repo
stringlengths 8
116
| tasks
stringlengths 8
117
| titles
stringlengths 17
302
| dependencies
stringlengths 5
372k
| readme
stringlengths 5
4.26k
| __index_level_0__
int64 0
4.36k
|
|---|---|---|---|---|---|
emerali/LottoRBM | ['network pruning'] | ['The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks'] | masked_rbm.py MaskedBinaryRBM | # LottoRBM Playing around with applying the concept of Lottery Tickets to RBMs Refs: https://arxiv.org/abs/1803.03635 and https://arxiv.org/abs/1905.01067 Requires the current master branch of [QuCumber](https://github.com/PIQuIL/QuCumber) installed, along with the data for [Tutorial 1](https://github.com/PIQuIL/QuCumber/tree/master/examples/Tutorial1_TrainPosRealWaveFunction). | 2,000 |
emmanueliarussi/generative3DSpongiosa | ['style transfer'] | ['Generative Modelling of 3D in-silico Spongiosa with Controllable Micro-Structural Parameters'] | code/utils3d.py code/optimize_sample.py code/synthesize_random_samples.py code/train.py code/sp.py code/layers.py code/trabecuar_dataset.py code/progressive_model3d.py PixelNormLayer conv WScaleLayer optimize_sample optimizer Generator param_number Discriminator _softplus2 VCNoSD _var VC _mean ReducedVC synthesize BonesPatchDataset train saveTensorBatch hypersphere denormalize_volume GradientPenalty weights_init exp_mov_avg normalize_volume Progress layer activ BatchNorm3d OrderedDict PixelNormLayer WScaleLayer range MSELoss clamp_ LBFGS netGs step content_weight opt_steps pretrained_model_path use_pixel_norm flatten use_weight_scale output_dir device save in_nch target_bmd target_bvtv load_state_dict to target_tmd format requires_grad_ eval use_batch_norm optimizer load join print reshape netGs numpy max_res makedirs parameters float type FloatTensor _mean mul _softplus2 _var sigmoid sqrt _mean cat enumerate mul _softplus2 sigmoid _mean cat enumerate size expand VC range cat randn use_pixel_norm flatten use_weight_scale output_dir device save in_nch load_state_dict to range hypersphere format eval model_path use_batch_norm load join print num_samples netGs numpy max_res makedirs format set_title add_subplot zfill imshow savefig figure range data netGs batchSize batch_size randn zero_grad use_pixel_norm use_weight_scale DataLoader model_checkpoints save lambda_gradient_penalty device resize in_nch saveTensorBatch batch_sizes p progress len step Adam netD append to sum state_dict format hypersphere n_iter gamma_gradient_penalty Compose mean lr GP item exp_mov_avg listdir Progress use_batch_norm enumerate deepcopy join BonesPatchDataset GradientPenalty backward print data_path parameters netG epsilon_drift max_res epochs array makedirs data bias normal_ xavier_normal_ weight constant_ data min add_ parameters zip | # Generative Modelling of 3D in-silico Spongiosa with Controllable Micro-Structural Parameters Code, data, and trained models from the [MICCAI 2020](https://miccai2020.org/en/default.asp) paper. We developed a method to create micro-structural bone samples in-silico that 1) contain realistic structures, and 2) allow to steer the properties to simulate changes of micro-structural parameters from deterioration or medical bone treatment.  ## Requirements Python 3.+ and PyTorch 1.6.0. The code has been tested only with PyTorch 1.6.0, there are no guarantees that it is compatible with older versions. ## Clone repo ``` $ git clone https://github.com/emmanueliarussi/generative3DSpongiosa.git $ cd generative3DSpongiosa | 2,001 |
emorynlp/bert-2019 | ['semantic parsing', 'part of speech tagging'] | ['Establishing Strong Baselines for the New Decade: Sequence Tagging, Syntactic and Semantic Parsing with BERT'] | bertsota/tagger/sequence_tagger_trainer.py test/script/wsj_noword_fe.py test/script/news_noword.py test/script/dm_pas_sharernn.py test/script/psd_id_500_rel.py test/script/ctb_nodropouttag.py test/script/train_embedding_psd_id.py test/script/debug_psd.py test/script/debug_chinese.py test/script/shrink_emb.py test/script/bert_psd_noword.py test/script/debug-refine.py test/script/bert_large_dm.py test/script/dm_pas.py test/script/evaluate.py test/script/overfit_psd_id.py bertsota/common/k_means.py test/script/ctb_word2vec.py bertsota/common/tarjan.py bertsota/common/data.py test/script/ptb_noword.py test/script/stair.py test/script/fasttext_dm.py test/script/ptb_notag.py test/script/bert_dm.py test/script/ctb_pos_bert.py test/preprocess/replace_pos.py test/script/wsj_att_fe.py test/script/debug_refine.py test/preprocess/en_bert.py bertsota/common/config.py test/script/ctb_noword.py test/script/bert_dm_noword.py test/preprocess/dump_conll03.py test/script/ctb_pos.py test/script/pas_psd_sharernn.py test/script/dm_psd.py test/script/wsj_belc_fe_way.py test/script/news.py test/script/beta_psd_id.py test/script/refine_bert_large.py test/script/ctb_debug.py test/script/bert_psd_base_sum.py test/script/bert_psd_large_sum.py test/script/eval_psd_id.py test/script/bert_large_psd.py test/script/ontoen_bert.py bertsota/parser/evaluate/__init__.py test/preprocess/cn_bert.py test/script/linear_psd_id.py test/script/text.py test/preprocess/extract_words_sdp2015.py test/script/cz.py test/script/evaluate_pos.py test/script/pas10.py test/script/conll03_be.py test/script/refine_bert_base_sum.py test/script/bert_psd.py test/script/ctb_pos_att_noword.py test/script/wsj_bert.py test/script/share_rnn_bert.py test/script/wsj_belc_fe_2layer.py bertsota/common/exponential_scheduler.py test/script/pas_psd_transfer_dm.py test/script/dm.py test/script/conll03_be_pe.py test/script/ontoen_pe_fe.py test/script/wsj_belc_ge.py test/script/wsj_bebu.py test/script/cz_noword.py test/script/__init__.py test/script/ptb_auto.py test/script/make_ptb.py test/script/ptbrefine.py test/script/conll03.py test/script/bert_pas_noword.py test/script/dm_pas_transfer_psd_refine.py test/script/ptb_noword_auto.py test/script/blend.py test/script/news_bert.py test/script/refine.py test/script/wsj_bebu_ge.py test/script/ctb.py test/script/simple_refine.py bertsota/common/__init__.py test/script/share_rnn.py bertsota/tagger/__init__.py test/script/bert_large_pas.py test/preprocess/extract_sent.py test/script/half_refine.py test/script/wsj_belc_fe.py test/preprocess/extract_words_sdp_cn.py bertsota/tagger/reduce_lr_on_plateau.py test/preprocess/limit_length.py test/script/refine_bert_large_sum.py test/script/form_psd_id.py test/script/conll03_be_fe.py test/script/eval_pas.py test/script/ctb_old_data.py test/script/text_noword.py bertsota/common/utils.py test/script/psd10.py test/preprocess/split_train_dev.py test/script/wsj_ge_fe.py test/script/verify_data.py test/script/ctb_bigtag.py test/script/double_cnnrefine.py test/script/psd_id_less_eval_every.py test/script/ptb_bert.py bertsota/tagger/embeddings.py test/script/dump_fasttext.py test/script/share_rnn_bert_large.py test/script/small_lr_psd.py test/script/ctb_pos_att.py test/script/conll03_pe_fe.py test/script/pas.py test/script/dm_pas_transfer_psd.py test/script/dm10.py test/script/refine_bert.py bertsota/parser/joint_parser.py bertsota/parser/evaluate/evaluate.py test/script/psd.py bertsota/parser/biaffine_parser.py test/script/wsj_belc_ge_fe.py test/script/pas_psd.py bertsota/tagger/lm_config.py test/__init__.py test/script/cz2.py test/script/dm_psd_transfer_pas.py test/script/wsj_noword.py bertsota/tagger/contextual_string_model.py bertsota/tagger/sequence_tagger_model.py test/script/cz_bert.py bertsota/parser/dep_parser.py test/script/conll03_be_pe_fe.py test/script/bert_pas.py test/script/sp.py test/script/wsj_bebu_ge_fe.py test/script/ctb_bert_notag.py bertsota/parser/__init__.py test/script/bert_psd_sum.py test/script/ptb.py test/script/wsj_bert_fe.py test/script/gold_arc_psd_id.py test/script/eval_dm.py test/script/tensor.py test/script/ontoen_bert_fe.py test/script/wsj_att_belc.py test/script/wsj_debug.py bertsota/tagger/corpus.py test/script/ctb_pos_noword.py test/script/wsj.py test/script/ptb_bert_auto.py bertsota/common/savable.py bertsota/__init__.py test/script/debug_bert.py test/script/linearrefine.py test/script/ctb_bert.py test/preprocess/extract_words_ctb_pos.py test/script/dm_psd_sharernn.py test/script/cnnrefine.py test/script/text_bert.py test/preprocess/enbert.py test/script/wsj_belc.py test/script/ctb_notag.py _Config DepDataLoader ConllSentence split_train_dev slice_train_file ParserVocabulary ConllWord DataLoader ExponentialScheduler KMeans Savable Tarjan leaky_relu arc_argmax reshape_fortran biLSTM embedding_from_numpy parameter_init freeze to_nd orthonormal_VanillaLSTMBuilder orthonormal_initializer rel_argmax init_logger flatten_numpy bilinear mxnet_prefer_gpu parameter_from_numpy Progbar BiAffineDep DoubleBiAffine RNN BasicParser StairParser BlendParser MLP BiaffineDepParser BiaffineParser SharedRNNParser UnlabeledBiAffine RefineParser BiAffine SharedPrivateRNNParser SDPParser DepParser JointParser prf evaluate_chinese_sdp output_sent evaluate_official_script compute_F1 dep_evaluate_official_script evaluate_joint_official_script ContextualStringModelTrainer ContextualStringModel _load _train _convert_dumped_model Word Label Token ioblu_iobes Vocabulary iob2 StringIdMapper iob_iobes TaggedCorpus get_chunks read_pretrained_embeddings TSVCorpus NLPTaskDataFetcher TextCorpus Sentence NLPTask Dictionary make_language_model_dataset StackedEmbeddings CharLMEmbeddings Embeddings TwoWayEmbeddings BERTEmbeddings TokenEmbeddings WordEmbeddings LanguageModelConfig ReduceLROnPlateau SequenceTagger log_sum_exp argmax_batch pad_tensors to_scalar LockedDropout argmax log_sum_exp_batch SequenceTaggerTrainer Metric embed_last_token embed_sum generate_bert dump embed_bert make_bert_for make_bert_text embed_sum make_bert_for make_bert_text embed extract_sent extract_words extract_words extract_words limit_wsj limit_msr limit_ctb limit_onto replace load dump stat_gold_and_test_data stat_one_tree stat_one_node_heads_and_deprels eval text align load dump get int join setFormatter format getLogger addHandler makedirs StreamHandler Formatter setLevel INFO FileHandler orthonormal_initializer zeros VariationalDropoutCell LSTMCell concat unroll zip flip Dropout reshape_fortran ones transpose concat dot batch_dot randn dot sqrt eye sum range pop update list remove arange set Tarjan add eye argmax array SCCs len ROOT argmax arange Embedding setattr get get setattr ndarray isinstance idx_sequence DepDataLoader time get_batches join format remove name reduce system dirname forward len idx_sequence sum time get_batches prf print transpose equal greater DataLoader compute_F1 zip forward asnumpy len tag2id idx_sequence time get_batches transpose range DataLoader zip forward asnumpy len F1 forward idx_sequence transpose append sum next format replace zip equal enumerate join time print greater compute_F1 asnumpy len namedtuple join str format write append range load_dumped_model initialize save ContextualStringModelTrainer ContextualStringModel dictionary TextCorpus train load_language_model join int format print close makedirs enumerate split append replace enumerate append replace enumerate append enumerate split exp tile sum max log list fill_ type_ max enumerate zip print startswith split encode append enumerate print exit mean stack startswith zip encode enumerate append split print embedding print dirname makedirs print dirname makedirs zip print startswith split encode append enumerate lower print set print set FastText int print startswith append split zip index len list keys stat_one_node_heads_and_deprels print set | # BERT SOTA Baseline Source code for paper [Establishing Strong Baselines for the New Decade: Sequence Tagging, Syntactic and Semantic Parsing with BERT](https://arxiv.org/pdf/1908.04943.pdf), to be published in The Thirty-Third International Flairs Conference proceedings. ## Requirements - Python>=3.6 - [mxnet](https://mxnet.apache.org/)>=1.4.1 - [pyhanlp](https://github.com/hankcs/pyhanlp) (optional for preprocessing Chinese) ## Datasets See [`data`](https://github.com/emorynlp/bert-2019/tree/master/data). All preprocessing scripts are placed in [`test/preprocess`](https://github.com/emorynlp/bert-2019/tree/master/test/preprocess). ## How to Run All experiment entrypoints are placed in [`test/script`](https://github.com/emorynlp/bert-2019/tree/master/test/script). One example is: | 2,002 |
empty16/hardnet.pytorch | ['image retrieval', 'patch matching'] | ["Working hard to know your neighbor's margins: Local descriptor learning loss"] | examples/extract_hardnet_desc_from_hpatches_file.py hardnet/W1BS.py hardnet/models/HardNet.py benchmarks/hpatches_extract_HardNet.py hardnet/config.py hardnet/dataset.py hardnet/HardNetMultipleDatasets.py hardnet/HardNetHPatchesSplits.py hardnet/dataloaders/HPatchesDatasetCreator.py hardnet/HardNetClassicalHardNegMiningSiftInit.py hardnet/pytorch_sift.py hardnet/Losses.py hardnet/models/HardNet_SRN.py hardnet/models/TFNet.py hardnet/HardNet_exp.py examples/caffe/extract_hardnetCaffe_desc_from_hpatches_file.py hardnet/scripts/check_gor_HardNet.py hardnet/scripts/check_gor_triplet.py hardnet/Loggers.py hardnet/Utils.py hardnet/dataloaders/TotalDataLoader.py hardnet/models/__init__.py hardnet/EvalMetrics.py hardnet/scripts/download_all_datasets.py examples/caffe/convert_weights_to_caffe.py hardnet/HardNetClassicalHardNegMining.py L2Norm L1Norm HardNet HpatchesSequence L1Norm HardNet L2Norm L1Norm copy_weights HardNet L2Norm extract_tfeats preprocess_patch get_config save_config np2torch TripletDataLoader HPatchesDM read_patch_file read_image_dir TripletPhotoTour find_files TotalDatasetsLoader ErrorRateAt95Recall BuildKNNGraphByFAISS_GPU get_hard_negatives PhototourTrainingData create_optimizer showImagesHorizontally create_loaders test pre_init_with_sift remove_descriptors_with_same_index adjust_learning_rate CorrelationPenaltyLoss TripletPhotoTour main train TripletPhotoTourHardNegatives get_descriptors_for_dataset create_optimizer create_loaders test adjust_learning_rate weights_init TripletPhotoTour main TotalDatasetsLoader train HardNet create_optimizer create_loaders test adjust_learning_rate CorrelationPenaltyLoss weights_init TripletPhotoTour main TotalDatasetsLoader train HardNet main create_loaders train test FileLogger Logger loss_random_sampling global_orthogonal_regularization loss_HardNet intermediate_featuremap_regularization distance_matrix_vector CorrelationPenaltyLoss loss_L2Net distance_vectors_pairwise get_bin_weight_kernel_size_and_stride SIFTNet getPoolingKernel create_optimizer L1Norm adjust_learning_rate resume_pretrained_model str2bool L2Norm w1bs_extract_descs_and_save mean_image std_image HPatches TotalDatasetsLoader HardNet weights_init HardNet_SRN weights_init TFNet weights_init ContextManager usage ContextManger usage str isinstance print Conv2d BatchNorm2d numpy astype float32 int reshape ceil zeros float range len parse_known_args print join log_dir experiment_name endswith join listdir append from_numpy rollaxis uint8 size convert astype mean append