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
/
decagon-pytorch
forked from sadaszewski/decagon-pytorch
1
0
Fork 0
Code Issues 0 Pull Requests 0 Releases 0 Wiki Activity

Commit Graph

  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • 5cf1e3d (HEAD -> master) torch_stablesort CUDA version works and is fast. by Stanislaw Adaszewski 2020-09-01 18:32:58 +0200
  • 897a6f0 torch_stablesort_cuda.cu getting there... by Stanislaw Adaszewski 2020-09-01 18:10:36 +0200
  • c135316 CUDA performance needs improvement. by Stanislaw Adaszewski 2020-09-01 13:08:14 +0200
  • 62af8fc Progress with CUDA version of torch_stablesort. by Stanislaw Adaszewski 2020-09-01 11:24:07 +0200
  • f8f2901 Start working on CUDA implementation of torch_stablesort. by Stanislaw Adaszewski 2020-08-31 20:40:55 +0200
  • b439a46 Use if constexpr to make stable_sort_impl() more compact. by Stanislaw Adaszewski 2020-08-28 22:31:44 +0200
  • 283d938 Add torch_stablesort.md. by Stanislaw Adaszewski 2020-08-24 18:31:26 +0200
  • 8795384 Try to implement gradient. by Stanislaw Adaszewski 2020-08-24 13:28:53 +0200
  • 747156a torch_stablesort_cpp seems to work. by Stanislaw Adaszewski 2020-08-24 10:21:06 +0200
  • 14321da Improved dispatch. by Stanislaw Adaszewski 2020-08-21 23:31:02 +0200
  • 37c5db1 Start torch_stablesort.0 by Stanislaw Adaszewski 2020-08-21 22:56:28 +0200
  • 356f3af With this weird little trick the new fixed_unigram_candidate_sampler() finally seems to be fully robust and to actually work. by Stanislaw Adaszewski 2020-08-21 18:51:27 +0200
  • 79b0890 Mini fix. by Stanislaw Adaszewski 2020-08-21 17:34:15 +0200
  • d7d442c New fixed_unigram_candidate_sampler() still requires work. by Stanislaw Adaszewski 2020-08-21 17:33:13 +0200
  • 490b4f9 New robust implementation of fixed_unigram_candidate_sampler() seems to be working. by Stanislaw Adaszewski 2020-08-21 17:24:07 +0200
  • 689dbee Add mergesort switch. by Stanislaw Adaszewski 2020-08-21 17:19:45 +0200
  • 670237c Work on fixed_unigram_candidate_sampler_new(). by Stanislaw Adaszewski 2020-08-21 17:15:58 +0200
  • 346f99f Introduce fixed_unigram_candidate_sampler_new(). by Stanislaw Adaszewski 2020-08-21 17:00:46 +0200
  • 6cf8251 fixed_unigram_candidate_sampler() still requires some work to be perfect. by Stanislaw Adaszewski 2020-08-21 12:44:08 +0200
  • 58b0aac Fixes in sampling. by Stanislaw Adaszewski 2020-08-21 11:07:27 +0200
  • 2ff358f Protect against endless loop. by Stanislaw Adaszewski 2020-08-20 22:06:40 +0200
  • 8fbad74 Debug negative sampling. by Stanislaw Adaszewski 2020-08-20 18:56:52 +0200
  • 9defd6c Add test_dropout. by Stanislaw Adaszewski 2020-08-20 18:43:20 +0200
  • 9ff1038 Add extra checks in Data. by Stanislaw Adaszewski 2020-08-20 18:30:12 +0200
  • 7e83fed Add test_normalize. by Stanislaw Adaszewski 2020-08-20 17:30:28 +0200
  • e184699 Add test_split_data_01/02(). by Stanislaw Adaszewski 2020-08-20 17:00:51 +0200
  • 6d27e26 Add test_split_edge_type_01/02/03/04(). by Stanislaw Adaszewski 2020-08-20 13:41:47 +0200
  • 604b816 Add test_split_adj_mat_(01|02|03)(). by Stanislaw Adaszewski 2020-08-20 12:45:51 +0200
  • 7094ae8 Fix negative_sample_adj_mat(). by Stanislaw Adaszewski 2020-08-20 12:26:06 +0200
  • 7fa7b73 Work on loop, split and sampling. by Stanislaw Adaszewski 2020-08-20 12:21:32 +0200
