kmkurn/pytorch-crf

Issues

• Cross Entropy as a loss function

  #60 opened 3 years ago
  2

• issue with imbalanced data

  #58 opened 5 years ago
  1

• Support half precission

  #56 opened 5 years ago
  0

• loss unstable

  #55 opened 5 years ago
  2

• Transition score for masked timesteps

  #53 opened 3 years ago
  3

• what pad symbol to use in tags tensor

  #52 opened 5 years ago
  3

• Get the score for the best_tags_list

  #51 opened 3 years ago
  4

• so has anyone got this pytorch crf to work yet ?

  #50 opened 5 years ago
  12

• About model's output

  #49 opened 5 years ago
  2

• How to get probability for each tag when Viterbi decode?

  #48 opened 5 years ago
  2

• How can one add weights for different tags?

  #47 opened 3 years ago
  5

• Why can't I mask the first token?

  #46 opened 5 years ago
  5

• The f1 score only increases when the batch size =1

  #45 opened 5 years ago
  2

• Unexpected IndexError is result of dtype change

  #44 opened 5 years ago
  3

• Computer Vision usage

  #43 opened 5 years ago
  3

• Loss not decreasing!

  #42 opened 3 years ago
  2

• Index out of bound error

  #41 opened 3 years ago
  4

• Loss decreases but f1 score remains unchanged

  #40 opened 3 years ago
  10

• Torch 1.1.0 Support?

  #39 opened 5 years ago
  7

• RuntimeError: tensors used as indices must be long or byte tensors

  #38 opened 4 years ago
  5

• How can we get the probabilities out?

  #36 opened 3 years ago
  1

• Size of the model with pytorch-crf.

  #35 opened 5 years ago
  2

• Multi-label results?

  #34 opened 5 years ago
  2

• Error with _validate on gpu

  #33 opened 5 years ago
  2

• some thing wrong

  #32 opened 3 years ago
  2

• The decoding module does not support multiple GPUs.

  #31 opened 3 years ago
  2

• A question about "denominator" obtained by the function <_compute_normalizer>

  #30 opened 6 years ago
  5

• [Question] Actual usage examples?

  #29 opened 6 years ago
  7

• is it necessary to use format function which is only supported by 3.6?

  #28 opened 6 years ago
  1

• Multiply -1 before loss.backward()?

  #27 opened 6 years ago
  2

• START TAG and STOP TAG?

  #23 opened 6 years ago
  1

• Allow parameters to be manually set

  #19 opened 3 years ago
  0

• transfer to pytorch 0.4

  #15 opened 6 years ago
  1

• Setup proper documentation

  #14 opened 6 years ago
  0

• Test for Python 3.6.x

  #13 opened 6 years ago
  1

• Test for Pytorch 0.3.x

  #12 opened 6 years ago
  0

• Add `batch_first` argument to `__init__`

  #11 opened 6 years ago
  0

• Add `size_average` argument in `forward`

  #10 opened 6 years ago
  1

• Add shape comments on every line that produces tensor

  #9 opened 6 years ago
  0

• Move dims and sizes checking to another method

  #8 opened 6 years ago
  0

• Fix `_new` return type.

  #7 opened 6 years ago
  0

• Rename kwargs to match PyTorch's convention

  #2 opened 7 years ago
  0

• Unchecked Initialize params in Init

  #1 opened 7 years ago
  0