range len append print read_patch_file find_files cat sum argmax add_subplot axis imshow figure range len remove training_set Compose copy TripletPhotoTour format loss_random_sampling backward decor len zero_grad dataset tqdm set_description adjust_learning_rate save step cuda enumerate format enable_logging print reshape size len tqdm sqrt eval numpy set_description ErrorRateAt95Recall log_value append dataset sum cuda enumerate batch_size param_groups n_triplets lr float epochs parameters Adam SGD GpuIndexFlatConfig search GpuIndexFlatL2 add shape StandardGpuResources data DataLoader resize cuda FloatTensor append expand_dims astype eval SIFT type enumerate PhototourTrainingData Variable SIFTNet float32 tqdm extend numpy data transform PhototourTrainingData model print extend tqdm eval DataLoader append numpy cuda enumerate len append all range len BuildKNNGraphByFAISS_GPU items list print labels remove_descriptors_with_same_index hardnegatives create_indices get_hard_negatives draw_and_save_plots pre_init_with_sift w1bs_extract_descs_and_save DataLoader save features cuda load_state_dict draw_and_save_plots_with_loggers TripletPhotoTourHardNegatives range format create_optimizer replace Compose log_string test get_list_of_patch_images match_descriptors_and_save_results start_epoch eval lr resume vars load enable_logging print isfile train epochs get_descriptors_for_dataset data constant isinstance orthogonal Conv2d TotalDatasetsLoader DataLoader model loss_HardNet model_dir loss_L2Net gor stat enable_logging Variable TripletDataLoader log_string log_string start_epoch epochs resample_dataset_triplets resume_pretrained_model unsqueeze sqrt sum exp print clamp min exit mean distance_vectors_pairwise exp print exit mean distance_matrix_vector sum diag exp view print clamp squeeze type_as exit min gather mean distance_matrix_vector float cuda diag mul clamp size mean pow sum exp view size squeeze log mean sum diag T arange hstack maximum outer floor float int float round floor load join format print model_dir resume load_state_dict isfile experiment_name model resize cuda from_numpy shape savetxt dirname append imread range replace concatenate astype eval int time Variable reshape print float32 zeros makedirs mean std Linear print exit | # HardNet model implementation HardNet model implementation in PyTorch for NIPS 2017 paper ["Working hard to know your neighbor's margins: Local descriptor learning loss"](https://arxiv.org/abs/1705.10872) [poster](http://cmp.felk.cvut.cz/~mishkdmy/posters/hardnet2017.pdf), [slides](http://cmp.felk.cvut.cz/~mishkdmy/slides/HardNet2017.pdf) ## Important update After two major updates of PyTorch library, we were not able to reproduce the results. Therefore, we did a hyperparameter search and we found that increasing the learning rate to 10 and dropout rate to 0.3 leads to better results, see figure below. We have obtained the same results on different machines . We will update arXiv version soon. Pretrained weights in PyTorch format are updated. We also release version, trained on all subsets of Brown dataset, called [HardNet6Brown]('pretrained/6Brown). ## Benchmark on [HPatches](https://github.com/hpatches/hpatches-benchmark), mAP  | 2,003 |
emtiaz-ahmed/COVID19-Twitter-Reopen | ['sentiment analysis'] | ['COVID-19: Social Media Sentiment Analysis on Reopening'] | config.py get_ibm_tone_format_tweet join format replace upper RX_URL sub startswith RX_HASHTAG RX_MENTION RX_EMAIL split | # COVID19-Twitter-Reopen The novel coronavirus (COVID-19) pandemic is the most talked topic in social media platforms in 2020. People are using social media such as Twitter to express their opinion and share information on a number of issues related to the COVID-19 in this stay at home order. In this paper, we investigate the sentiment and emotion of peoples in the United States on the subject of reopening. We choose the social media platform Twitter for our analysis and study the Tweets to discover the sentimental perspective, emotional perspective, and triggering words towards the reopening. During this COVID-19 pandemic, researchers have made some analysis on various social media dataset regarding lockdown and stay at home. However, in our analysis, we are particularly interested to analyse public sentiment on reopening. Our major finding is that when all states resorted to lockdown in March, people showed dominant emotion of fear, but as reopening starts people have less fear. While this may be true, due to this reopening phase daily positive cases are rising compared to the lockdown situation. Overall, people have a less negative sentiment towards the situation of reopening. - - - • Preprint: [COVID-19: Social Media Sentiment Analysis on Reopening](https://arxiv.org/abs/2006.00804) • Data: https://github.com/emtiaz-ahmed/COVID19-Twitter-Reopen/tree/master/Filtered_Dataset - - - | 2,004 |
endo-angel/ct-angel | ['lesion segmentation'] | ['Comparative study of deep learning methods for the automatic segmentation of lung, lesion and lesion type in CT scans of COVID-19 patients'] | ct-angel-train/unet_pp/backbones/classification_models/classification_models/resnet/builder.py ct-angel-train/unet_pp/xnet/blocks.py ct-angel-train/unet_pp/backbones/inception_resnet_v2.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/resnext/models.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/weights.py ct-angel-train/unet_pp/utils.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/resnext/preprocessing.py ct-angel-train/unet_pp/xnet/__init__.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/resnext/__init__.py ct-angel-train/unet_pp/xnet/model.py ct-angel-train/unet_pp/__init__.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/__init__.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/utils.py ct-angel-train/unet_pp/backbones/backbones.py ct-angel-train/unet_pp/backbones/inception_v3.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/resnext/blocks.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/resnet/__init__.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/resnext/builder.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/resnet/preprocessing.py ct-angel-train/unet_pp/backbones/classification_models/tests/test_imagenet.py ct-angel-train/unet_pp/xnet/builder.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/resnext/params.py ct-angel-train/unet_pp/backbones/preprocessing.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/resnet/blocks.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/resnet/models.py ct-angel-train/unet_pp/backbones/__init__.py ct-angel-train/unet_pp/backbones/classification_models/classification_models/resnet/params.py set_trainable extract_outputs to_tuple get_layer_number freeze_model reverse add_docstring recompile get_backbone InceptionResNetV2 preprocess_input inception_resnet_block conv2d_bn InceptionV3 conv2d_bn preprocess_input get_preprocessing load_model_weights find_weights handle_block_names conv_block basic_identity_block basic_conv_block identity_block build_resnet ResNet18 ResNet34 ResNet50 ResNet101 ResNet152 get_bn_params get_conv_params preprocess_input handle_block_names identity_block GroupConv2D conv_block build_resnext ResNeXt50 ResNeXt101 get_bn_params get_conv_params preprocess_input main test_model get_top is_equal handle_block_names Upsample2D_block Transpose2D_block ConvRelu build_xnet Xnet layers enumerate insert output loss metrics optimizer compile layers layers recompile isscalar isinstance str conv2d_bn _obtain_input_shape get_file Input get_source_inputs warn Model conv2d_bn load_weights inception_resnet_block range _obtain_input_shape get_file concatenate get_source_inputs warn Model conv2d_bn load_weights Input range list filter name load_weights get_file find_weights format _obtain_input_shape get_conv_params get_source_inputs Model get_bn_params Input range enumerate load_model_weights build_resnet load_model_weights build_resnet load_model_weights build_resnet load_model_weights build_resnet load_model_weights build_resnet update update resize _obtain_input_shape get_conv_params get_source_inputs Model get_bn_params Input range enumerate load_model_weights build_resnext load_model_weights build_resnext squeeze range len get_top decode_predictions print is_equal preprocessing_func expand_dims predict format model print test_model imread output to_tuple Model input range len freeze_model get_backbone format build_xnet |  # CT-Angel --- ## Table of Contents * Background * Data * Model Development * Usage * Directory Structure * About the author | 2,005 |
enkiwang/Imperceptible-fake-face-antiforensic | ['face detection', 'adversarial attack'] | ['Perception Matters: Exploring Imperceptible and Transferable Anti-forensics for GAN-generated Fake Face Imagery Detection'] | attacks/attack_ens.py models/DiscNet.py models/MobileNet.py attacks/utils.py attacks/attack.py models/FakeFaceNet.py models/__init__.py attacks/eval.py models/AlexNet.py models/VGG.py ycbcr2rgb_np get_file_list load_model read_image_ycbcr get_file_name ycbcr_to_rgb attack rgb2ycbcr_np ycbcr_to_tensor read_image save_image ycbcr2rgb_np get_file_list load_model read_image_ycbcr get_file_name attack_ens ycbcr_to_rgb rgb2ycbcr_np ycbcr_to_tensor read_image save_image get_file_list load_model get_acc read_image save_records get_file_list read_image_pil save_image_pil read_image_np read_sz_save save_image_tensor get_pair_list read_image AlexNet DiscNet FakeNet conv_1x1_bn InvertedResidual conv_bn MobileNetV2 VGGNet glob join sort preprocess unsqueeze open T putmask astype dot float array dot T array rgb2ycbcr_np astype open Tensor transpose view clamp squeeze image_width unsqueeze image_height permute tensor float mm fromarray squeeze transpose astype save numpy clip load to load_state_dict data deepcopy zeros_like model criterion clamp backward abs sign ycbcr_to_rgb sum range detach data deepcopy zeros_like criterion backward clamp abs sign ycbcr_to_rgb sum range detach str max_epsilon2 print model_num max_epsilon3 max_epsilon1 open get_file_list load_model model len eval to max read_image enumerate join get_file_list append range len open open save str get_file_list read_image_pil LANCZOS print save_image_pil resize enumerate fromarray numpy astype save | ## An Imperceptible Anti-forensic Method, Models and Dataset **Perception matters: exploring imperceptible and transferable anti-forensics forGAN-generated fake face imagery detection** <img src="https://github.com/enkiwang/Imperceptible-fake-face-antiforensic/blob/master/example.png/" width=800> This repository contains the code, models and dataset for the project "Perception matters: exploring imperceptible and transferable anti-forensics for GAN-generated fake face imagery detection". ## Implementation This code has been tested on Ubuntu 16.04 system, with following pre-requisites. ### Pre-requisites 1. python >=3.6.10 | 2,006 |
enrongtsai/Horovod-practice | ['stochastic optimization'] | ['Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour'] | solve_cudnn_error.py original_cat_dog.py horovod_cat_dog.py create_model create_model solve_cudnn_error compile size Sequential SGD add DistributedOptimizer Dense MaxPooling2D Conv2D Flatten print list_physical_devices set_memory_growth list_logical_devices | # Horovod practice
This repository keeps a simple note and code for distributed training practice using horovod.
## Intro
### Horovod
Horovod is an open source distributed training framework which supports TensorFlow, Keras, PyTorch and MXNet. It implemented with `OpenMPI` & `ring all-reduce algorithm` and was easy to submit jobs on modern supercomputers. The main benifit is it requires only a few lines of modification to the original code.
### Code modification
- Initializing Horovod: `hvd.init()`
- Control the calling process on specified worker (e.g. worker 0) with horovod rank: `if hvd.rank() == 0`
| 2,007 |
ensv/TS-Net | ['patch matching'] | ['TS-Net: Combining modality specific and common features for multimodal patch matching'] | data/vedai/vedai_dataset.py data/nirscene/nirscene_dataset.py data/abstract_dataset.py network/tsnet.py network/utils.py data/cuhk/cuhk_dataset.py network/abstract.py network/matchnet.py Abstract_dataset NCUHK NNIR NVEDAI AbstractNetwork ErrorRateAt95Recall _parse_function _errorRateAt95 MatchNet main train_multimodal_additional_loss train_multimodal train_multimodal main train_multimodal_additional_loss TSNet check_result_valid print_matchnet_table print_tsnetbackup_size_table print_tsnet_bottleneck correct_path smooth check_done print_tsnet_size_table print_std_mean combine_data correct_cpt_path should_correct_cpt print_matchnetlecun_table _errorRateAt95 uint8 decode_raw int64 cast int32 parse_single_example sorted list sum zip MatchNet process MatchNet process chdir getcwd print MatchNet train_multimodal realpath dirname train_multimodal_additional_loss process TSNet TSNet TSNet print listdir print listdir startswith mean zeros range len print argmin min smooth append listdir exists split readlines abspath listdir should_correct_cpt abspath correct_path print print_std_mean print print_std_mean print print_std_mean print print_std_mean append print_std_mean print load dump | # TS-Net: Combining Modality Specific and Common Features for Multimodal Patch Matching # Setting up the environement: - Python3 - Tensorflow 1.4+ (GPU version only) - Numpy - Scipy - matplotlib - skimage # Dataset: To generate new patches for VeDAI, CUHK and NIR-Scene dataset, run the following command in the "data" folder: | 2,008 |
eogns282/IMODE | ['time series'] | ['Neural Ordinary Differential Equations for Intervention Modeling'] | datasets/exp_decay/decay_loader.py utils/solvers.py datasets/collision/collision_loader.py trainer.py main_eicu.py evaluator.py main_decay.py utils/visualization.py main_collision.py models/imode.py datasets/eICU/eicu_loader.py Evaluator_eicu Evaluator_collision Evaluator_decay boolean_string boolean_string boolean_string Trainer_eicu Trainer_decay Trainer_collision collision_collate collision_dataset eicu_dataset eicu_collate decay_collate decay_dataset ODEFunc ode_function_for_adaptive vis_collision_traj vis_decay_traj dict tensor stack values dict tensor stack values dict tensor stack values str plot close scatter savefig mkdir numpy len str plot close scatter savefig mkdir numpy len |
# Neural Ordinary Differential Equations for Intervention Modeling (IMODE)
This repository is the official implemtation of [Neural Ordinary Differential Equations for Intervention Modeling](https://arxiv.org/abs/2010.08304)
Real-world systems often involves external interventions that cause changes in the system dynamics such as a patient being administered with particular drug. We propose a novel neural ODE-based approach (IMODE) that properly model the effect of external interventions by employing two ODE functions to separately handle the observations and the interventions.
