-
Currently the library only allows the creation of basic feedforward networks (essentially a stack of layers stored as a list). But some tasks requires a more complex architecture with mixed connection…
-
I'm not an expert, but it looks like it's a feedforward neural network. Am I missing something? In any case, keep up the good work! So glad this code exists.
-
I used the C++ API and found the second derivative of the path parameter is not equal to zero at the start and end of the computed trajectory which is not friendly to the acceleration feedforward cont…
-
Hi,
I was not able to find proper place to post this questions. Please give me an insight.
I have Lasagne network and can export network as a list of lists using
lasagne.layers.get_all_param_values…
-
The following minimal program illustrates two alignment problems with decorators put on connections:
On the feedback path, the decorator is put above the connection rather than on it, and on th…
-
import torch
import torch.nn as nn
import nltk
from gensim.models import Word2Vec
from nltk.tokenize import word_tokenize, sent_tokenize
from collections import Counter
nltk.download('stopwords')
nl…
-
Hi, I have read with a lot if interest you [report on DAS](https://github.com/wedeling/EasySurrogate/blob/master/tutorials/deep_active_subspaces/report_DAS.pdf). I can see that it ends with some impor…
-
It would be nice to have a "reworked" [as in https://github.com/k2-fsa/icefall/pull/288] version of the CTC+attention setup (in a separate directory).
That would be a good reference for cases where w…
-
### 🐛 Describe the bug
Get orch.onnx.errors.UnsupportedOperatorError: Exporting the operator 'aten::unflatten' to ONNX opset version 12 is not supported.
When I execute
`torch.onnx.export(t…
-
# ❓ Questions and Help
Hi,
I've installed the xformers library (windows, install via pip)
When running the microGPT.py script, I have this error (see trace bellow). What's the problem?
regards,
…