Skip to content

issues Search Results · repo:harvardnlp/pytorch-struct language:"Jupyter Notebook"

Filter by

55 results
 (74 ms)

55 results

inharvardnlp/pytorch-struct (press backspace or delete to remove)

harvardnlp/pytorch-struct
Broken Alignmnent for CTC example ?

Hi, Firstly, just wanted to say this is a really cool library. I have been working on some CTC/alignment research and when i saw this trick with the parallel-scan and semi-ring it struck me as a very ...
  • angusturner
  • 2
  • Opened 
    on Jul 5, 2023
  • #127

harvardnlp/pytorch-struct
[Question] How to apply pytorch-struct for 2 dimensional data?

I could find examples of pytorch struct usage for 1d sequence data like text or video frame. But I m trying to parse tables structure in pdf documents. Could you provide some hints where to start?
  • YuriyPryyma
  • 4
  • Opened 
    on Feb 9, 2022
  • #126

harvardnlp/pytorch-struct
end_class support for Autoregressive

end_class is not used for the Autoregressive module: https://github.com/harvardnlp/pytorch-struct/blob/7146de5659ff17ad7be53023c025ffd099866412/torch_struct/autoregressive.py#L49
  • urchade
  • 1
  • Opened 
    on Feb 4, 2022
  • #125

harvardnlp/pytorch-struct
Update examples to use newer torchtext APIs

I m in the process of updating the pytorch benchmarks repo to use newer torchtext APIs. These benchmarks use an adapted version of this example notebook. If it makes sense, I m happy to submit a PR to ...
  • erip
  • 2
  • Opened 
    on Dec 27, 2021
  • #122

harvardnlp/pytorch-struct
RuntimeError when using pytorch-struct on GPU

When I try to use topk or kmax on the GPU, the following error occurs: ipdb dist.kmax(5) *** RuntimeError: Expected condition, x and y to be on the same device, but condition is on cpu and x and y are ...
  • goncalomcorreia
  • 2
  • Opened 
    on Dec 20, 2021
  • #121

harvardnlp/pytorch-struct
pytorch-struct shows a warning for torch versions >= 1.8.0

When using pytorch-struct while using torch =1.8.0, the following warning appears: In [1]: import torch ...: from torch_struct import DependencyCRF, LinearChainCRF In [2]: vals = torch.zeros(2, 10, ...
  • goncalomcorreia
  • 2
  • Opened 
    on Dec 20, 2021
  • #120

harvardnlp/pytorch-struct
Instable learning with SemiMarkov CRF

HI, First, thank you for fixing #110 (@da03), the SemiCRF works better now, I was able to get good results on span extraction tasks. However, I still encounter a learning instability where the loss (neg ...
  • urchade
  • Opened 
    on Oct 27, 2021
  • #117

harvardnlp/pytorch-struct
Low memory implementations

A lot of the approaches could also have low-memory versions (like CKY does). Figure out how to enable that in a clean manner (v0.6)
  • srush
  • Opened 
    on Oct 20, 2021
  • #115

harvardnlp/pytorch-struct
Device mismatch in LinearChainCRF

This line below causes device mismatch error if chart is in GPU because init will always be created in CPU. https://github.com/harvardnlp/pytorch-struct/blob/e51fecc1473925e4c44de135c4a3240fcb20fa40/torch_struct/linearchain.py#L56 ...
  • kmkurn
  • 1
  • Opened 
    on Oct 4, 2021
  • #111

harvardnlp/pytorch-struct
Mini-batch setting with Semi Markov CRF

I encounter learning instability when using a batch size 1 with the semi-markovian CRF (loss goes to very large negative number), even when explicitly providing lengths . I think the bug comes from the ...
  • urchade
  • 5
  • Opened 
    on Oct 2, 2021
  • #110
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Press the
/
key to activate the search input again and adjust your query.
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Press the
/
key to activate the search input again and adjust your query.
Issue search results · GitHub