<p align="center">
<b> A simulation of the trajectory in a 2D plane </b>
</p>
| 2,009 |
eowjd0512/VITAMIN-E | ['visual tracking'] | ['VITAMIN-E: VIsual Tracking And MappINg with Extremely Dense Feature Points'] | thirdParty/eigen-master/debug/gdb/printers.py thirdParty/eigen-master/scripts/relicense.py thirdParty/eigen-master/debug/gdb/__init__.py EigenQuaternionPrinter _MatrixEntryIterator EigenSparseMatrixPrinter lookup_function register_eigen_printers build_eigen_dictionary EigenMatrixPrinter update append strip_typedefs search tag target type | Email <[email protected]> My CV <http://eowjd0512.github.io/> # VITAMIN-E Re-implementation of VITAMIN-E SLAM This project was started to re-produce the paper "M. Yokozuka et al, VITAMIN-E: VIsual Tracking And MappINg with Extremely Dense Feature Points, CVPR 2019" <https://arxiv.org/pdf/1904.10324.pdf> ## Requirements | 2,010 |
epfl-dlab/GoogleTrendsAnchorBank | ['time series'] | ['Calibration of Google Trends Time Series'] | gtab/command_line.py gtab/__init__.py gtab/core.py example/generate_leads.py setup.py GroupedAction set_hitraffic init_dir set_options set_blacklist create_gtab set_active_gtab list_gtabs _load_dir_cl delete_gtab rename_gtab new_query print_options GTAB print print add_argument path GTAB abspath ArgumentParser parse_args print _load_dir_cl GTAB add_argument _load_dir_cl GTAB ArgumentParser vars parse_args add_argument blacklist _load_dir_cl GTAB ArgumentParser parse_args hitraffic add_argument _load_dir_cl GTAB ArgumentParser parse_args _load_dir_cl GTAB set_active_gtab src set_active_gtab add_argument dst _load_dir_cl GTAB ArgumentParser parse_args src set_active_gtab add_argument _load_dir_cl GTAB ArgumentParser parse_args src add_argument _load_dir_cl GTAB ArgumentParser parse_args _load_dir_cl GTAB create_anchorbank deepcopy join kws set_active_gtab print add_argument results_file dumps _load_dir_cl GTAB loads ArgumentParser parse_args makedirs |  _[For a full technical description, see [this paper](https://arxiv.org/abs/2007.13861). When using `gtab` in your own work, please cite the paper.]_ # Quick install The package is available on pip, so you just need to call ~~~python python -m pip install gtab ~~~ The explicit list of requirements can be found in [`requirements.txt`](requirements.txt). The code was developed and tested in Python 3.8.1. # Which problem does `gtab` solve? | 2,011 |
epfl-dlab/KLearn | ['text summarization'] | ['KLearn: Background Knowledge Inference from Summarization Data'] | klearn/algorithms.py klearn/IT.py klearn/utils.py experiments/learn_optimal_K.py experiments/learn_annotator_K.py experiments/learn_domain_K.py klearn/Convertor.py model_comparison.py klearn/__init__.py klearn/KModels.py add_JS_KL_baselines cross_validation_split run_CV_experiment chunk_lst write write_K write_K optimal_K hPL load_idfs cut_off unique_key_set MS_U restrict_key_set extract_tensors MS_D baseline get_batch_PL IDF_baseline DatasetLDA LDAConvertor DataConvertor NgramConvertor pre_process_documents Entropy JS theta avg normalize theta_lst KL Cross_Entropy kl_lst js_lst ThetaPL evaluate_topic evaluate_JS_KL_baseline convertor_load convertor_loads ConvertorUnpickler load_data to_serializable evaluate_K evaluate_MR ts_float32 extend append evaluate_K evaluate_MR method enumerate append float len append list keys chunk_lst append evaluate_JS_KL_baseline dict join print hPL write_K append range flush all_keys items list extend sum dict len restrict_key_set unique_key_set list sum array values restrict_key_set unique_key_set items sorted list unique_key_set restrict_key_set normalize_keysets zeros sum array len items sorted list min divide unique_key_set annotations restrict_key_set normalize_keysets zeros sum array len items list get_vectors reshape smooth min_freq repeat normalize_keysets max len list choice vstack sample float keys range append set_detect_anomaly zero_grad SGD numpy theta sorted list restrict_key_set extract_tensors range format unique_key_set ThetaPL zip get_batch_PL BCELoss criterion backward print dict step len append get_text dict sum min values mean list union list values items list get items list get union keys set append KL Entropy append append BytesIO doc_freq beta annotations K alpha append theta_lst items list doc_freq annotations append theta_lst items sorted list doc_freq annotations index append theta_lst range items sorted list annotations index append doc_freq range kl_lst js_lst | # KLearn Code accompanying the paper: `KLearn: Background Knowledge Inference from Summarization Data` published at *Findings of EMNLP 2020* ### Abstract The goal of text summarization is to compress documents to the relevant information while excluding background information already known to the receiver. So far, summarization researchers have given considerably more attention to relevance than to background knowledge. In contrast, this work puts background knowledge in the foreground. Building on the realization that the choices made by human summarizers and annotators contain implicit information about their background knowledge, we develop and compare techniques for inferring background knowledge from summarization data. Based on this framework, we define summary scoring functions that explicitly model background knowledge, and show that these scoring functions fit human judgments significantly better than baselines. We illustrate some of the many potential applications of our framework. First, we provide insights into human information importance priors. | 2,012 |
epfl-lts2/mulan | ['information retrieval'] | ['MULAN: A Blind and Off-Grid Method for Multichannel Echo Retrieval'] | measurement_tools.py fri.py noise_tools.py algorithm.py test_tools.py test_set.py baseline_methods.py estimate_filters_from_signal get_initial_value get_reconstruction_error run_benchmark mulan InitializationType get_relative_delays_and_amplitudes run estimate_signal_from_filters polyscale cross_relations LASSO_approach LASSO CR pick_echoes get_annihilating_filter get_weights get_locations FilterType load_experimental_rirs InputSignalType select_frequency_set get_spectral_coefficients load_simulated_rirs get_artificial_input_signal get_measurements generate_artificial_rirs load_input_signal denoise add_white_noise benchmark phase_transition save_results_to_latex process_results init_folder_structure get_reconstruction_error save_recovery_details_3_methods run_experiments save_figure_and_data load_data save_recovery_details print_experiment_parameters run_experiments_benchmark list range normal norm exp multiply pi range mean shape power sum zeros len sort argsort amin shape zeros abs max range norm estimate_filters_from_signal get_relative_delays_and_amplitudes get_initial_value roots multiply power flipud get_locations info zeros get_weights range estimate_signal_from_filters polyscale multiply toeplitz flipud get_annihilating_filter zeros range concatenate diag hstack toeplitz get_annihilating_filter zeros range roots divide flipud abs range poly get_measurements range add_white_noise mean_squared_error abs range convolve LASSO get_spectral_coefficients CR load_simulated_rirs get_artificial_input_signal get_reconstruction_error mulan generate_artificial_rirs linspace zeros load_input_signal range len LASSO_approach concatenate min info pick_echoes cross_relations concatenate min info pick_echoes max coef_ concatenate reshape toeplitz Lasso fit toeplitz reshape matmul concatenate sort argsort append zeros array range svd reshape vander reshape transpose flipud len list range list range load_experimental_rirs get_spectral_coefficients load_simulated_rirs get_artificial_input_signal generate_artificial_rirs load_input_signal linspace convolve exp select_frequency_set multiply pi linspace zeros sum range len int sort rand min maximum zeros range min argsort info zeros range ones zeros str read join info all randn abs reshape shape sqrt sum len svd diag mean toeplitz zeros range array name run_experiments array makedirs mean shape zeros sum std range arange set_yticklabels clim str list ylabel colorbar strftime imshow title savefig gca range set_xticklabels reversed flipud int join xlabel reshape set_yticks text set_xticks figure len join remove str name exists join remove str name exists join style rc concat append DataFrame range print asarray process_results init_folder_structure squeeze save_recovery_details_3_methods strftime zeros range str asarray product init_folder_structure print get_reconstruction_error squeeze tqdm save_figure_and_data print_experiment_parameters save_recovery_details append matrix array range | This repository contains implementation for **MULAN** algorithm presented at NeurIPS 2018: <br /> MULAN: A Blind and Off-Grid Method for Multichannel Echo Retrieval <br /> https://nips.cc/Conferences/2018/Schedule?showEvent=11229 <br /> <br /> A <a href="https://arxiv.org/pdf/1810.13338.pdf">link</a> to the paper PDF on ArXiv. <br /> <br />  <br /> <br /> | 2,013 |
epfml/ChocoSGD | ['stochastic optimization'] | ['Decentralized Stochastic Optimization and Gossip Algorithms with Compressed Communication'] | dl_code/pcode/models/mlp.py dl_code/pcode/optim/parallel_choco_v.py convex_code/base_logistic.py convex_code/experiment_epsilon_final.py convex_code/logistic.py dl_code/auto_extract.py dl_code/run.py dl_code/pcode/create_dataset.py dl_code/pcode/models/data_parallel_wrapper.py dl_code/pcode/models/lenet.py dl_code/pcode/optim/utils.py dl_code/pcode/tools/show_results.py dl_code/pcode/optim/dcd_psgd.py dl_code/pcode/utils/tensor_buffer.py dl_code/pcode/utils/communication.py dl_code/pcode/utils/stat_tracker.py dl_code/pcode/utils/timer.py convex_code/experiment.py dl_code/pcode/datasets/loader/serialize.py dl_code/pcode/datasets/loader/utils.py dl_code/pcode/distributed_running_cv.py dl_code/pcode/create_scheduler.py convex_code/parameters.py dl_code/pcode/distributed_running_nlp.py dl_code/pcode/datasets/loader/svhn_folder.py dl_code/pcode/models/rnn_lm.py dl_code/pcode/create_optimizer.py dl_code/pcode/utils/topology.py dl_code/tmux_cluster/utils.py dl_code/pcode/tools/tune_consensus_stepsize.py dl_code/pcode/datasets/loader/imagenet_folder.py dl_code/pcode/datasets/loader/preprocess_toolkit.py dl_code/pcode/utils/auxiliary.py convex_code/constants.py dl_code/pcode/utils/checkpoint.py dl_code/pcode/datasets/partition_data.py dl_code/pcode/models/resnet.py dl_code/pcode/tools/plot.py dl_code/pcode/models/vgg.py dl_code/pcode/datasets/loader/epsilon_or_rcv1_folder.py dl_code/pcode/utils/logging.py dl_code/pcode/utils/op_files.py dl_code/pcode/models/__init__.py dl_code/pcode/optim/dgc.py dl_code/tmux_cluster/tmux.py dl_code/pcode/optim/deep_squeeze.py dl_code/pcode/utils/mathdict.py convex_code/utils.py dl_code/pcode/models/wideresnet.py convex_code/baselines.py dl_code/pcode/tools/plot_utils.py dl_code/pcode/create_model.py dl_code/pcode/models/densenet.py dl_code/pcode/optim/parallel_choco.py dl_code/parameters.py dl_code/pcode/optim/sgd.py dl_code/pcode/optim/ecd_psgd.py convex_code/pickle_datasets.py dl_code/pcode/utils/sparsification.py dl_code/pcode/create_metrics.py dl_code/main.py dl_code/pcode/utils/error_handler.py dl_code/pcode/utils/op_paths.py dl_code/pcode/optim/ef_sign_sgd.py dl_code/pcode/datasets/prepare_data.py loss BaseLogistic run_logistic run_experiment qsgd_quantize LogisticDecentralizedSGD Parameters to_percent unpickle_dir pickle_it main get_args check_args main init_distributed_world init_config get_args str2bool build_mpi_script main_nccl_or_gloo read_hostfile run_cmd map_slot main_mpi build_nccl_script create_job_on_nodes get_random_port _get_cv_data_stat _get_nlp_data_stat _define_cv_dataset load_data_batch define_cv_dataset define_nlp_dataset define_dataset Metrics consistent_model define_nlp_model define_model define_cv_model get_model_stat define_optimizer _linear_scale _get_lr_epoch_fields _get_scheduling_setup_for_onecycle _get_scheduling_setup_for_convex_decay _get_scheduling_setup_for_multistep _get_lr_scale_indicators DeterministicLRScheduler _get_lr_fields _get_scheduling_setup AdaptiveLRScheduler _build_multistep_lr_change_epochs _convex_scale _get_scheduling_setup_for_strict _build_multistep_lr_fields Scheduler _poly_scale inference validate do_validate train_and_validate inference validate do_validate train_and_validate Partition DataPartitioner _get_svhn _get_text get_dataset _get_nlp_lm_dataset _get_cifar _get_mnist _get_epsilon_or_rcv1 _get_stl10 _get_imagenet define_epsilon_or_rcv1_folder define_imagenet_folder Lighting Grayscale pad_random_crop Saturation Contrast scale_crop scale_random_crop get_transform Brightness inception_preproccess RandomOrder inception_color_preproccess ColorJitter loads_pyarrow loads_msgpack create_dummy_func dumps_msgpack dumps_pyarrow define_svhn_folder SVHN LMDBPT uint8_to_float be_ncwh_pt LMDBPTClass AllReduceDataParallel BasicLayer densenet Transition DenseNet Bottleneck lenet LeNet mlp MLP Bottleneck conv3x3 ResNet_imagenet ResNetBase resnet BasicBlock ResNet_cifar RNNLM vgg VGG wideresnet BasicBlock NetworkBlock WideResNet DCDSignCompressor DCDSparsificationCompressor DCD_PSGD DCDQuantizationCompressor DCDCompressor DeepSqueezeSignCompressor DeepSqueezeSparsificationCompressor DeepSqueezeQuantizationCompressor DeepSqueeze DeepSqueezeCompressor DGC compress_or_quantize ECDSignCompressor ECDQuantizationCompressor ECD_PSGD ECDCompressor ECDSparsificationCompressor EFSignCompressor EF_SignSGD CHOCOSparsificationCompressor CHOCOQuantizationCompressor CHOCOSignCompressor ParallelCHOCO CHOCOCompressor CHOCOSparsificationCompressor CHOCOQuantizationCompressor ParallelCHOCO_V CHOCOSignCompressor CHOCOCompressor SGD HelperThread update_params_from_neighbor recover_params apply_gradient join_thread plot_curve_wrt_time plot_by_global_minibatch_size build_legend smoothing_func groupby_indices rebuild_runtime_record drop_first_few configure_figure plot_one_case sample_from_records add_communication_delay find_same_num_sync determine_color_and_lines reject_outliers _parse_runtime_info reorganize_multi_records _parse_runtime_infos _summarize_info extract_list_of_records get_pickle_info reorganize_records reorder_records summarize_info is_meet_conditions _is_same load_raw_info_from_experiments _get_arguments _get_info_from_the_folder quantize evaluate_consensus_stepsize get_graph_topology search_gamma get_n_params get_diff_weights get_fullname str2time deepcopy_model is_float get_model_difference get_diff_states dict2obj list_to_vec save_arguments get_checkpoint_folder_name init_checkpoint save_to_checkpoint _save_to_checkpoint maybe_resume_from_checkpoint recover_device unflatten elementwise_min Aggregation global_average EfficientDecentralizedAggregation _get_data _get_shape get_data flatten get_aggregators DecentralizedAggregation CentralizedAggregation broadcast abort global_except_hook display_args dispaly_best_test_stat display_test_stat Logger display_training_stat _mathdict_binary_in_place_op _ifloordiv MathDict _mathdict_binary_op _itruediv _isub _iadd _imul _mathdict_map_op write_txt load_pickle output_string read_json read_txt write_cpickle load_cpickle write_pickle read_text_withoutsplit build_dir list_files remove_folder get_current_path build_dirs SparsificationCompressor SignCompressor QuantizationCompressor get_n_bits AverageMeter MaxMeter MinMeter BestPerf RuntimeTracker TensorBuffer Timer SocialNetworkGraph PhysicalLayout CompleteGraph MargulisExpanderGraph RingExtGraph UndirectedGraph Edge ExpanderGraph define_graph_topology RingGraph TorusGraph exec_on_node Task Run Job environ ossystem wait_for_file load_yaml squeeze sum logaddexp format print score LogisticDecentralizedSGD fit print format pickle_it makedirs minimum sum square sqrt floor abs len glob join parse_args add_argument ArgumentParser check_args join in_dir out_name in_dir load_raw_info_from_experiments out_path write_pickle join init_process_group timestamp build_dirs checkpoint train_and_validate_fn define_model save_arguments init_config Metrics DataParallel BestPerf init_distributed_world define_optimizer Timer maybe_resume_from_checkpoint Scheduler on_cuda CrossEntropyLoss set_best_tracker define_dataset checkpoint_dir display_args batch_size set_device define_graph_topology n_sub_process init_checkpoint Logger manual_seed on_cuda join sorted get_checkpoint_folder_name dict read_txt int list range items print system items list list items python_path format join list items print run_cmd make_job format print dict clean_python n_mpi_process build_nccl_script create_job_on_nodes append range dict clean_python create_job_on_nodes build_mpi_script on_cuda print define_cv_dataset define_nlp_dataset data join format splits print data_dir _get_nlp_data_stat get_dataset rnn_n_hidden rank rnn_use_pretrained_emb build_vocab num_iterations format batch_size print text n_nodes lr_warmup_epochs num_batches_train_per_device_per_epoch num_epochs num_batches_val_per_device_per_epoch len data format _get_cv_data_stat print _define_cv_dataset rank data format use batch_size print data_dir n_nodes