window=1, output_gate=True,
use_cuda=True):
super(BiDirQRNNLayer, self).__init__()
assert window in [1,
2], "This QRNN implementation currently only handles convolutional window of size 1 or size 2"
self.window = window
self.input_size = input_size
self.hidden_size = hidden_size if hidden_size else input_size
self.zoneout = zoneout
self.save_prev_x = save_prev_x
self.prevX = None
self.output_gate = output_gate
self.use_cuda = use_cuda
self.forward_qrnn = QRNNLayer(input_size, hidden_size=hidden_size, save_prev_x=save_prev_x, zoneout=zoneout, window=window,
output_gate=output_gate, use_cuda=use_cuda)
self.backward_qrnn = QRNNLayer(input_size, hidden_size=hidden_size, save_prev_x=save_prev_x, zoneout=zoneout, window=window,
output_gate=output_gate, use_cuda=use_cuda)
def forward(self, X, hidden=None):
if not hidden is None:
fwd, h_fwd = self.forward_qrnn(X, hidden=hidden)
bwd, h_bwd = self.backward_qrnn(torch.flip(X, [0]), hidden=hidden)
else:
fwd, h_fwd = self.forward_qrnn(X)
bwd, h_bwd = self.backward_qrnn(torch.flip(X, [0]))
bwd = torch.flip(bwd, [0])
return torch.cat([fwd, bwd], dim=-1), torch.cat([h_fwd, h_bwd], dim=-1)
in the same file in the "class QRNN(torch.nn.Module):"
replace :
self.layers = torch.nn.ModuleList(
layers if layers else [QRNNLayer(input_size if l == 0 else hidden_size, hidden_size, **kwargs) for l in
range(num_layers)])
with :
if bidirectional:
self.layers = torch.nn.ModuleList(
layers if layers else [BiDirQRNNLayer(input_size if l == 0 else hidden_size*2, hidden_size, **kwargs) for l in
range(num_layers)])
else:
self.layers = torch.nn.ModuleList(
layers if layers else [QRNNLayer(input_size if l == 0 else hidden_size, hidden_size, **kwargs) for l in
range(num_layers)])
and remove the assert statement above that deals with bidirectional.:
assert bidirectional == False, 'Bidirectional QRNN is not yet supported'
from pytorch-qrnn.
I didn't test super thoroughly but it works for me on a basic use case. If you need to fix something please post it here also :)
@salesforce do feel free to incorporate this into your code. I don't think any paper work is needed for this.
from pytorch-qrnn.
Related Issues (20)
- Problem with QRNN num_layers=2, layers=None, and input_size != hidden_size HOT 1
- Bad squeeze in CPUForgetMult HOT 2
- RuntimeError: matrix and matrix expected HOT 1
- Could you update the code to Pytorch 0.4?
- Any updates on bidirectional QRNN?
- Strings are encoded twice by both QRNN and Pynvrtc? HOT 1
- RuntimeError: size mismatch if use the window size of 2
- Is there a sample dataset to demo the project on seq2seq model HOT 1
- Can QRNN be used in a online manner?
- Error when running your example and on AWD-LSTM-LM HOT 3
- ForgetMult equation in code is different from the paper HOT 2
- AttributeError: 'bytes' object has no attribute 'encode' HOT 4
- [WinError 126] The specified module could not be found - Any idea of the error source? HOT 1
- Legacy autograd Runtime error HOT 2
- Is there tensorflow code for QRNN?
- Support local development by removing dependency on PyCy when not used
- Multi-GPU [Torch DataParallel] HOT 7
- Package install ASCII error from long_description=open('README')? HOT 3
- Error in executing QRNN HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from pytorch-qrnn.