  • 2260f55 Add test_merge_pos_neg_batch_01/02(). by Stanislaw Adaszewski 2020-08-20 10:30:59 +0200
  • cee0c23 Add test_cumcount. by Stanislaw Adaszewski 2020-08-19 12:56:30 +0200
  • ebcd38c Add first triacontagon experiment. by Stanislaw Adaszewski 2020-08-18 17:44:57 +0200
  • 4c8c06c Add triacontagon.TrainLoop. by Stanislaw Adaszewski 2020-08-18 16:53:00 +0200
  • c206df6 Add tests for _same_data_org() and DualBatcher. by Stanislaw Adaszewski 2020-08-18 14:46:23 +0200
  • 6b15ec8 Finish DualBatcher. by Stanislaw Adaszewski 2020-08-18 13:59:29 +0200
  • 0a061af Attempt DualBatcher. by Stanislaw Adaszewski 2020-08-18 12:25:29 +0200
  • 8f86fc0 More batcher tests. by Stanislaw Adaszewski 2020-08-10 17:58:42 +0200
  • 14e3d14 Add Batcher. by Stanislaw Adaszewski 2020-08-10 17:55:06 +0200
  • 56fdd9f Make negative sampling more correct and more efficient at the same time. by Stanislaw Adaszewski 2020-08-07 18:45:54 +0200
  • bf39e7d Small fix. by Stanislaw Adaszewski 2020-08-07 16:35:18 +0200
  • 4ed3715 Add cumcount(). by Stanislaw Adaszewski 2020-08-07 16:28:05 +0200
  • e557b18 Add trainprep and negative_sample_(adj_mat|data)(). by Stanislaw Adaszewski 2020-08-07 14:50:32 +0200
  • 7ea82a6 Add test_model_decode_01(). by Stanislaw Adaszewski 2020-08-07 11:21:58 +0200
  • b1a5f0d Add common_one_hot_encoding(). by Stanislaw Adaszewski 2020-08-07 09:56:50 +0200
  • 82d5c06 Fix aggregation of representations from different edge types. by Stanislaw Adaszewski 2020-08-06 15:25:50 +0200
  • bd81434 Make work convolve() in new model. by Stanislaw Adaszewski 2020-08-06 15:12:09 +0200
  • b607a72 Rename _per_layer_required_rows to _vertices. by Stanislaw Adaszewski 2020-08-06 11:58:10 +0200
  • 4ed6626 Work on required vertices per layer. by Stanislaw Adaszewski 2020-08-06 10:24:06 +0200
  • f8a0219 _clear_adjacency_matrix_except_rows() might yet be fast enough on the GPU. by Stanislaw Adaszewski 2020-08-04 15:10:27 +0200
  • 4f953fd Work on triacontagon. by Stanislaw Adaszewski 2020-08-04 14:11:59 +0200
  • d57af9c Start triacontagon. by Stanislaw Adaszewski 2020-08-03 19:12:28 +0200
  • 2eff854 Start working on FastLoop. by Stanislaw Adaszewski 2020-07-30 12:46:20 +0200
  • 8edf8ce Add test_fast_model_01(). by Stanislaw Adaszewski 2020-07-28 12:17:00 +0200
  • f99f0bb Add FastModel. by Stanislaw Adaszewski 2020-07-28 11:39:34 +0200
  • b9b97f3 Add test_fast_conv_layer_01() and test_fast_conv_layer_02(). by Stanislaw Adaszewski 2020-07-28 10:20:23 +0200
  • aa3cc1f Reimplement FastConvLayer. by Stanislaw Adaszewski 2020-07-28 10:20:09 +0200
  • c6f4b87 Add test_fast_graph_conv_03() and test_fast_graph_conv_04(). by Stanislaw Adaszewski 2020-07-27 17:01:53 +0200
  • 1a303f1 Add test_fast_graph_conv_01() and test_fast_graph_conv_02(). by Stanislaw Adaszewski 2020-07-24 14:20:29 +0200
  • fcaccc8 Fix a typo. by Stanislaw Adaszewski 2020-07-24 13:28:11 +0200
  • e2ae97a Work on FastGraphConv. by Stanislaw Adaszewski 2020-07-24 13:26:30 +0200
  • 271fba0 Add test_cat_01() and test_cat_02(). by Stanislaw Adaszewski 2020-07-24 12:27:21 +0200
  • 5e2818f Add test_sparse_diag_cat_02(). by Stanislaw Adaszewski 2020-07-24 11:27:16 +0200
  • 058d4d4 Add _sparse_diag_cat(). by Stanislaw Adaszewski 2020-07-24 11:02:09 +0200
  • 3ecc5a7 Update matrix-multiply. by Stanislaw Adaszewski 2020-07-23 20:57:03 +0200