get_dataset DataLoader rank DataPartitioner len num_iterations format batch_size print n_nodes lr_warmup_epochs num_batches_train_per_device_per_epoch num_epochs num_batches_val_per_device_per_epoch len consistent_model get_model_stat distributed on_cuda cuda vectors consistent_model format print get_model_stat copy_ arch RNNLM on_cuda cuda format print rank arch sum data format print all_reduce parameters rank _get_lr_epoch_fields lr_scale_indicators _get_lr_scale_indicators lr_change_epochs _get_lr_fields lr_fields format format join learning_rate lr_change_epochs init_warmup_lr lr_warmup lr_decay _build_multistep_lr_change_epochs lr_warmup_epochs _build_multistep_lr_fields num_epochs join split format learning_rate validate agg_model timer display_training_stat log do_validate save_json abort RuntimeTracker define_dataset deepcopy reshuffle_per_epoch collect print is_stop barrier reset summary train step evaluate update_metrics criterion model update is_best validate print dispaly_best_test_stat save_to_checkpoint epoch_ log_metric deepcopy agg_model _evaluate enumerate view n_tokens Compose CIFAR100 Normalize CIFAR10 Compose Compose join format join format load add_special_case Field splits _get_text join print format print get_transform format int join isinstance permute int replace DenseNet data int replace ResNet_imagenet ResNet_cifar data data int replace WideResNet compress get_n_bits data add_ zeros_like add get_data TensorBuffer buffer join build_legend smoothing_func list set_ylim log add set enumerate configure_figure plot_one_case reorder_records range determine_color_and_lines len update set_xscale list set_xticklabels extract_values set scatter set_xticks configure_figure determine_color_and_lines enumerate append smoothing range len sorted groupby list max len set_title set_xlabel set_ylabel legend set_tick_params lineplot plot split update vars dict list reduce values append _get_info_from_the_folder print join sorted format _parse_runtime_info join format append range len list isinstance Number keys enumerate values len _parse list defaultdict map _parse zip append print format len sorted split sort_values DataFrame reorder_records define_graph_topology int qsgd_quantize_numpy arange zeros_like choice range zeros_like quantize copy mean range normal format print get_graph_topology get_n_params logspace deepcopy track_model_aggregation clone parameters zip append parameters get_diff_weights zip ndarray isinstance view numel append tensor detach float str time format int comm_op quantize_level compress_ratio optimizer consensus_stepsize data join str checkpoint_dir checkpoint_root build_dirs timestamp rank arch checkpoint split join save write_pickle join _save_to_checkpoint copyfile load join str format checkpoint_root checkpoint_index checkpoint_dir collect print remove_folder rank resume update_from_checkpoint load_state_dict isfile Logger empty_cache helper all_reduce append _get_shape _get_data view nelement zip append empty view zip print_exception format write Get_rank flush Abort str format print graph node rank getattr device vars log_metric strftime log_metric save_json strftime log_metric format epoch_ strftime local_index best_perf get_best_perf_loc log print format print format write_txt dict split rmtree mkdir makedirs rmtree listdir graph_class ossystem isinstance print system tqdm time sleep | # Choco-SGD This repository provides code for **communication-efficient decentralized ML training** (both deep learning, compatible with [PyTorch](https://pytorch.org/), and traditional convex machine learning models. We provide code for the main experiments in the papers - [Decentralized Stochastic Optimization and Gossip Algorithms with Compressed Communication](https://arxiv.org/abs/1902.00340) and - [Decentralized Deep Learning with Arbitrary Communication Compression](https://arxiv.org/abs/1907.09356). Please refer to the folders `convex_code` and `dl_code` for more details. # References If you use the code, please cite the following papers: ``` @inproceedings{koloskova2019choco, | 2,014 |
epic-kitchens/action-models | ['action recognition', 'action classification'] | ['An Evaluation of Action Recognition Models on EPIC-Kitchens'] | ops/temporal_shift.py ops/trn.py tests/test_checkpoints.py demo.py ops/__init__.py tests/test_hub.py hubconf.py archs/bn_inception.py ops/basic_ops.py ops/utils.py archs/__init__.py ops/non_local.py model_loader.py tools/update_checkpoints.py tsn.py tools/print_checkpoint_details.py pretrained_settings.py tsm.py main extract_settings_from_args make_model make_trn get_model_settings_from_checkpoint load_checkpoint make_tsn make_tsm InvalidPretrainError TSM _vis_main TSN MTRN TRN bninception BNInception ConsensusModule get_base_net _NonLocalBlockND make_non_local NONLocalBlock2D NONLocalBlock3D NONLocalBlock1D NL3DWrapper TemporalShift TemporalPool make_temporal_shift make_temporal_pool InplaceShift return_TRN RelationModule RelationModuleMultiScaleWithClassifier __main RelationModuleMultiScale softmax class_accuracy log_add get_grad_hook find_checkpoints test_demo extract_model_details_from_filename test_can_load_all_checkpoints test_epic_pretrained_models test_imagenet_pretrained_models get_info main print_checkpoint_details main strip_module_prefix update_checkpoint get_model_variant vars basicConfig extract_settings_from_args make_model isinstance randn print load_checkpoint model print_model exit flow_length checkpoint endswith update load make_model load_state_dict get_model_settings_from_checkpoint format arch_diagram isinstance randn noun_count model print TSN base_model segment_count render make_dot verb_count modality cat format BNInception print load_url load_state_dict DataParallel isinstance Sequential NL3DWrapper layer1 make_block_temporal format isinstance layer3 layer4 info layer2 isinstance ResNet TemporalPool info layer2 lower RelationModuleMultiScale RelationModule model Variable print randn RelationModuleMultiScale exp astype confusion_matrix mean sum diag dict lower defaultdict split items extract_model_details_from_filename load_checkpoint getattr len main parse_args load load items ndarray Number isinstance append Tensor print items isinstance get_info load weights print_checkpoint_details img_feature_dim copy strip_module_prefix get_model_variant consensus_type save updated_checkpoint update_checkpoint | # EPIC-KITCHENS-55 action recognition models [](https://arxiv.org/abs/1908.00867) This is a set of models trained for EPIC-KITCHENS-55 baselines. We support: - [TSN](https://github.com/yjxiong/tsn-pytorch) - [TRN](https://github.com/metalbubble/TRN-pytorch) - [TSM](https://github.com/MIT-HAN-LAB/temporal-shift-module) Many thanks to the authors of these repositories. You can use the code provided here in one of two ways: 1. [PyTorch hub](#pytorch-hub) (**recommended**) 1. [Local installation](#local-installation) | 2,015 |
epic-kitchens/epic-kitchens-55-action-models | ['action recognition', 'action classification'] | ['An Evaluation of Action Recognition Models on EPIC-Kitchens'] | ops/temporal_shift.py ops/trn.py tests/test_checkpoints.py demo.py ops/__init__.py tests/test_hub.py hubconf.py archs/bn_inception.py ops/basic_ops.py ops/utils.py archs/__init__.py ops/non_local.py model_loader.py tools/update_checkpoints.py tsn.py tools/print_checkpoint_details.py pretrained_settings.py tsm.py main extract_settings_from_args make_model make_trn get_model_settings_from_checkpoint load_checkpoint make_tsn make_tsm InvalidPretrainError TSM _vis_main TSN MTRN TRN bninception BNInception ConsensusModule get_base_net _NonLocalBlockND make_non_local NONLocalBlock2D NONLocalBlock3D NONLocalBlock1D NL3DWrapper TemporalShift TemporalPool make_temporal_shift make_temporal_pool InplaceShift return_TRN RelationModule RelationModuleMultiScaleWithClassifier __main RelationModuleMultiScale softmax class_accuracy log_add get_grad_hook find_checkpoints test_demo extract_model_details_from_filename test_can_load_all_checkpoints test_epic_pretrained_models test_imagenet_pretrained_models get_info main print_checkpoint_details main strip_module_prefix update_checkpoint get_model_variant vars basicConfig extract_settings_from_args make_model isinstance randn print load_checkpoint model print_model exit flow_length checkpoint endswith update load make_model load_state_dict get_model_settings_from_checkpoint format arch_diagram isinstance randn noun_count model print TSN base_model segment_count render make_dot verb_count modality cat format BNInception print load_url load_state_dict DataParallel isinstance Sequential NL3DWrapper layer1 make_block_temporal format isinstance layer3 layer4 info layer2 isinstance ResNet TemporalPool info layer2 lower RelationModuleMultiScale RelationModule model Variable print randn RelationModuleMultiScale exp astype confusion_matrix mean sum diag dict lower defaultdict split items extract_model_details_from_filename load_checkpoint getattr len main parse_args load load items ndarray Number isinstance append Tensor print items isinstance get_info load weights print_checkpoint_details img_feature_dim copy strip_module_prefix get_model_variant consensus_type save updated_checkpoint update_checkpoint | # EPIC-KITCHENS-55 action recognition models [](https://arxiv.org/abs/1908.00867) This is a set of models trained for EPIC-KITCHENS-55 baselines. We support: - [TSN](https://github.com/yjxiong/tsn-pytorch) - [TRN](https://github.com/metalbubble/TRN-pytorch) - [TSM](https://github.com/MIT-HAN-LAB/temporal-shift-module) Many thanks to the authors of these repositories. You can use the code provided here in one of two ways: 1. [PyTorch hub](#pytorch-hub) (**recommended**) 1. [Local installation](#local-installation) | 2,016 |
epierson9/multiphenotype_methods | ['time series'] | ['Inferring Multidimensional Rates of Aging from Cross-Sectional Data'] | laplacian_variational_autoencoder.py multiphenotype_utils.py mortality_weighted_variational_age_autoencoder.py variational_autoencoder.py variational_longitudinal_monotonic_rate_of_aging_autoencoder.py sparse_variational_age_autoencoder.py setup.py general_autoencoder.py variational_rate_of_aging_autoencoder.py variational_rate_of_aging_monotonic_autoencoder.py sparse_correlation_variational_age_autoencoder.py variational_age_autoencoder.py dimreducer.py standard_autoencoder.py CPCA LinearAgePredictor NeuralNetAgePredictor LinearDimReducer PCA TibshiraniMixedCriterion MahalanobisDistance DimReducer GeneralAutoencoder VariationalLaplacianAutoencoder MortalityWeightedVariationalAgeAutoencoder make_age_bins remove_id_and_get_mat compute_correlation_matrix_with_incomplete_data get_continuous_features_as_matrix cluster_and_plot_correlation_matrix partition_dataframe_into_binary_and_continuous compute_column_means_with_incomplete_data divide_idxs_into_batches assert_zero_mean move_last_col_to_first add_id SparseCorrelationVariationalAgeAutoencoder SparseVariationalAgeAutoencoder StandardAutoencoder VariationalAgeAutoencoder VariationalAutoencoder VariationalLongitudinalMonotonicRateOfAgingAutoencoder VariationalRateOfAgingAutoencoder VariationalRateOfAgingMonotonicAutoencoder tolist deepcopy sum print float64 astype isnan cov corr DataFrame array range len print columns append values array set_cmap xticks abs max yticks show len colorbar imshow scatter range mean linkage deepcopy print argsort fcluster figure squareform partition_dataframe_into_binary_and_continuous print get_continuous_features_as_matrix mean list index move_last_col_to_first DataFrame values append arange print ceil range append len | This code contains methods for inferring low-dimensional representations of high-dimensional phenotype spaces. To see an example, please look at `toy_data_example.ipynb`. We are happy to make this code available for any commercial or non-commercial actor to use. This code is available under the MIT License (see LICENSE.md). | 2,017 |
epochx/elmo4irony | ['common sense reasoning'] | ['Deep contextualized word representations for detecting sarcasm and irony'] | eval_sarc_v2.py src/train.py prepare/prepare_riloff.py src/corpus/corpus.py run.py src/utils/ops.py src/layers/pooling.py src/seq_utils/data_manager.py base_args.py src/utils/logger.py src/layers/elmo.py evaluate.py download/twitter_web_crawler.py src/config.py prepare/prepare_iac_v2.py src/corpus/batch_iterator.py src/utils/io.py src/corpus/lang.py ensemble.py preprocess.py prepare/prepare_semeval_2018_irony.py prepare/prepare_sarc_v2.py src/seq_utils/pad.py prepare/prepare_iac_v1.py unicode_codes_py3.py src/utils/torch.py src/optim/optim.py prepare/split.py prepare/prepare_sarc_v2_pol.py augment_prepared.py twokenize.py src/models/classifier.py prepare/prepare_platek.py download/twitter_api_crawler.py extract_users_list extract_hashtags_list open_prepared_file is_english open_semeval_2018_file open_extra_file CustomArgumentParser CustomFormatter get_ids_and_ndarray_from_prob_file run_main get_ids_and_ndarray_from_prob_file preprocess open_prepared_file filter_hashtags main tokenizeRawTweetText simpleTokenize addAllnonempty splitToken squeezeWhitespace splitEdgePunct regex_or tokenize normalizeTextForTagger clean_tweet clean_tweet read_iac_v1_dataset read_iac_v2_dataset read_platek_dataset read_riloff_dataset read_csv_file read_csv_file split_list Trainer BaseNLPBatch BatchIterator IESTBatch POSCorpus BaseCorpus ClassificationCorpus Lang ElmoWordEncodingLayer SumPoolingLayer MaxPoolingLayer CombinedPoolingLayer PoolingLayer GatherLastLayer MeanPoolingLayer BLSTMEncoder SentenceEncodingLayer WordEncodingLayer Classifier ScheduledOptim OptimWithDecay split_map_sequences2d split_map_pad_sequences2d map_sequences split_map_sequences map_sequence split_map_pad_sequences split_map_sequence split_sequences2d Padder get_hyperparams_from_model load_pickle get_name_from_path get_hash_from_model write_hyperparams load_or_create _repr2str update_in_db write_output_details write_sent_reprs write_metrics read_jsonl write_probs write_output get_datetime_from_model save_pickle Logger get_commit_hash get_machine_id get_server_name columnwise_cosine_similarity np_softmax embed_context_window simple_columnwise_cosine_similarity batch_full_cosine_similarity mp_matching_op distance context_window full_cosine_similarity matrix_mp_matching_op to_var to_torch_embedding pack_forward get_gpu_memory_map get_free_gpu_index set guess_language detect string argmax get_ids_and_ndarray_from_prob_file format print DataFrame tolist output files mean stack append parse_args sum fillna append zip filter_hashtags np_softmax test_batches save_model batch_size write_architecture Trainer Logger torch_save_file cuda seed str basename update_learning_rate_nie corpus exit train_batches updt_lr_accuracy savetxt parse_args CrossEntropyLoss update_results state_dict manual_seed_all range SummaryWriter Classifier _update_in_db glob ClassificationCorpus test lstm_hidden_size dev_batches write_current_run_details manual_seed load join model_hash evaluate shuffle_examples write_hyperparams RESULTS_PATH print run_savepath write train_epoch tqdm parameters corpora_dict OptimWithDecay epochs label2id len sub addAllnonempty len splitEdgePunct append range finditer split append strip search replace unescape tokenize normalizeTextForTagger join listdir min append join Tweet zip int shuffle len get append append map_sequence seq_splitter map_sequence map_sequences pad1d split_map_sequences append append map_sequences split_sequences2d pad2d split_map_sequences2d print dirname makedirs DATABASE_CONNECTION_STRING join insert connect makedirs DATABASE_CONNECTION_STRING update connect join join join dirname get_hyperparams_from_model get_hyperparams_from_model replace genre map index2word id_tuples zip append write_output __getitem__ enumerate pop load_pickle print enable disable function_or_object callable save_pickle append loads strip sum exp amax embedding_dim view size contiguous transpose embeddings pow sum norm mul squeeze div sum norm mul squeeze div sum norm mul view size contiguous squeeze repeat norm mul size squeeze unsqueeze repeat mul transpose repeat columnwise_cosine_similarity mul view size transpose contiguous unsqueeze list requires_grad_ cuda pack_padded_sequence tolist argsort index_select numpy module check_output dict zip range len get_gpu_memory_map debug items Parameter Embedding Tensor | # elmo4irony: Deep contextualized word representations for detecting sarcasm and irony This repo contains the official implementation of the paper "Deep contextualized word representations for detecting sarcasm and irony" WASSA 2018. # Main Requirements * Python 3 * Pytorch 0.4.0 * conda # Installation 1. Clone this repo in your home directory ```bash git clone https://github.com/epochx/elmo4irony-dev | 2,018 |
