sadaszewski1 - ADARED:Code

IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an email to s dot adaszewski at gmail dot com. User accounts are meant only to report issues and/or generate pull requests. This is a purpose-specific Git hosting for ADARED projects. Thank you for your understanding!

sadaszewski1

Joined on May 07, 2020

Loading Heatmap…

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

5cf1e3d1b5 torch_stablesort CUDA version works and is fast.
897a6f0722 torch_stablesort_cuda.cu getting there...
Compare 2 commits »

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

c1353168f3 CUDA performance needs improvement.

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

62af8fc5f1 Progress with CUDA version of torch_stablesort.

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

f8f2901eec Start working on CUDA implementation of torch_stablesort.

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

b439a46fc3 Use if constexpr to make stable_sort_impl() more compact.

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

283d9387a8 Add torch_stablesort.md.

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

87953842a6 Try to implement gradient.

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

747156a971 torch_stablesort_cpp seems to work.
14321daf62 Improved dispatch.
Compare 2 commits »

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

37c5db1330 Start torch_stablesort.0

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

356f3af3c3 With this weird little trick the new fixed_unigram_candidate_sampler() finally seems to be fully robust and to actually work.

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

79b0890a70 Mini fix.
d7d442c5e3 New fixed_unigram_candidate_sampler() still requires work.
490b4f9281 New robust implementation of fixed_unigram_candidate_sampler() seems to be working.
689dbeed20 Add mergesort switch.
670237c3f8 Work on fixed_unigram_candidate_sampler_new().
Compare 11 commits »

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

9ff1038728 Add extra checks in Data.

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

7e83fed37b Add test_normalize.

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

e184699365 Add test_split_data_01/02().

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

6d27e26fa7 Add test_split_edge_type_01/02/03/04().
604b81675e Add test_split_adj_mat_(01|02|03)().
7094ae87db Fix negative_sample_adj_mat().
7fa7b7372c Work on loop, split and sampling.
2260f55399 Add test_merge_pos_neg_batch_01/02().
Compare 5 commits »

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

cee0c23bd7 Add test_cumcount.

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

ebcd38c910 Add first triacontagon experiment.

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

4c8c06c63c Add triacontagon.TrainLoop.

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

c206df638a Add tests for _same_data_org() and DualBatcher.
6b15ec8c10 Finish DualBatcher.
0a061af526 Attempt DualBatcher.
Compare 3 commits »

3 years ago

sadaszewski1 pushed to master at sadaszewski1/decagon-pytorch

8f86fc0833 More batcher tests.

3 years ago