|
Changing warp-ctc by PyTorch CTCLoss provides blank label prediction
|
|
2
|
615
|
January 5, 2022
|
|
Adding tgt_mask or tgt_key_padding_mask gives nan output
|
|
0
|
33
|
January 4, 2022
|
|
Seq2Seq best loss function for bag of sentence output
|
|
0
|
33
|
January 3, 2022
|
|
How to implement BERT using torch.nn.Transformer?
|
|
3
|
630
|
January 3, 2022
|
|
Attention does not seem to be applied at TransformerEncoderLayer and MultiheadAttention PyTorch
|
|
1
|
36
|
January 2, 2022
|
|
Implementing custom loss function in pytorch with weights
|
|
6
|
89
|
January 1, 2022
|
|
Trained a BERT model on my domain specific dataset. Got improvement in scores. However, when combined together with TFIDF, the trained BERT model performs poorer than the untrained BERT model. Why?
|
|
0
|
56
|
January 1, 2022
|
|
Vector using pre-trained Glove vector
|
|
0
|
37
|
December 29, 2021
|
|
How to transpose a tensor and apply it to cross entropy loss?
|
|
2
|
76
|
December 29, 2021
|
|
Constructing a tensor by taking mean over index list of another tensor
|
|
1
|
61
|
December 28, 2021
|
|
CapaGNN program: No module named 'torch_scatter.scatter_cuda'
|
|
1
|
66
|
December 24, 2021
|
|
Extracting self-attention maps from nn.TransformerEncoder
|
|
1
|
85
|
December 22, 2021
|
|
Weird behavior when rying to export fp16 model to onnx
|
|
0
|
44
|
December 22, 2021
|
|
Constant training/validation accuracy
|
|
3
|
76
|
December 21, 2021
|
|
Loss.backward() - IndexError: select(): index 1 out of range for tensor of size [1, 32, 100] at dimension 0
|
|
13
|
149
|
December 21, 2021
|
|
Encoder-decoder text summarisation
|
|
0
|
61
|
December 21, 2021
|
|
Problem in finetuning the pretrained model
|
|
1
|
67
|
December 21, 2021
|
|
Loss Decreasing but Accuracy Remaining the Sane
|
|
0
|
53
|
December 20, 2021
|
|
Train multi-class dataset NN with one custom embedding layer and one single linear layer
|
|
0
|
50
|
December 20, 2021
|
|
RuntimeError: CUDA error: device-side assert triggered CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect
|
|
4
|
644
|
December 20, 2021
|
|
Model loss difference between BertForSequenceClassification and Bert + nn.Linear
|
|
1
|
217
|
December 15, 2020
|
|
LSTM Autoencoders in pytorch
|
|
1
|
92
|
December 19, 2021
|
|
Correct way to connect an RNN to a Linear layer
|
|
4
|
80
|
December 18, 2021
|
|
Tensor and Sequence dataset
|
|
3
|
70
|
December 17, 2021
|
|
Using different feature size between source and target nn.Transformer
|
|
1
|
79
|
December 17, 2021
|
|
Getting probability score
|
|
12
|
161
|
December 17, 2021
|
|
IndexError: select(): index #num out of range for tensor of size #num at dimension 0 in backward for rnn based model
|
|
1
|
1272
|
December 16, 2021
|
|
Trying to use a Transformer for a simple contrived task doesn't work
|
|
2
|
80
|
December 15, 2021
|
|
Transformers in O(log(n))
|
|
0
|
59
|
December 15, 2021
|
|
RuntimeError: Expected hidden[0] size (2, 20, 256), got (2, 50, 256)
|
|
13
|
11703
|
December 13, 2021
|