eracah/self-supervised-survey | ['imitation learning'] | ['Supervise Thyself: Examining Self-Supervised Representations in Interactive Environments'] | evaluations/linear_model.py submit_scripts/predict_run.py models/vae.py submit_scripts/embed_run.py evaluations/pca_corr_model.py models/tdc.py models/base_encoder.py training/control_trainer.py data/setup.py models/setup.py submit_scripts/infer_test.py evaluations/utils.py models/rec_env_sim.py training/base_trainer.py training/inference_trainer.py utils.py evaluations/fmap_superimpose.py models/random_baselines.py data/env_utils/wrappers.py submit_scripts/infer_run.py models/inverse_model.py training/prediction_trainer.py submit_scripts/predict_test.py main.py training/embed_trainer.py evaluations/predict_model.py data/utils.py evaluations/infer_model.py models/base_linear_model.py data/env_utils/env_setup.py data/datasets.py data/env_utils/__init__.py data/env_utils/get_state_params.py models/shuffle_n_learn.py EnvDataset EpisodeDataset setup_dataset setup_train_data setup_data setup_test_data setup_viz_data convert_frame appendabledict convert_frames setup_env get_wrapper get_env_type get_gym_module flappybird_get_latent_dict snake_get_nclasses_table atari_get_nclasses_table waterworld_get_latent_dict catcher_get_latent_dict atari_get_latent_dict lunarlander_get_latent_dict bucket_coord flappybird_get_nclasses_table sonic_get_nclasses_table waterworld_get_nclasses_table lunarlander_get_nclasses_table snake_get_latent_dict catcher_get_nclasses_table sonic_get_latent_dict monster_kong_get_nclasses_table monster_kong_get_latent_dict FlappyBirdWrapper InfoWrapper AtariWrapper LunarLanderWrapper SonicWrapper InfoActionWrapper superimpose_seq_frames superimpose_fmaps InferModel LinearModel compute_pca_corr PredictModel OneStepForwardModel classification_acc get_encoder Encoder LinearModel InverseModel RandomWeightCNN RandomLinearProjection RawPixelsEncoder RecEnvSimulator get_weights_path setup_infer_model setup_viz_model setup_model load_weights setup_embed_model setup_predict_model ShuffleNLearn TDC Decoder VAE BaseTrainer base_evaluate ControlTrainer do_rollout EmbedTrainer InferenceTrainer PredictionTrainer setup_env EnvDataset env_name setup_dataset random_split print val_size tr_size manual_seed test_size setup_dataset dict getattr setup_dataset_fn mode fromarray asarray Compose extend to transforms partial seed make wrapper get_gym_module get_wrapper getattr get_env_type hasattr unwrapped floor dict bucket_coord num_buckets lookup_value dict bucket_coord position bucket_coord getRAM id dict shape dict num_buckets id dict dict dict shape creeps player dict partial dict height pos_y min bucket_coord index dict num_buckets dict num_buckets bucket_coord centerx log_figure list tuple reshape transpose choice shape imshow interpolate figure fmap_model numpy xs enumerate str transpose fmap_model axis imshow title numpy clf interpolate figure savefig __iter__ next data list print PCA dict shape zip fit_transform explained_variance_ratio_ data size float argmax ELU Sequential ReLU Conv2d task to setup_fn setup_env dict getattr device embedder_name n encoder load_weights encoder to load_weights device encoder to load_weights device get_weights_path load str print load_state_dict deepcopy task transfer_level transfer_env inf embed_env parent print list glob embed_level Path float iterdir rollouts mean device append to range do_rollout data make convert_frame render reset env_name step ctlr | # supervise-thyself Code for the ICML Self-Supervised Workshop paper: `Supervise Thyself: Examining Self-Supervised Representations in Interactive Environments` https://arxiv.org/abs/1906.11951 | 2,019 |
eracah/supervise-thyself | ['imitation learning'] | ['Supervise Thyself: Examining Self-Supervised Representations in Interactive Environments'] | evaluations/linear_model.py submit_scripts/predict_run.py models/vae.py submit_scripts/embed_run.py evaluations/pca_corr_model.py models/tdc.py models/base_encoder.py training/control_trainer.py data/setup.py models/setup.py submit_scripts/infer_test.py evaluations/utils.py models/rec_env_sim.py training/base_trainer.py training/inference_trainer.py utils.py evaluations/fmap_superimpose.py models/random_baselines.py data/env_utils/wrappers.py submit_scripts/infer_run.py models/inverse_model.py training/prediction_trainer.py submit_scripts/predict_test.py main.py training/embed_trainer.py evaluations/predict_model.py data/utils.py evaluations/infer_model.py models/base_linear_model.py data/env_utils/env_setup.py data/datasets.py data/env_utils/__init__.py data/env_utils/get_state_params.py models/shuffle_n_learn.py EnvDataset EpisodeDataset setup_dataset setup_train_data setup_data setup_test_data setup_viz_data convert_frame appendabledict convert_frames setup_env get_wrapper get_env_type get_gym_module flappybird_get_latent_dict snake_get_nclasses_table atari_get_nclasses_table waterworld_get_latent_dict catcher_get_latent_dict atari_get_latent_dict lunarlander_get_latent_dict bucket_coord flappybird_get_nclasses_table sonic_get_nclasses_table waterworld_get_nclasses_table lunarlander_get_nclasses_table snake_get_latent_dict catcher_get_nclasses_table sonic_get_latent_dict monster_kong_get_nclasses_table monster_kong_get_latent_dict FlappyBirdWrapper InfoWrapper AtariWrapper LunarLanderWrapper SonicWrapper InfoActionWrapper superimpose_seq_frames superimpose_fmaps InferModel LinearModel compute_pca_corr PredictModel OneStepForwardModel classification_acc get_encoder Encoder LinearModel InverseModel RandomWeightCNN RandomLinearProjection RawPixelsEncoder RecEnvSimulator get_weights_path setup_infer_model setup_viz_model setup_model load_weights setup_embed_model setup_predict_model ShuffleNLearn TDC Decoder VAE BaseTrainer base_evaluate ControlTrainer do_rollout EmbedTrainer InferenceTrainer PredictionTrainer setup_env EnvDataset env_name setup_dataset random_split print val_size tr_size manual_seed test_size setup_dataset dict getattr setup_dataset_fn mode fromarray asarray Compose extend to transforms partial seed make wrapper get_gym_module get_wrapper getattr get_env_type hasattr unwrapped floor dict bucket_coord num_buckets lookup_value dict bucket_coord position bucket_coord getRAM id dict shape dict num_buckets id dict dict dict shape creeps player dict partial dict height pos_y min bucket_coord index dict num_buckets dict num_buckets bucket_coord centerx log_figure list tuple reshape transpose choice shape imshow interpolate figure fmap_model numpy xs enumerate str transpose fmap_model axis imshow title numpy clf interpolate figure savefig __iter__ next data list print PCA dict shape zip fit_transform explained_variance_ratio_ data size float argmax ELU Sequential ReLU Conv2d task to setup_fn setup_env dict getattr device embedder_name n encoder load_weights encoder to load_weights device encoder to load_weights device get_weights_path load str print load_state_dict deepcopy task transfer_level transfer_env inf embed_env parent print list glob embed_level Path float iterdir rollouts mean device append to range do_rollout data make convert_frame render reset env_name step ctlr | # supervise-thyself Code for the ICML Self-Supervised Workshop paper: `Supervise Thyself: Examining Self-Supervised Representations in Interactive Environments` https://arxiv.org/abs/1906.11951 | 2,020 |
eragonruan/text-detection-ctpn | ['scene text detection'] | ['Detecting Text in Natural Image with Connectionist Text Proposal Network'] | utils/rpn_msr/anchor_target_layer.py utils/rpn_msr/config.py utils/rpn_msr/proposal_layer.py utils/text_connector/text_connect_cfg.py utils/text_connector/text_proposal_connector.py nets/vgg.py utils/bbox/setup.py utils/dataset/data_util.py utils/text_connector/text_proposal_graph_builder.py utils/dataset/data_provider.py utils/prepare/utils.py main/demo.py utils/text_connector/text_proposal_connector_oriented.py main/train.py utils/bbox/bbox_transform.py utils/text_connector/detectors.py utils/text_connector/other.py nets/model_train.py utils/rpn_msr/generate_anchors.py utils/prepare/split_label.py main resize_image get_images main anchor_target_layer model Bilstm lstm_fc smooth_l1_dist mean_image_subtraction make_var loss vgg_16 vgg_arg_scope clip_boxes bbox_transform bbox_transform_inv generator get_training_data load_annoataion get_batch GeneratorEnqueuer pickTopLeft shrink_poly orderConvex _unmap _compute_targets anchor_target_layer Config generate_anchors generate_basic_anchors scale_anchor _filter_boxes proposal_layer _filter_irregular_boxes TextDetector clip_boxes threshold Graph Config TextProposalConnector TextProposalConnector TextProposalGraphBuilder join test_data_path format endswith print append walk len int min shape resize float max exists output_path rmtree gpu makedirs trainable_variables checkpoint_path pretrained_model_path moving_average_decay Saver get_variable global_variables merge_all strftime placeholder apply apply_gradients get_default_graph FileWriter logs_path get_trainable_variables assign_from_checkpoint_fn ConfigProto int learning_rate Variable now float32 AdamOptimizer ExponentialMovingAverage global_variables_initializer scalar range split Bilstm reshape lstm_fc conv2d shape mean_image_subtraction softmax anchor_target_layer REGULARIZATION_LOSSES not_equal reshape get_collection float32 where sparse_softmax_cross_entropy_with_logits reduce_sum shape smooth_l1_dist reduce_mean cast add_n gather equal scalar transpose log dtype exp astype shape zeros minimum maximum join format endswith print append walk len append list map split load_annoataion arange subplots show shape imshow get_training_data imread format close shuffle tight_layout splitext join print reshape set_yticks rectangle set_xticks array split generator get is_running start sleep GeneratorEnqueuer argsort reshape pickTopLeft convex_hull int min append max range arange RPN_BBOX_INSIDE_WEIGHTS _unmap argmax RPN_FG_FRACTION generate_anchors ones transpose array meshgrid sum RPN_BATCHSIZE format hstack ascontiguousarray choice sqrt fill empty RPN_POSITIVE_WEIGHT int EPS print reshape RPN_CLOBBER_POSITIVES _compute_targets zeros bbox_overlaps fill empty zeros array int32 scale_anchor copy append nms generate_anchors RPN_POST_NMS_TOP_N format arange print reshape meshgrid transpose clip_boxes bbox_transform_inv _filter_boxes hstack shape RPN_NMS_THRESH RPN_PRE_NMS_TOP_N RPN_MIN_SIZE threshold | # text-detection-ctpn text detection mainly based on ctpn (connectionist text proposal network). It is implemented in tensorflow. I use id card detect as an example to demonstrate the results, but it should be noticing that this model can be used in almost every horizontal scene text detection task. The origin paper can be found [here](https://arxiv.org/abs/1609.03605). Also, the origin repo in caffe can be found in [here](https://github.com/tianzhi0549/CTPN). For more detail about the paper and code, see this [blog](http://slade-ruan.me/2017/10/22/text-detection-ctpn/). If you got any questions, check the issue first, if the problem persists, open a new issue. *** # roadmap - [x] freeze the graph for convenient inference - [x] pure python, cython nms and cuda nms - [x] loss function as referred in paper - [x] oriented text connector - [x] BLSTM *** | 2,021 |
eraserNut/MTMT | ['shadow detection'] | ['A Multi-Task Mean Teacher for Semi-Supervised Shadow Detection'] | train.py networks/MTMT.py utils/losses.py utils/ramps.py dataloaders/joint_transforms.py test_MT_util.py networks/resnext/resnext101_5out.py test_MT.py networks/resnext/resnext101_regular.py config.py dataloaders/joint_transforms_edge.py networks/resnext/config.py networks/resnext/resnext_101_32x4d_.py utils/util.py networks/resnext/__init__.py dataloaders/SBU.py test_calculate_metric test_all_case cal_acc worker_init_fn get_current_consistency_weight update_ema_variables create_model Resize RandomHorizontallyFlip Compose Resize RandomHorizontallyFlip Compose make_union_dataset relabel_dataset SBU make_labeled_dataset MergeLayer2 MergeLayer1 build_model MergeLayer1_FPN extra_layer ResidualBlockLayer TUN_bone weights_init xavier ConvertLayer ResNeXt101 ResNeXt101 Lambda LambdaMap LambdaBase LambdaReduce get_resnext_101_32x4d binary_xloss flatten_binary_scores_weight softmax_kl_loss softmax_dice_loss lovasz_grad dice_loss softmax_mse_loss lovasz_hinge_flat StableBCELoss lovasz_hinge sigmoid_mse_loss mean lovasz_hinge_flat_weight entropy_loss_map bce2d_new symmetric_mse_loss flatten_binary_scores isnan kl_loss lovasz_hinge_weight mse_loss entropy_loss sigmoid_rampup linear_rampup cosine_rampdown crf_refine _sigmoid load_model AverageMeter iterate_once UnifLabelSampler relabel_dataset TwoStreamBatchSampler grouper Logger learning_rate_decay iterate_eternally cal_subitizing load format print eval load_state_dict test_all_case cuda ToPILImage save resize cal_acc cuda crf_refine format size Compose Normalize to_pil net join print convert tqdm sigmoid cpu array float sum astype data min add_ parameters zip parameters detach_ cuda build_model seed append append sorted set imgs append range len MergeLayer2 MergeLayer1 xavier_uniform_ normal_ zero_ isinstance Conv2d AvgPool2d LambdaMap Sequential Lambda MaxPool2d Conv2d ReLU BatchNorm2d LambdaReduce Linear float sum mean sum cuda log softmax range sum cuda log softmax sigmoid log_softmax softmax float sum mean lovasz_hinge_flat_weight data lovasz_grad relu Variable sort dot float view mean lovasz_hinge_flat data lovasz_grad relu Variable sort dot float view Variable float flatten_binary_scores cumsum sum len filterfalse next iter enumerate clip load format print size load_state_dict isfile param_groups sqrt addPairwiseGaussian _sigmoid DenseCRF2D reshape addPairwiseBilateral astype flatten zeros setUnaryEnergy max convert min label remove_small_objects sum array | # A Multi-task Mean Teacher for Semi-supervised Shadow Detection by Zhihao Chen, Lei Zhu, Liang Wan, Song Wang, Wei Feng, and Pheng-Ann Heng [[paper link](http://openaccess.thecvf.com/content_CVPR_2020/papers/Chen_A_Multi-Task_Mean_Teacher_for_Semi-Supervised_Shadow_Detection_CVPR_2020_paper.pdf)] #### News: In 2020.9.17, We release the unsorted code for other researchers. The sorted code will be released after. *** ## Citation @inproceedings{chen20MTMT, author = {Chen, Zhihao and Zhu, Lei and Wan, Liang and Wang, Song and Feng, Wei and Heng, Pheng-Ann}, title = {A Multi-task Mean Teacher for Semi-supervised Shadow Detection}, booktitle = {CVPR}, year = {2020} | 2,022 |
eric-erki/ID-Card-Segmentation | ['face detection'] | ['MIDV-500: A Dataset for Identity Documents Analysis and Recognition on Mobile Devices in Video Stream'] | test_model.py model/iou_loss.py model/train.py model/unet_model.py dataset/download_dataset.py dataset/stack_npy.py main read_image main IoU v_generator t_generator get_model load COLOR_BGR2GRAY fillPoly reshape shape resize zeros imread array cvtColor open str remove sorted replace print glob system rmtree save zip append expand_dims listdir array read_image load vstack range sum zeros range zeros range clear_session concatenate print Model summary Input compile | # ID-Card-Segmentation Segmentation of ID Cards using U-Net ### U-Net Architecture <img src="http://deeplearning.net/tutorial/_images/unet.jpg" width="500" height="400" alt="U-net"> ### Our Result's   ### Requirements - Tensorflow-GPU 1.12 - Keras 2.1 | 2,023 |
eric4421/MTWI-2018 | ['scene text detection'] | ['Detecting Text in Natural Image with Connectionist Text Proposal Network'] | handler.py train.py lib/utils.py net/loss.py predict.py config.py net/network.py evaluate.py lib/setup_cython.py model.py evaluate write_data prehandler scale_image list2str LMDB_dataset read_txt check_image create_dataset reorganize_dataset scale_image_only gen_VGG16 threshold get_text_recs fit_y predict_one gen_test_pair get_anchor_h random_test clip_box meet_v_iou succession plot_loss loop_files get_train_val valid_anchor tag_anchor draw_box_4pt draw_poly_4pt get_y anchor_y np2base64 draw_box_2pt base642img init_weight get_range draw_slice get_iou get_anchor_in_image y_range trans_to_2pt gen_anchor CTPN_Loss VGG_16 CTPN Im2col BLSTM scale_image read_txt cuda FloatTensor transpose imread format tag_anchor splitext info sample float net join time deepcopy criterion print split gen_anchor join readline format print close open listdir range split readlines close open append split VGG_16 write_data clear format str np_img2base64 Environment print len scale_image list2str read_txt imread range makedirs int min resize append float range len float min resize append join str remove print sort rename zip listdir array open load update load_state_dict save append range vgg16 state_dict threshold polyfit poly1d max fabs fit_y min float32 polyfit mean sqrt clip_box zeros array enumerate join splitext sample listdir append pow list len set get_anchor_h append range enumerate imwrite succession scale_image_only draw_ploy_4pt transpose shape append imread range trans_to_2pt get_text_recs draw_box_2pt net deepcopy print cpu_nms pow Tensor array split imwrite gen_test_pair succession scale_image_only exists draw_ploy_4pt basename transpose shape append imread range trans_to_2pt get_text_recs draw_box_2pt mkdir net deepcopy join print cpu_nms rmtree pow Tensor array append join listdir join loop_files subplot list plot xlabel ylabel title savefig range len line reshape array int32 rectangle int float b64encode squeeze b64decode fromstring COLOR_RGB2BGR imdecode len float int int float bias constant_ normal_ weight range all_weights len int draw_box_4pt min anchor_y floor draw_slice ceil float max range append len deepcopy draw_box_4pt append range len readlines close open append gen_anchor array split count_nonzero max size min zeros y_range range int float y_range min where log10 get_iou append zeros float max range | # README The Python implementation of Connectionist Text Proposal Network (CTPN) targeting at MTWI 2018 Challenge 2. ## Competition Details https://tianchi.aliyun.com/competition/entrance/231685/introduction ## Environment * Python 3.x with PyTorch (0.4 or newer) * cython, configphaser, lmdb, matplotlib, numpy, opencv-python are required. ## Run before training * Run setup_cython.py FIRST: ```bash | 2,024 |
