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!

156 Revīzijas (346f99fa83068bf07bc6099f6d3691f5e071ef13)

Autors SHA1 Ziņojums Datums
  Stanislaw Adaszewski 346f99fa83 Introduce fixed_unigram_candidate_sampler_new(). pirms 4 gadiem
  Stanislaw Adaszewski 58b0aac18b Fixes in sampling. pirms 4 gadiem
  Stanislaw Adaszewski 8fbad74dfa Debug negative sampling. pirms 4 gadiem
  Stanislaw Adaszewski 9defd6c9e4 Add test_dropout. pirms 4 gadiem
  Stanislaw Adaszewski 9ff1038728 Add extra checks in Data. pirms 4 gadiem
  Stanislaw Adaszewski 7e83fed37b Add test_normalize. pirms 4 gadiem
  Stanislaw Adaszewski e184699365 Add test_split_data_01/02(). pirms 4 gadiem
  Stanislaw Adaszewski 6d27e26fa7 Add test_split_edge_type_01/02/03/04(). pirms 4 gadiem
  Stanislaw Adaszewski 604b81675e Add test_split_adj_mat_(01|02|03)(). pirms 4 gadiem
  Stanislaw Adaszewski 7fa7b7372c Work on loop, split and sampling. pirms 4 gadiem
  Stanislaw Adaszewski 2260f55399 Add test_merge_pos_neg_batch_01/02(). pirms 4 gadiem
  Stanislaw Adaszewski cee0c23bd7 Add test_cumcount. pirms 4 gadiem
  Stanislaw Adaszewski c206df638a Add tests for _same_data_org() and DualBatcher. pirms 4 gadiem
  Stanislaw Adaszewski 8f86fc0833 More batcher tests. pirms 4 gadiem
  Stanislaw Adaszewski 14e3d14b36 Add Batcher. pirms 4 gadiem
  Stanislaw Adaszewski 56fdd9ffeb Make negative sampling more correct and more efficient at the same time. pirms 4 gadiem
  Stanislaw Adaszewski bf39e7d71f Small fix. pirms 4 gadiem
  Stanislaw Adaszewski 4ed3715b83 Add cumcount(). pirms 4 gadiem
  Stanislaw Adaszewski e557b18762 Add trainprep and negative_sample_(adj_mat|data)(). pirms 4 gadiem
  Stanislaw Adaszewski 7ea82a6267 Add test_model_decode_01(). pirms 4 gadiem
  Stanislaw Adaszewski b1a5f0d0ee Add common_one_hot_encoding(). pirms 4 gadiem
  Stanislaw Adaszewski bd8143469c Make work convolve() in new model. pirms 4 gadiem
  Stanislaw Adaszewski b607a727ae Rename _per_layer_required_rows to _vertices. pirms 4 gadiem
  Stanislaw Adaszewski 4ed6626e02 Work on required vertices per layer. pirms 4 gadiem
  Stanislaw Adaszewski f8a02191cb _clear_adjacency_matrix_except_rows() might yet be fast enough on the GPU. pirms 4 gadiem
  Stanislaw Adaszewski 4f953fd203 Work on triacontagon. pirms 4 gadiem
  Stanislaw Adaszewski 2eff854467 Start working on FastLoop. pirms 4 gadiem
  Stanislaw Adaszewski 8edf8ce4f9 Add test_fast_model_01(). pirms 4 gadiem
  Stanislaw Adaszewski b9b97f3dd7 Add test_fast_conv_layer_01() and test_fast_conv_layer_02(). pirms 4 gadiem
  Stanislaw Adaszewski c6f4b8779d Add test_fast_graph_conv_03() and test_fast_graph_conv_04(). pirms 4 gadiem
  Stanislaw Adaszewski 1a303f1a51 Add test_fast_graph_conv_01() and test_fast_graph_conv_02(). pirms 4 gadiem
  Stanislaw Adaszewski 271fba0004 Add test_cat_01() and test_cat_02(). pirms 4 gadiem
  Stanislaw Adaszewski 5e2818fb8d Add test_sparse_diag_cat_02(). pirms 4 gadiem
  Stanislaw Adaszewski 058d4d43fb Add _sparse_diag_cat(). pirms 4 gadiem
  Stanislaw Adaszewski 5bc276eb6b Add BatchedData support to DecodeLayer. pirms 4 gadiem
  Stanislaw Adaszewski c210be41b3 Add databatch. pirms 4 gadiem
  Stanislaw Adaszewski cef1d1110a Add more tests in test_bulkdec. pirms 4 gadiem
  Stanislaw Adaszewski e74eb25a31 Add bulkdec. pirms 4 gadiem
  Stanislaw Adaszewski fc8f9726af Remove unnecessary .detach() call in dropout_dense(). pirms 4 gadiem
  Stanislaw Adaszewski 1ee89509ee Fixes for bool sparse tensors. pirms 4 gadiem
  Stanislaw Adaszewski 4a9ddd540f Fix in test_decode_layer_05(). pirms 4 gadiem
  Stanislaw Adaszewski 94314d2e52 Fix in test_decagon_layer_04(). pirms 4 gadiem
  Stanislaw Adaszewski cd4c34ea7d Fix for non-leaf .grad warning. pirms 4 gadiem
  Stanislaw Adaszewski 431c871af7 Add test_timing_05(). pirms 4 gadiem
  Stanislaw Adaszewski 13940671cc Performance tests. pirms 4 gadiem
  Stanislaw Adaszewski a5b8701a0d Add some debug output for profiling, the bottleneck is in DecodeLayer but also comes generally from computing always all nodes. pirms 4 gadiem
  Stanislaw Adaszewski 3391e15da1 Fix for sparse_coo_tensor(). pirms 4 gadiem
  Stanislaw Adaszewski 0a0a524dc8 Fixes for the GPU. pirms 4 gadiem
  Stanislaw Adaszewski d4dab52e82 Make _equal() much faster. pirms 4 gadiem
  Stanislaw Adaszewski 1f977eee4c Fix test names in test_trainprep. pirms 4 gadiem