  • 5770f7c Update matrix-multiply with an idea that will actually work. by Stanislaw Adaszewski 2020-07-23 20:46:31 +0200
  • dd41ba9 Add matrix-multiply illustration. by Stanislaw Adaszewski 2020-07-22 14:28:17 +0200
  • 36fe246 Start working on fastconv. by Stanislaw Adaszewski 2020-07-22 12:51:30 +0200
  • 54c7f1a Fix a typo. by Stanislaw Adaszewski 2020-07-21 15:01:33 +0200
  • 5bc276e Add BatchedData support to DecodeLayer. by Stanislaw Adaszewski 2020-07-21 14:56:43 +0200
  • 45a18a4 Add shuffle to DataBatcher. by Stanislaw Adaszewski 2020-07-21 14:26:16 +0200
  • c210be4 Add databatch. by Stanislaw Adaszewski 2020-07-21 14:14:07 +0200
  • cef1d11 Add more tests in test_bulkdec. by Stanislaw Adaszewski 2020-07-21 12:51:51 +0200
  • e74eb25 Add bulkdec. by Stanislaw Adaszewski 2020-07-21 10:57:50 +0200
  • fc8f972 Remove unnecessary .detach() call in dropout_dense(). by Stanislaw Adaszewski 2020-07-21 09:08:20 +0200
  • 47645ef Add Citing note. by Stanislaw Adaszewski 2020-07-17 18:46:49 +0200
  • 6cf3175 Remove an extra newline. by Stanislaw Adaszewski 2020-07-17 14:17:01 +0200
  • 1ee8950 Fixes for bool sparse tensors. by Stanislaw Adaszewski 2020-07-17 13:59:30 +0200
  • 4a9ddd5 Fix in test_decode_layer_05(). by Stanislaw Adaszewski 2020-07-17 13:27:30 +0200
  • 94314d2 Fix in test_decagon_layer_04(). by Stanislaw Adaszewski 2020-07-17 13:26:27 +0200
  • cd4c34e Fix for non-leaf .grad warning. by Stanislaw Adaszewski 2020-07-17 13:20:06 +0200
  • 431c871 Add test_timing_05(). by Stanislaw Adaszewski 2020-07-17 12:58:39 +0200
  • 1394067 Performance tests. by Stanislaw Adaszewski 2020-07-17 11:55:38 +0200
  • 82c0b5d Update .gitignore by Stanislaw Adaszewski 2020-07-16 21:35:39 +0200
  • a5b8701 Add some debug output for profiling, the bottleneck is in DecodeLayer but also comes generally from computing always all nodes. by Stanislaw Adaszewski 2020-07-16 21:32:43 +0200
  • 749ff00 Shuffle. by Stanislaw Adaszewski 2020-07-16 20:47:57 +0200
  • b9e7e39 Add shuffle param to TrainLoop. by Stanislaw Adaszewski 2020-07-16 20:23:26 +0200
  • 3391e15 Fix for sparse_coo_tensor(). by Stanislaw Adaszewski 2020-07-16 20:21:47 +0200
  • 0a0a524 Fixes for the GPU. by Stanislaw Adaszewski 2020-07-16 18:41:40 +0200
  • 055b4b3 run_epoch() progresses at a crawl. by Stanislaw Adaszewski 2020-07-16 16:49:19 +0200
  • a34a0f5 Done data loading in decagon_run. by Stanislaw Adaszewski 2020-07-16 15:15:40 +0200
  • 8aafb6f Clean up _equal(). by Stanislaw Adaszewski 2020-07-16 14:09:20 +0200
  • d4dab52 Make _equal() much faster. by Stanislaw Adaszewski 2020-07-16 14:08:37 +0200
  • 373a12f Start working on experiments/decagon_run. by Stanislaw Adaszewski 2020-07-16 12:44:48 +0200
  • 59edeaa Add docker/. by Stanislaw Adaszewski 2020-07-15 16:58:03 +0200
  • 1f977ee Fix test names in test_trainprep. by Stanislaw Adaszewski 2020-06-20 18:18:38 +0200
  • 2cb561b Add test_prepare_relation_type_[01,02](). by Stanislaw Adaszewski 2020-06-20 18:16:49 +0200
  • e2a57e8 Add test_prep_rel_two_node_types_asym_01(). by Stanislaw Adaszewski 2020-06-20 18:00:15 +0200
  • 5c24883 Add test_prep_rel_two_node_types_sym_01(). by Stanislaw Adaszewski 2020-06-20 17:56:20 +0200
  • 2a2aecb Add test_prep_rel_one_node_type_01(). by Stanislaw Adaszewski 2020-06-20 17:50:13 +0200