ericholgate/vulgartwitter | ['sentiment analysis'] | ['Expressively vulgar: The socio-dynamics of vulgarity and its effects on sentiment analysis in social media'] | clean_data.py bilstm.py BILSTM read_csv_curse_words check_regex read_regex find punctuation strip close translate maketrans open append split len append close split bool fullmatch | # Expressively vulgar: The socio-dynamics of vulgarity and its effects on sentiment analysis in social media **[Please check this paper for details regarding annotation and modeling](http://aclweb.org/anthology/C18-1248)** **Citation:** ``` @InProceedings{cachola2018vulgar, author = {Cachola, Isabel and Holgate, Eric and Preo\c{t}iuc-Pietro, Daniel and Li, Junyi Jessy}, title = {Expressively vulgar: The socio-dynamics of vulgarity and its effects on sentiment analysis in social media}, booktitle = {Proceedings of the 27th International Conference on Computational Linguistics}, year = {2018}, pages = {2927--2938}, | 2,025 |
erichson/koopmanAE | ['time series'] | ['Forecasting Sequential Data using Consistent Koopman Autoencoders'] | train.py plot_pred_error.py tools.py model.py driver.py read_dataset.py decoderNet dynamics_back koopmanAE gaussian_init_ encoderNet dynamics moving_average data_from_name rescale pendulum_lin pendulum set_seed get_device weights_init add_channels train Tensor Normal cumsum min max seed T arange ptp standard_normal min copy dot sol qr seed T arange ptp standard_normal min copy dot sol qr seed manual_seed print device isinstance bias xavier_uniform_ weight constant_ Linear model clip_grad_norm_ zero_grad weight numpy abs hasattr step append to sum range dynamics eig item enumerate lr_scheduler criterion backward AdamW print parameters get_device mm | # Consistent Koopman Autoencoders Research code that demonstrates consistent Koopman autoencoders on the nonlinear pendulum with no friction. Here is an example: ``` python driver.py --dataset pendulum --folder results_back_pendulum --bottleneck 6 --backward 1 ``` You can also create a baseline model: ``` python driver.py --dataset pendulum --folder results_pendulum --bottleneck 6 --backward 0 ``` | 2,026 |
ermongroup/BiasAndGeneralization | ['density estimation'] | ['Bias and Generalization in Deep Generative Models: An Empirical Study'] | clevr/pixelcnn/pixel_cnn_pp/model.py DotsAndPie/dataset/__init__.py clevr/pixelcnn/data/clevr_data.py clevr/pixelcnn/train.py clevr/pixelcnn/pixel_cnn_pp/nn.py clevr/clevr/image_generation/render_images.py clevr/clevr/image_generation/utils.py DotsAndPie/dataset/dataset_dots.py DotsAndPie/train.py clevr/gan/utils.py clevr/pixelcnn/utils/plotting.py Evaluate/combination_dataset.py DotsAndPie/gan.py clevr/gan/ops.py clevr/gan/sc_model.py DotsAndPie/models.py DotsAndPie/dataset/dataset_pie.py clevr/clevr/image_generation/generate_combinations.py DotsAndPie/vae.py DotsAndPie/utils.py DotsAndPie/dataset/generate/dots_generator.py clevr/gan/num.py clevr/gan/sc.py clevr/gan/num_model.py render_shadeless render_scene compute_all_relationships add_random_objects check_visibility main render_shadeless render_scene compute_all_relationships add_random_objects check_visibility main get_camera_coords set_layer add_material extract_args delete_object parse_args load_materials add_object main parse_time lrelu batch_norm linear concat conv2d deconv2d conv_cond_concat main parse_time get_image make_gif to_json visualize save_images transform image_manifold_size center_crop merge_images show_all_variables inverse_transform imread imsave merge sample_from_model parse_time make_feed_dict DataLoader model_spec int_shape energy_distance get_var_maybe_avg down_shifted_conv2d log_prob_from_logits conv2d gated_resnet discretized_mix_logistic_loss adam_updates get_name right_shift down_shifted_deconv2d down_shift dense down_right_shifted_conv2d concat_elu log_sum_exp nin deconv2d sample_from_discretized_mix_logistic down_right_shifted_deconv2d get_vars_maybe_avg plot_img tile_raster_images scale_to_unit_interval conv_filter_tile img_stretch img_tile GAN generator_conv64large discriminator_fc64 encoder_conv64large generator_conv64 encoder_fc64 generator_conv64small discriminator_conv28 discriminator_conv64large generator_fc64 discriminator_conv64 encoder_conv64small discriminator_conv64small sample_z generator_conv28 encoder_conv64 encoder_conv28 lrelu create_multi_display conv2d_t_relu convert_to_display create_display make_model_path fc_bn_lrelu fc_lrelu conv2d_lrelu conv2d_bn_lrelu fc_bn_relu fc_relu conv2d_t_bn_relu VAE DotsDataset compute_proportion gen_pie compute_radius PieDataset compute_location gen_image_count fig2data CombinationDataset join time min_objects render_scene output_blend_dir max_objects num_images gen_list start_idx split append randint range makedirs tuple object open_mainfile to_quaternion Vector save_as_mainfile render normalized delete_object width add_random_objects range normal height render_max_bounces primitive_plane_add render_min_bounces render_tile_size preferences compute_all_relationships material_dir load_materials render_num_samples shape_dir object abs list add_material location uniform delete_object append range choice sqrt items print check_visibility get_camera_coords add_object min_pixels_per_object split items sorted list add set append sum enumerate load render_shadeless pixels list remove Counter mkstemp most_common set_layer engine new set add render zip append filepath use_antialiasing enumerate output_scene_dir print output_image_dir argv index delete objects world_to_camera_view int resolution_y scene resolution_x round resolution_percentage layers range len join resize translate objects startswith append append join listdir active_object materials new inputs nodes append len checkpoint_dir input_height sample_dir output_height pprint open __flags ConfigProto test_sample_dir get_shape as_list trainable_variables analyze_vars imread zeros enumerate squeeze merge int round center_crop imresize VideoClip write_gif make_gif int arange save_images batch_size print sampler strftime choice uniform gmtime run tile ceil zeros range append enumerate int nr_gpu range run update nr_gpu split get_shape len get_shape reduce_max len get_shape reduce_max len range len int tanh exp softplus reshape concat maximum where reduce_sum sigmoid log_prob_from_logits int_shape zeros log minimum tanh get_shape exp one_hot reshape maximum reduce_sum random_uniform int_shape argmax log average get_variable append get_var_maybe_avg get_shape gradients Variable name square pow assign sqrt assign_add zip append zeros str get_name get_name int_shape get_name int_shape reshape dense dropout nonlinearity sigmoid conv int_shape initialized_value split int_shape int_shape pad int_shape deconv2d pad int_shape deconv2d axis tight_layout imshow title figure astype int sqrt ceil img_stretch float empty array range shape transpose img_stretch resize copy isinstance reshape scale_to_unit_interval zeros range call makedirs isdir split lrelu convolution2d lrelu convolution2d batch_norm convolution2d_transpose batch_norm relu convolution2d_transpose relu lrelu fully_connected fully_connected relu fully_connected batch_norm relu lrelu fully_connected batch_norm reshape transpose int slice reshape transpose sqrt pad floor float int slice reshape transpose stack pad floor float range len int concatenate cumsum zeros insert ones pi atan2 copy uniform append randint range len reshape pi sqrt argwhere float len reshape argwhere mean argwhere get_width_height tostring_argb fromstring draw roll fill_diagonal Circle fig2data axis square tight_layout add_artist close sqrt uniform figure tile gca expand_dims sum range | ermongroup/BiasAndGeneralization | 2,027 |
ermongroup/GraphScoreMatching | ['graph generation'] | ['Permutation Invariant Graph Generation via Score-Based Generative Modeling'] | model/gin.py model/edp_gnn.py model/gnn.py process_dataset.py evaluation/stats.py train.py utils/arg_helper.py model/gcn_layer.py model/pix2pix.py model/cond_layers.py model/score_network.py utils/loading_utils.py sample.py evaluation/setup.py utils/visual_utils.py model/langevin_mc.py model/mlp.py utils/data_generators.py utils/graph_utils.py gen_data.py model/gcn.py evaluation/mmd.py graph_load_batch graph_load parse_index_file citeseer_ego save_dataset sample_main train_main loss_func fit emd kernel_parallel_worker l2 gaussian_tv gaussian compute_mmd disc gaussian_emd test process_tensor kernel_parallel_unpacked compute_emd eval_torch_batch spectral_stats clustering_worker degree_worker adjs_to_graphs is_lobster_graph degree_stats edge_list_reindexed spectral_worker clustering_stats orca orbit_stats_all eval_acc_lobster_graph add_tensor eval_graph_list ConditionalLayer1d EdgeDensePredictionGNNLayer EdgeDensePredictionGraphScoreNetwork GCN GraphConvolution GraphAttentionLayer doubly_stochastic_norm GIN GraphNeuralNetwork LangevinMCSampler func_x MLP UnetSkipConnectionBlockWithResNet get_norm_layer PixelDiscriminator GANLoss ResnetGenerator ResnetBlock define_D UnetGenerator UnetSkipConnectionBlock init_weights get_scheduler init_net NLayerDiscriminator cal_gradient_penalty define_G ScoreNetwork ConvScore UNetResScore MLPScoreNetwork node_feature_to_matrix OneLayerScoreNetwork edict2dict load_model set_seed_and_logger graphs_to_dataloader get_config graphs_to_tensor parse_arguments mkdir load_data process_config n_community load_dataset GraphGenerator gen_graph_list remove_self_loop_if_exists check_adjs_symmetry add_self_loop_if_not_exists gen_list_of_data pad_adjs add_gaussian_noise toggle_edge_np get_corrupt_k mask_adjs get_norm_fit_stats get_score_model gen_random_feature get_mc_sampler prepare_test_model eval_sample_batch plot_graphs_list_new plot_multi_channel_numpy_adjs_1b1 plot_graphs_list plot_curve plot_multi_channel_numpy_adjs plot_graphs_adj save_fig str list add_edges_from number_of_nodes max isolates arange print loadtxt Graph subgraph astype map number_of_edges append remove_nodes_from range add_node append int strip open load list format lil_matrix from_dict_of_lists tolil tuple sort len min parse_index_file append max range open connected_component_subgraphs graph_load selfloop_edges number_of_nodes convert_node_labels_to_integers remove_edges_from append max range ego_graph print join plot_graphs_list mkdir run_sample batch_size save_dir sorted load_model eps set_seed_and_logger max_node_num copy prepare_test_model mkdir info join isinstance print grad_step_size get_mc_sampler load_data edict makedirs mean zip append item __next__ model zero_grad loss_func dev save append to range gen_list_of_data chunk mean eval item info model_save_dir join time ExponentialLR isinstance backward grad_step_size array2string train step len str requires_grad join get_score_model set_seed_and_logger size fit Adam named_parameters parameters sample_main get_mc_sampler load_data info append sum max len process_tensor astype float max len norm emd process_tensor float astype norm process_tensor float sum astype print array hstack max len process_tensor print compute_mmd now append range array degree_histogram len sum todense histogram eigvalsh print compute_mmd now spectral_worker append range len list histogram values list print compute_mmd now histogram append range values len dict nodes append edges join str remove number_of_nodes check_output strip write close len edge_list_reindexed number_of_edges find array open number_of_nodes print compute_mmd orca append sum array len list selfloop_edges isolates remove_edges_from from_numpy_matrix remove_nodes_from append add_node is_lobster_graph is_tree remove_nodes_from nodes len adjs_to_graphs numpy eval_graph_list print bmm transpose sum BatchNorm2d partial InstanceNorm2d LambdaLR CosineAnnealingLR ReduceLROnPlateau StepLR print apply init_weights to DataParallel ResnetGenerator UnetGenerator get_norm_layer NLayerDiscriminator PixelDiscriminator get_norm_layer view size rand grad mean requires_grad_ netD to size transpose cat expand parse_args add_argument ArgumentParser load print config_file process_config edict open model_save_dir str join dump edict2dict folder_name print exp_dir getpid mkdir pformat save_dir exp_name open items list isinstance makedirs seed join manual_seed_all basicConfig RandomState info log_level comment lower manual_seed save_dir setLevel FileHandler DataLoader TensorDataset graphs_to_tensor data list asarray isinstance concatenate pad_adjs append tensor to_numpy_matrix len int split load_dataset get_score_model eval load_state_dict to edict connected_component_subgraphs add_edge list print len nodes number_of_edges range disjoint_union_all str locals join number_of_nodes print graph_generator dumps number_of_edges info append max GraphGenerator update join map set info values unsqueeze min geometric min choice transpose add_gaussian_noise repeat append cat len check_adjs_symmetry transpose mask_adjs triu abs concatenate to repeat randn cdf norm kstest fit LangevinMCSampler update str info feature_nums dict dev in_feature to ones_like zeros_like where info plot_graphs_adj model_save_dir load join print sort append listdir model_files spring_layout subplot list isolates save_fig min draw axis copy figure remove_nodes_from range len set_text axis max subplot list draw_networkx_nodes edges append draw_networkx_edges range asarray zip remove_nodes_from enumerate spring_layout int isolates suptitle print from_numpy_array figure save_fig len rstrip number_of_nodes twilight_shifted axis vstack linspace max list Reds draw_a_graph range Blues_r asarray mean from_list remove_nodes_from enumerate spring_layout int isolates print min from_numpy_array figure pcolor std save_fig len show join close tight_layout subplots_adjust savefig makedirs asarray plot axvline add_subplot axhline figure fill_between DataFrame number_of_nodes set_text subplot list range copy sqrt number_of_edges remove_nodes_from spring_layout int isolates suptitle min draw number_of_selfloops figure save_fig len selfloop_edges number_of_nodes set_text subplot list from_numpy_matrix ceil range sqrt number_of_edges remove_edges_from remove_nodes_from isinstance isolates suptitle min draw number_of_selfloops figure Tensor numpy save_fig | # Permutation Invariant Graph Generation via Score-Based Generative Modeling This repo contains the official implementation for the paper [Permutation Invariant Graph Generation via Score-Based Generative Modeling](http://proceedings.mlr.press/v108/niu20a) (AISTATS 2020), Authors: Chenhao Niu, Yang Song, Jiaming Song, Shengjia Zhao, Aditya Grover, Stefano Ermon ------------------------------------------------------------------------------------- We propose a permutation invariant approach to modeling graphs, using the framework of score-based generative modeling. In particular, we design a permutation equivariant, multi-channel graph neural network to model the gradient of the data distribution at the input graph (_a.k.a_, the score function). This permutation equivariant model of gradients implicitly defines a permutation invariant distribution for graphs. We can train this graph neural network with score matching and sample from it with annealed Langevin dynamics. ## Dependencies First, install PyTorch following the steps on its [official website](https://pytorch.org/). The code has been tested over PyTorch 1.3.1 and 1.8.1. Then run the following command to install the other dependencies. ```shell | 2,028 |
ermongroup/InfoGAIL | ['imitation learning'] | ['InfoGAIL: Interpretable Imitation Learning from Visual Demonstrations'] | wgail_info_0/utils.py wgail_info_0/wgail_info.py wgail_info_0/gym_torcs.py wgail_info_1/wgail_info.py wgail_info_1/utils.py wgail_info_0/drive.py wgail_info_1/snakeoil_gym.py wgail_info_1/gym_torcs.py wgail_info_1/models.py wgail_info_1/drive.py wgail_info_1/preprocess.py wgail_info_0/models.py wgail_info_1/behavior_clone.py wgail_info_0/snakeoil_gym.py wgail_info_0/behavior_clone.py wgail_info_0/preprocess.py get_feat main calc_loss main get_state clip TorcsEnv Generator Posterior TRPOAgent main normalize collect_demo ServerState destringify Client drive_example clip DriverAction gauss_sample GetFlat gauss_ent get_feat conjugate_gradient get_state ReplayBuffer linesearch numel SetFromFlat discount dict2 gauss_selfKL_firstfixed TimeDependentBaseline pathlength NNBaseline rollout_contin LinearBaseline gauss_log_prob gauss_prob_val explained_variance gauss_prob gauss_KL var_shape flatgrad playGame get_feat main calc_loss main get_state clip TorcsEnv Generator Posterior TRPOAgent main normalize collect_demo ServerState destringify Client drive_example clip DriverAction gauss_sample GetFlat gauss_ent get_feat conjugate_gradient get_state ReplayBuffer linesearch numel SetFromFlat discount dict2 gauss_selfKL_firstfixed TimeDependentBaseline pathlength NNBaseline rollout_contin LinearBaseline gauss_log_prob gauss_prob_val explained_variance gauss_prob gauss_KL var_shape flatgrad playGame average abs preprocess_input astype float32 predict arange save_weights get_feat Session seed set_session Generator Model ResNet50 sum range predict shuffle ConfigProto zeros calc_loss load print summary randint train array speedZ preprocess_input speedX img astype float32 damage pre_action_1 pre_action_0 resize zeros speedY predict get_state end load_weights reset TorcsEnv step dict2 str img_to_array uint8 load_img concatenate print readlines astype float32 resize zeros range len normalize savez_compressed collect_demo d clip exp square pi exp square pi constant exp log pi stop_gradient list map exp square reduce_sum constant e pi reduce_sum float32 log gradients get_feat expand_dims array get_state dict2 act print end reset append zeros step array range var f arange enumerate zeros_like copy dot f_Ax range seed load set_session learn TRPOAgent print load_weights TorcsEnv ConfigProto Session | # [InfoGAIL: Interpretable Imitation Learning from Visual Demonstrations](https://arxiv.org/abs/1703.08840) By Yunzhu Li, Jiaming Song, Stefano Ermon ### Introduction Modified codebase of TORCS, with the ability to extract dashboard views. InfoGAIL implementation, attached with two examples: *pass* & *turn*. ### Citing InfoGAIL If you find this codebase useful in your research, please consider citing: @article{li2017inferring, title={InfoGAIL: Interpretable Imitation Learning from Visual Demonstrations}, author={Li, Yunzhu and Song, Jiaming and Ermon, Stefano}, | 2,029 |
ermongroup/dail | ['imitation learning'] | ['Domain Adaptive Imitation Learning'] | reacher_env/mujoco/reacher_3dof_corner.py reacher_env/mujoco/reacher_2dof.py environments/transfer_env/reacher_env.py reacher_env/mujoco/tp_reacher_2dof.py reacher_env/mujoco/write_reacher_2dof.py reacher_env/mujoco/reacher_2dof_corner.py reacher_env/mujoco/reacher_6dof.py reacher_env/mujoco/reacher_2dof_act.py reacher_env/mujoco/reacher_5dof.py compgraph.py environments/reacher_generator.py train.py reacher_env/mujoco/reacher_3dof.py reacher_env/mujoco/tp_write_reacher_2dof.py utils.py environments/asset_generator.py reacher_env/mujoco/reacher_4dof.py sample.py environment.py environments/snake_generator.py reacher_env/mujoco/reacher_2dof_act_corner.py reacher_env/mujoco/reacher_3dof_push.py agents/base.py environments/transfer_env/snake_env.py reacher_env/mujoco/reacher_2dof_act_wall.py environments/centipede_generator.py graphs/ddpg/ddpg_graph_with_goal.py environments/multitask_env/walkers.py agents/ddpg.py model.py environments/transfer_env/invpendulum_env.py environments/transfer_env/antS.py reacher_env/mujoco/reacher_2dof_sto.py reacher_env/__init__.py reacher_env/mujoco/reacher_2dof_wall.py replaymemory.py environments/register.py environments/transfer_env/centipede_env.py reacher_env/mujoco/reacher_2dof_act_push.py reacher_env/mujoco/reacher_3dof_wall.py saved_params/reacher_params.py __init__.py graphs/ddpg/transfer_ddpg_graph_no_goal.py graphs/ddpg/transfer_ddpg_graph_with_goal.py reacher_env/mujoco/reacher_2dof_push.py build_compgraph get_ddpg_ph PermutedPendulumEnv TruncatedPendulumEnv ModifiedPendulumEnv InvertedPendulumEnv CustomMountainCarEnv CustomPendulumEnv MountainCarEnv create_env CustomCartPoleEnv logit_bernoulli_entropy feedforward convnet scale_state logsigmoid scale_action get_regularizer get_activation get_initializer FIFOdeque create_replay_memory eps_greedy_sample gaussian_sample open_file_and_save save_frames_as_video print_metrics create_hybrid_dataset create_dataset render_policy GracefulKiller DDPGAgent save_xml_files _add_body generate_centipede _add_leg _add_custom _add_actuators generate_reacher _add_actuators _add_body get_name_list get_mujoco_model_settings _add_actuators generate_snake _add_body WalkersHopperEnv WalkersHalfcheetahtwoEnv WalkersFullcheetahoneEnv WalkersHoppertwoEnv WalkersHalfhumanoidthreeEnv WalkersHalfcheetahzeroEnv WalkersHalfhumanoidfiveEnv WalkersFullcheetahfourEnv WalkersKangarooEnv WalkersOstrichfourEnv WalkersOstrichzeroEnv WalkersHalfcheetahoneEnv WalkersHalfhumanoidEnv WalkersOstrichtwoEnv WalkersOstrichEnv WalkersOstrichoneEnv WalkersHalfcheetahthreeEnv WalkersFullcheetahfiveEnv WalkersHalfhumanoidfourEnv WalkersFullcheetahtwoEnv WalkersHalfhumanoidtwoEnv WalkersHalfhumanoidoneEnv WalkersHalfcheetahEnv WalkersHopperoneEnv WalkersFullcheetahEnv WalkersFullcheetahthreeEnv WalkersFullcheetahzeroEnv WalkersHalfcheetahfourEnv WalkersHopperzeroEnv WalkersOstrichfiveEnv WalkersHopperfourEnv WalkersHalfcheetahfiveEnv WalkersHopperthreeEnv WalkersHalfhumanoidzeroEnv WalkersOstrichthreeEnv WalkersHopperfiveEnv modify_xml AntEnv CentipedeTwelveEnv CpCentipedeFourteenEnv CpCentipedeFourEnv CentipedeOnehundredEnv CentipedeThirtyEnv CentipedeFourteenEnv CpCentipedeTwelveEnv CentipedeFortyEnv CentipedeTwentyEnv CpCentipedeSixEnv CentipedeSevenEnv CentipedeThreeEnv CpCentipedeEightEnv CentipedeSixEnv CentipedeEnv CentipedeFiveEnv CentipedeTenEnv CpCentipedeTenEnv CentipedeFiftyEnv CentipedeFourEnv CentipedeEightEnv InvPendulumEnv InvPendulumOneEnv InvPendulumSixEnv InvPendulumFiveEnv InvPendulumThreeEnv InvPendulumFourEnv InvPendulumEightEnv InvPendulumNineEnv InvPendulumSevenEnv InvPendulumTwoEnv SwitcherTwoEnv AvoiderThreeEnv AvoiderOneEnv ReacherZeroEnv ReacherEightEnv AvoiderSevenEnv ReacherTwoEnv SwitcherSevenEnv AvoiderNineEnv ReacherFourEnv ReacherEnv SwitcherFourEnv AvoiderFourEnv AvoiderSixEnv ReacherSixEnv SwitcherEnv AvoiderEightEnv SwitcherEightEnv ReacherOneEnv AvoiderTwoEnv AvoiderZeroEnv ReacherThreeEnv ReacherNineEnv SwitcherFiveEnv AvoiderFiveEnv AvoiderEnv ReacherFiveEnv SwitcherOneEnv SwitcherNineEnv SwitcherThreeEnv ReacherSevenEnv SwitcherZeroEnv SwitcherSixEnv SnakeSixEnv BackSnakeEnv SnakeNineEnv SnakeEightEnv BackSnakeFiveEnv BackSnakeSixEnv SnakeSevenEnv SnakeThreeEnv SnakeEnv SnakeFourEnv SnakeTwentyEnv SnakeTenEnv BackSnakeNineEnv BackSnakeThreeEnv CrippledSnakeEnv BackSnakeFourEnv BackSnakeSevenEnv BackSnakeEightEnv SnakeFiveEnv get_ddpg_with_goal_targets get_ddpg_with_goal_vars ddpg_graph_with_goal transfer_ddpg_graph_no_goal get_transfer_ddpg_no_goal_targets get_transfer_ddpg_no_goal_vars transfer_ddpg_graph_with_goal get_transfer_ddpg_with_goal_vars get_transfer_ddpg_with_goal_targets Reacher2DOFEnv Reacher2DOFActEnv Reacher2DOFActCornerEnv Reacher2DOFActPushEnv Reacher2DOFActWallEnv Reacher2DOFCornerEnv Reacher2DOFPushEnv Reacher2DOFStoEnv Reacher2DOFWallEnv Reacher3DOFEnv Reacher3DOFCornerEnv Reacher3DOFPushEnv Reacher3DOFWallEnv Reacher4DOFEnv Reacher5DOFEnv Reacher6DOFEnv TP_Reacher2DOFEnv TP_WRITE_Reacher2DOFEnv WRITE_Reacher2DOFEnv generate_params items list placeholder items list format print get_ddpg_with_goal_vars exit get_ddpg_with_goal_targets get_ddpg_ph ddpg_graph_with_goal append InvertedPendulumEnv CustomMountainCarEnv CustomCartPoleEnv seed list TruncatedPendulumEnv exit shape TimeLimit prod update format PermutedPendulumEnv ModifiedPendulumEnv CustomPendulumEnv make items print array sigmoid logsigmoid variance_scaling_initializer format truncated_normal print exit append xavier_initializer format l2_regularizer print exit append l1_l2_regularizer l1_regularizer enumerate format leaky_relu relu print exit sigmoid append sigmoid high low sigmoid high low scale_fn get_regularizer get_activation get_initializer len list format print keys FIFOdeque multinomial cast int32 random_uniform mkdir remove exists print mimwrite asarray enumerate items list print capitalize mean around format print save_frames_as_video mean render reset append step range run format savez print save_frames_as_video squeeze mean render reset append step array range run normal format concatenate print save_frames_as_video close mean set_state_from_obs render reset open deque step append run join write close upper num2words open _add_custom _add_actuators _add_leg _add_body str replace str replace _add_leg range range _add_actuators _add_body update replace upper TASK_DICT range _add_actuators _add_body num2words replace sorted format feedforward print concat append keys list get_collection GLOBAL_VARIABLES keys concat assign list assign_add apply_gradients append expand_dims range format group square compute_gradients keys enumerate constant sigmoid_cross_entropy_with_logits print Variable AdamOptimizer reduce_mean append sorted keys feedforward list get_collection GLOBAL_VARIABLES keys gradients concat assign RMSPropOptimizer list reduce_sum assign_add apply_gradients append expand_dims range group square sqrt compute_gradients keys enumerate constant sigmoid_cross_entropy_with_logits Variable AdamOptimizer reduce_mean sorted feedforward concat append keys list print get_collection GLOBAL_VARIABLES keys gradients concat assign RMSPropOptimizer list reduce_sum assign_add apply_gradients append expand_dims range group square sqrt compute_gradients keys enumerate constant sigmoid_cross_entropy_with_logits Variable AdamOptimizer reduce_mean | # Domain Adaptive Imitation Learning (DAIL) This repo contains the official implementation for the paper [Domain Adaptive Imitation Learning](https://arxiv.org/abs/1910.00105). Authors: Kuno Kim, Yihong Gu, Jiaming Song, Shengjia Zhao, Stefano Ermon. For any questions, please correspond with Kuno Kim ([email protected]) ----------------------------------------------------------------------------------------- ### Dependencies Core python dependencies are `tensorflow==1.15.0, mujoco-py==0.5.7, gym==0.7.4`. Also be sure to have [tmux](https://github.com/tmux/tmux) installed to run the provided shell scripts and [virtualenv](https://pypi.org/project/virtualenv/) for package management. On a GPU machine, run the following to install all necessary python packages for our code. ```bash pip install -r requirements.txt ``` | 2,030 |
ermongroup/generative_adversary | ['adversarial attack'] | ['Provable defenses against adversarial examples via the convex outer adversarial polytope'] | models/resnet_model.py models/libs/ops.py models/aditi_mnist.py models/madry_mnist.py models/libs/resnet_ops.py models/vgg16.py train_acgan.py adv_utils.py models/libs/sn.py ops.py utils.py main.py models/acwgan_gp.py models/zico_mnist.py get_weights_path random_flip_left_right normalize_image median_filtering_2x2 feature_squeeze label_smooth unnormalize_image lrelu linear concat conv2d deconv2d conv_out_size_same conv_cond_concat bn main load_svhn4classifier save_images gradients check_folder load_mnist load_celebA4classifier write_labels load_celebA inverse_transform save_scattered_image center_crop imread show_all_variables imsave load_svhn get_image per_image_standardization discrete_cmap label_images load_mnist4classifier merge Rop Lop merge_images transform ACWGAN_GP AditiMNIST MadryModel ResNet vgg_16 ZicoMNIST lrelu batch_norm linear conv2d deconv2d gan_cond_batch_norm gan_batch_norm scope_has_variables UpsampleConv ResidualBlock ConvMeanPool ResidualBlockDisc ResidualBlock_celebA conv2d MeanPoolConv _l2normalize spectral_normed_weight get_weights_path random_flip_left_right normalize_image median_filtering_2x2 feature_squeeze label_smooth unnormalize_image lrelu linear concat conv2d deconv2d conv_out_size_same conv_cond_concat bn main load_svhn4classifier save_images gradients check_folder load_mnist load_celebA4classifier write_labels load_celebA inverse_transform save_scattered_image center_crop imread show_all_variables imsave load_svhn get_image per_image_standardization discrete_cmap label_images load_mnist4classifier merge Rop Lop merge_images transform ACWGAN_GP AditiMNIST MadryModel ResNet vgg_16 ZicoMNIST lrelu batch_norm linear conv2d deconv2d gan_cond_batch_norm gan_batch_norm scope_has_variables UpsampleConv ResidualBlock ConvMeanPool ResidualBlockDisc ResidualBlock_celebA MeanPoolConv _l2normalize spectral_normed_weight where reverse cast sample bool multiply rint div median_filtering_layer_2x2 join adv_gen trained makedirs classifier dataset adv get_shape as_list print ConfigProto seed join int asarray concatenate reshape extract_data astype shuffle zeros enumerate seed join concatenate print extract_data shuffle zeros print seed shuffle print seed join print extract_data shuffle zeros seed join int reshape extract_data astype shuffle zeros makedirs trainable_variables analyze_vars imread print format range str uint8 LINE_AA copyMakeBorder putText squeeze astype FONT_HERSHEY_SIMPLEX float32 resize append enumerate zeros enumerate squeeze merge int round center_crop imresize set_xlim grid colorbar scatter savefig figure gca set_ylim str name linspace base get_cmap sqrt moments maximum isinstance zeros_like range len ones_like isinstance sqrt sqrt sqrt as_list embedding_lookup batch_normalization moments get_variable as_list batch_normalization moments get_variable conv2d add_n add_n concat depth_to_space conv2 conv_shortcut partial relu gan_cond_batch_norm conv1 conv2 conv_shortcut partial relu gan_cond_batch_norm conv1 conv2 conv_shortcut partial relu conv1 as_list while_loop reshape add_to_collection assign get_variable | # Constructing Unrestricted Adversarial Examples with Generative Models This repo contains necessary code for reproducing main results in the paper [Constructing Unrestricted Adversarial Examples with Generative Models](https://arxiv.org/abs/1805.07894), NIPS 2018, Montréal, Canada. by [Yang Song](https://yang-song.github.io/), [Rui Shu](https://ruishu.io//), [Nate Kushman](http://www.kushman.org/) and [Stefano Ermon](https://cs.stanford.edu/~ermon/), Stanford AI Lab. --- We propose **Unrestricted Adversarial Examples**, a new kind of adversarial examples to machine learning systems. Different from traditional adversarial examples that are crafted by adding norm-bounded perturbations to clean images, unrestricted adversarial examples are _realistic images that are synthesized entirely from scratch_, and not restricted to small norm-balls. This new attack demonstrates the danger of a stronger **threat model**, where traditional defense methods for perturbation-based adversarial examples fail. ## Datasets Here are links to the datasets used in our experiments: * [CelebA (gender)](https://drive.google.com/open?id=1coLQbEZW6zshHVYi00IYSRiexq4RkA2x) * [SVHN](https://drive.google.com/open?id=1uPxNdW4K-GLFhqhOgtfI1jFFNEqp2eZn) ## Running Experiments | 2,031 |
ermongroup/neuralsort | ['stochastic optimization'] | ['Stochastic Optimization of Sorting Networks via Continuous Relaxations'] | tf/run_median.py pytorch/utils.py pytorch/pl.py pytorch/dknn_layer.py pytorch/models/easy_net.py tf/mnist_input.py tf/run_sort.py pytorch/run_baseline.py pytorch/models/preact_resnet.py pytorch/neuralsort.py pytorch/run_dknn.py tf/multi_mnist_cnn.py tf/sinkhorn.py pytorch/dataset.py tf/util.py DataSplit get_shapes ClassicDataset DKNN NeuralSort PL train experiment_id test experiment_id new_predict test majority acc train dknn_loss one_hot ConvNet PreActBlock PreActResNet50 PreActResNet PreActResNet18 test PreActResNet152 PreActBottleneck PreActResNet101 PreActResNet34 get_iterators get_iterator get_multi_mnist_input test_iterators split_digits select_digit max_pool_2x2 conv2d deepnn bias_variable weight_variable save_model load_model prnt test train get_median_probs save_model load_model prnt test vec_gradient train gumbel_sinkhorn sinkhorn_operator neuralsort prop_any_correct bl_matmul br_matmul prop_correct sample_gumbel pl_log_density sum ce_loss backward linear_layer print len zero_grad mean item append to step h_phi sum linear_layer print len mean eval mkdir save item append to argmax h_phi sum dknn_layer zip reshape dknn_loss flush argsort sum unsqueeze take new_predict unsqueeze scatter_ randn size PreActResNet18 net int median concatenate insert reshape astype stack append randint array range from_generator batch prefetch get_iterators truncated_normal constant reduce_sum print save latest_checkpoint restore print prnt range run save_model prnt run range trainable_variables concat gradients prnt reshape range reshape transpose tile sinkhorn_operator sample_gumbel ones transpose bl_matmul cast softmax expand_dims abs range reduce_mean argmax cast equal argmax reduce_all equal random_uniform exp squeeze reduce_sum add_n expand_dims | # Stochastic Optimization of Sorting Networks via Continuous Relaxations This repository provides a reference implementation for learning NeuralSort-based models as described in the paper: > Stochastic Optimization of Sorting Networks via Continuous Relaxations > [Aditya Grover](https://aditya-grover.github.io), [Eric Wang](https://ericjwang.com), Aaron Zweig and [Stefano Ermon](https://cs.stanford.edu/~ermon/). > International Conference on Learning Representations (ICLR), 2019. > Paper: https://openreview.net/pdf?id=H1eSS3CcKX ## Requirements The codebase is implemented in Python 3.7. To install the necessary requirements, run the following commands: ``` pip install -r requirements.txt | 2,032 |
ernestgong/data2text-three-dimensions | ['table to text generation', 'time series', 'text generation'] | ['Table-to-Text Generation with Effective Hierarchical Encoder on Three Dimensions (Row, Column and Time)'] | onmt/modules/CopyGenerator.py onmt/io/IO.py onmt/io/DatasetBase.py onmt/modules/AudioEncoder.py text2num.py onmt/io/__init__.py onmt/io/TextDataset.py onmt/modules/UtilClass.py onmt/modules/__init__.py onmt/modules/WeightNorm.py non_rg_metrics.py onmt/modules/Embeddings.py onmt/translate/Translation.py onmt/modules/SRU.py opts.py onmt/modules/ImageEncoder.py onmt/ModelConstructor.py onmt/Utils.py train.py onmt/translate/Beam.py onmt/modules/GlobalAttention.py onmt/io/BoxField.py onmt/io/ImageDataset.py onmt/Optim.py onmt/translate/Translator.py onmt/__init__.py onmt/translate/__init__.py onmt/modules/Conv2Conv.py onmt/Loss.py onmt/modules/Gate.py preprocess.py onmt/translate/Penalties.py onmt/modules/PointerAttention.py onmt/modules/StackedRNN.py onmt/Trainer.py translate.py onmt/modules/Transformer.py onmt/modules/StructuredAttention.py onmt/modules/GlobalSelfAttention.py onmt/io/AudioDataset.py onmt/Models.py onmt/modules/MultiHeadedAttn.py onmt/modules/ConvMultiStepAttention.py trip_match norm_dld same_ent dedup_triples calc_dld int_value calc_precrec get_triples add_md_help_argument MarkdownHelpFormatter train_opts MarkdownHelpAction preprocess_opts DeprecateAction model_opts translate_opts build_save_dataset build_save_text_dataset_in_shards build_save_vocab main parse_args makedir check_existing_pt_files NumberException text2num train_model make_loss_compute check_save_model_path make_dataset_iter tally_parameters collect_report_features build_optim build_model lazily_load_dataset DatasetLazyIter report_func main load_fields _report_rouge _report_bleu main makedir _report_score NMTLossCompute filter_shard_state shards LossComputeBase make_encoder make_base_model make_embeddings make_decoder load_test_model EncoderBase InputFeedRNNDecoder PositionwiseFeedForward RNNDecoderState RNNEncoder rnn_factory PositionEmb NMTModel StdRNNDecoder RNNDecoderBase DecoderState HierarchicalEncoder PositionalEncoding dumpEmb Optim Statistics Trainer use_gpu aeq sequence_mask AudioDataset BoxField ONMTDatasetBase ImageDataset collect_features _build_field_vocab merge_vocabs collect_feature_vocabs _make_examples_nfeats_tpl OrderedIterator save_fields_to_vocab load_fields_from_vocab _setstate get_num_features make_features _getstate build_dataset get_fields build_vocab TextDataset ShardedTextCorpusIterator AudioEncoder CNNEncoder CNNDecoderState shape_transform GatedConv StackedCNN CNNDecoder ConvMultiStepAttention seq_linear CopyGenerator CopyGeneratorCriterion CopyGeneratorLossCompute ProcessChar PositionalEncoding Embeddings dumpEmb SourceContextGate ContextGate context_gate_factory TargetContextGate BothContextGate GlobalAttention GlobalSelfAttention ImageEncoder MultiHeadedAttention PointerAttention check_sru_requirement SRU_Compute SRU SRUCell CheckSRU StackedLSTM StackedGRU MatrixTree PositionwiseFeedForward TransformerDecoderLayer TransformerDecoder TransformerDecoderState TransformerEncoder TransformerEncoderLayer LayerNorm Elementwise WeightNormLinear get_var_maybe_avg WeightNormConv2d WeightNormConvTranspose2d get_vars_maybe_avg Beam GNMTGlobalScorer PenaltyBuilder Translation TranslationBuilder Translator int text2num trip_match add set range len append dedup_triples print get_triples float sum enumerate join trip_match range len print get_triples float enumerate len add_argument_group add_argument add_argument_group add_argument add_argument_group add_argument add_argument_group add_argument add_argument makedirs glob write save_data exit seed add_md_help_argument save_data preprocess_opts ArgumentParser manual_seed makedir check_existing_pt_files TextDataset format src_seq_length_trunc save_data print num_feats save tgt_seq_length_trunc append getsize max_shard_size ShardedTextCorpusIterator valid_src format train_src_hist save_data build_dataset print train_src data_type save valid_src_hist train_tgt valid_tgt train_ptr save_data save_fields_to_vocab data_type tgt_vocab_size save share_vocab src_vocab_size tgt_words_min_frequency src_words_min_frequency build_vocab build_save_dataset train_src_hist print train_src build_save_vocab data_type get_num_features parse_args train_tgt get_fields get enumerate split log_tensorboard tensorboard Statistics output exp_host log generator use_gpu copy_attn_force copy_loss_by_seqlength NMTLossCompute cuda CopyGeneratorLossCompute make_loss_compute validate batch_size Trainer log accum_count make_dataset_iter drop_checkpoint epoch_step max_generator_batches normalization range vocab log_tensorboard gpuid ppl start_epoch lr print accuracy lazily_load_dataset truncated_decoder train epochs save_model dirname abspath makedirs print sum named_parameters glob sorted data load data vocab print train_from load_fields_from_vocab dict len print collect_features enumerate print make_base_model DataParallel use_gpu learning_rate Optim train_from print set_parameters named_parameters max_grad_norm load_state_dict optim state_dict load train_model tally_parameters collect_report_features check_save_model_path build_optim build_model train_from tensorboard close lazily_load_dataset next load_fields print print decode strip print decode strip stage1 verbose ArgumentParser TranslationBuilder cuda fields log GNMTGlobalScorer open count dump_beam report_rouge tgt set_device from_batch encode beam_accum build_dataset load_test_model dump __dict__ src_hist translate_batch length_penalty alpha OrderedIterator Translator flush _report_score _report_rouge join coverage_penalty _report_bleu src write output replace_unk model_opts beta n_best gpu report_bleu Variable items list data list backward filter_shard_state dict zip tgt_word_vec_size Embeddings src_word_vec_size feat_vec_size len load make_base_model use_gpu load_fields_from_vocab eval model2 make_encoder CopyGenerator share_embeddings cuda pre_word_vecs_dec hasattr make_embeddings fix_word_vecs_enc NMTModel pre_word_vecs_enc make_decoder model_type load_state_dict rnn_size encoder param_init collect_feature_vocabs vocab fix_word_vecs_dec dropout load_pretrained_vectors window_size uniform_ ImageEncoder decoder enc_layers print sample_rate AudioEncoder brnn parameters cpu weight SRU next numel update defaultdict stoi collect_features list defaultdict items dict stoi get_fields len items list dict stoi append sum Counter sorted isinstance append str count append str count vocab TextDataset _make_examples_nfeats_tpl make_text_examples_nfeats_tpl ImageDataset AudioDataset list vocab_cls fromkeys load update str examples _build_field_vocab merge_vocabs print n_tgt_feats Counter getattr range make_text_examples_nfeats_tpl make_audio_examples_nfeats_tpl make_image_examples_nfeats_tpl linear size view check_output compile getenv getattr append get_var_maybe_avg | # data2text-three-dimensions This repo contains code for [Table-to-Text Generation with Effective Hierarchical Encoder on Three Dimensions (Row, Column and Time)](https://www.aclweb.org/anthology/D19-1310.pdf) (Gong, H., Feng, X., Qin, B., & Liu, T.; EMNLP 2019); this code is based on [data2text-plan-py](https://github.com/ratishsp/data2text-plan-py). ## Requirement All dependencies can be installed via: ```bash pip install -r requirements.txt ``` Note that requirements.txt contains necessary version requirements of certain dependencies and Python version is 3.6. ## data and model Before executing commands in the following sections, the data and/or trained model need to be downloaded and extracted. They are available as a single tar.gz file at link https://www.dropbox.com/s/hnij20b0rm5dn9b/data_and_model.tar.gz?dl=0 (suited for global user) or https://pan.baidu.com/s/11mXCQnWFW9TLM2O3bfeHnQ (retrieval code: cqxx ). Please move the extracted folder rotowire and paper_model into this repo's folder. | 2,033 |
errec-sun/mnist | ['data augmentation'] | ['DENSER: Deep Evolutionary Network Structured Representation'] | utils/helper.py configs.py benchmark/convnet.py app.py benchmark/runner.py utils/argparser.py utils/mnist_reader.py visualization/project_zalando.py start_s3_sync get_json_logger touch touch_dir _get_logger main cnn_model_fn PredictJob JobWorker JobManager get_args_request parse_arg get_args_cli now_int upload_result_s3 get_sprite_image invert_grayscale create_sprite_image vector_to_matrix_mnist UploadS3Thread load_mnist UploadS3Thread start Event dirname makedirs makedirs setFormatter touch_dir DEBUG getLogger addHandler StreamHandler Formatter touch setLevel INFO FileHandler setFormatter getLogger addHandler Formatter touch setLevel INFO FileHandler dense max_pooling2d dropout one_hot minimize reshape GradientDescentOptimizer conv2d softmax_cross_entropy asarray evaluate print Estimator shuffle labels images numpy_input_fn train range read_data_sets int append items list defaultdict utcfromtimestamp info int isinstance ones sqrt ceil array range vector_to_matrix_mnist invert_grayscale join | # Fashion-MNIST [](https://github.com/zalandoresearch/fashion-mnist/) [](https://gitter.im/fashion-mnist/Lobby?utm_source=share-link&utm_medium=link&utm_campaign=share-link) [](README.zh-CN.md) [](README.ja.md) [](https://opensource.org/licenses/MIT) [](https://hanxiao.github.io/2018/09/28/Fashion-MNIST-Year-In-Review/) <details><summary>Table of Contents</summary><p> * [Why we made Fashion-MNIST](#why-we-made-fashion-mnist) * [Get the Data](#get-the-data) | 2,034 |
ervinK/mxnetpixel | ['scene text detection', 'text classification', 'instance segmentation', 'semantic segmentation'] | ['PixelLink: Detecting Scene Text via Instance Segmentation'] | train.py ImgLib/ImgFormat.py custom_softmax_cross_entropy_loss.py util.py ImgLib/util.py ImgLib/ImgTransform.py net.py predict.py config.py ImgLib/ImgShow.py datasets.py ImgLib/testImgTransform.py main.py postprocess.py loss.py _reshape_like SoftmaxCrossEntropyLoss _apply_weighting Loss PixelLinkDataset ICDAR15Dataset PixelLinkLoss main custom_conv2d Net sum_matrix mask_to_box is_valid_cord mask_filter func get_neighbors predict_main result_draw try_gpu train try_gpu get_colormap2label read_voc_images bilinear_kernel get_iter_num VOCSegDataset voc_rand_crop load_latest_params voc_label_indices ImgOrderFormat ImgColorFormat DrawLabels RotateImageWithLabel normalize_image ZeroMeanImage ReadImage CropImage UnzeroMeanImage _ParseImage _ReturnImage ResizeImage RotateImage ResizeImageWithLabel CropImageWithLabel main testTransformWithLabel ReadLabels testTransform find_contours broadcast_mul initialize try_gpu PixelLinkDataset epoch Xavier train Trainer Net DataLoader train_labels_dir train_images_dir load_latest_params collect_params load_parameters saving_model_dir shape pad max Sequential add BatchNorm Conv2D Activation uint8 astype logical_and shape softmax zeros range append uint8 print findContours boxPoints astype short_side_filter func append minAreaRect max range list fromkeys zeros_like find_root len shape joint zip get_neighbors asnumpy enumerate cpu drawContours try_gpu mask_to_box Xavier b_mean mask_filter initialize shape g_mean append expand_dims Net ReadImage as_in_context load_parameters net ZeroMeanImage r_mean result_draw array join time backward print PixelLinkLoss range as_in_context step save_parameters saving_model_dir enumerate array gpu findall compile fixed_crop random_crop zeros enumerate astype zeros abs imread enumerate sort join listdir get_iter_num drawContours reshape transpose ascontiguousarray array fromarray ReadImage uint8 randint _ParseImage rotate deepcopy reshape size tolist astype _ParseImage range array RotateImage int size min random sqrt randint _ParseImage round crop deepcopy reshape tolist astype CropImage array _ParseImage resize deepcopy reshape size tolist astype array _ParseImage ResizeImage array float astype float astype imwrite ReadImage CropImage ResizeImage RotateImage imwrite RotateImageWithLabel DrawLabels ReadImage ReadLabels ResizeImageWithLabel CropImageWithLabel testTransform findContours asarray copy CHAIN_APPROX_SIMPLE | # reference 《PixelLink Detecting Scene Text via Instance Segmentation》 https://arxiv.org/pdf/1801.01315.pdf # Pixellink-mxnet reconsitution pixellink by mxnet <br/> # Environment python3.5\python3.6\python3.7<br/> mxnet==1.5.1<br/> numpy==1.17.4<br/> cv2==4.1.1<br/> | 2,035 |
escorciav/Mixture-of-Embedding-Experts | ['video retrieval'] | ['Learning a Text-Video Embedding from Incomplete and Heterogeneous Data'] | LSMDC.py train.py loupe.py loss.py qcm_sampler.py MSR_sampler.py MSRVTT.py model.py MaxMarginRankingLoss NetVLAD NetRVLAD LSMDC LSMDC_qcm Net Gated_Embedding_Unit Context_Gating MEE MSRVTT MSRSampler QCMSampler verbose compute_metric make_tensor print sort where median float sum diag len zeros range len | # Mixture-of-Embeddings-Experts This github repo provides a Pytorch implementation of the Mixture-of-Embeddings-Experts model (MEE) [1]. ## Usage example Creating an MEE block: ```python from model import MEE ''' Initializig an MEE module Input: - video_modality_dim: dictionary of all video modality with input dimension and output embedding dimension. | 2,036 |
escorciav/Text-to-Clip_Retrieval | ['weakly supervised action localization', 'action localization'] | ['Guess Where? Actor-Supervision for Spatiotemporal Action Localization'] | experiments/Text_to_Clip/_init_paths.py caffe3d/tools/extra/extract_seconds.py lib/setup.py lib/tdcnn/__init__.py caffe3d/python/caffe/test/test_net_spec.py caffe3d/python/caffe/test/test_io.py caffe3d/python/caffe/pycaffe.py lib/rpn/__init__.py lib/utils/__init__.py lib/roi_data_layer/roidb.py caffe3d/python/caffe/io.py experiments/Text_to_Clip/test_fast/utils.py caffe3d/python/detect.py caffe3d/python/caffe/classifier.py experiments/train_rpn/train_net.py lib/roi_data_layer/layer_retrieval_pairwiserank_caption_share.py caffe3d/examples/c3d_finetune/check_data_and_model.py experiments/extract_HDF_for_LSTM/_init_paths.py lib/rpn/caption_target_layer.py experiments/Text_to_Clip/test_fast/evaluation/evaluation_recall.py caffe3d/scripts/download_model_binary.py caffe3d/python/caffe/test/test_solver.py lib/lstm/lstm_last_hidden_state_layer.py lib/utils/caption_generator.py caffe3d/examples/pycaffe/layers/pascal_multilabel_datalayers.py caffe3d/src/caffe/test/test_data/generate_sample_data.py caffe3d/examples/pycaffe/caffenet.py caffe3d/examples/pycaffe/tools.py experiments/train_rpn/_init_paths.py preprocess/util_2.py lib/rpn/anchor_target_layer.py preprocess/generate_roidb_modified_freq1.py preprocess/generate_roidb_modified_freq1_full_retrieval_test.py lib/rpn/proposal_target_layer_binary.py lib/tdcnn/test_retrieval_caption_share_fast.py lib/tdcnn/nms_wrapper.py caffe3d/examples/web_demo/app.py lib/rpn/pad_controller_layer.py caffe3d/python/caffe/test/test_coord_map.py caffe3d/python/classify.py caffe3d/tools/extra/summarize.py lib/rpn/generate_anchors.py lib/tdcnn/test_caption_extract_HDF_for_LSTM.py caffe3d/python/caffe/test/test_layer_type_list.py caffe3d/python/caffe/__init__.py lib/rpn/GTboxes_to_GTrois.py experiments/Text_to_Clip/test_fast/_init_paths.py lib/rpn/gt_caption_target_layer_hierarchical.py caffe3d/examples/finetune_flickr_style/assemble_data.py lib/roi_data_layer/layer_caption.py lib/tdcnn/config.py caffe3d/python/caffe/net_spec.py caffe3d/examples/c3d_ucf101/check_data_and_model.py lib/tdcnn/twin_transform.py lib/rpn/Prepare_Paired_Positive_Negative_Sim.py lib/utils/timer.py caffe3d/python/caffe/test/test_python_layer_with_param_str.py experiments/extract_HDF_for_LSTM/test_net.py caffe3d/python/caffe/coord_map.py caffe3d/tools/extra/py_plot_training_loss.py experiments/Text_to_Clip/train_net.py caffe3d/examples/c3d_finetune/convert_npy_to_binaryproto.py caffe3d/examples/web_demo/exifutil.py caffe3d/python/caffe/detector.py lib/roi_data_layer/minibatch.py caffe3d/scripts/copy_notebook.py caffe3d/python/caffe/test/test_python_layer.py experiments/Text_to_Clip/test_fast/test_net.py lib/rpn/GTbox_sorted.py lib/roi_data_layer/minibatch_caption.py caffe3d/python/draw_net.py lib/rpn/get_controller_hidden_layer.py lib/nms/py_cpu_nms.py caffe3d/scripts/cpp_lint.py lib/rpn/proposal_layer.py lib/roi_data_layer/__init__.py caffe3d/tools/extra/parse_log.py caffe3d/python/caffe/draw.py experiments/extract_HDF_for_LSTM/utils.py caffe3d/python/caffe/test/test_net.py lib/rpn/caption_layer.py caffe3d/tools/extra/resize_and_crop_images.py lib/utils/blob.py lib/tdcnn/train.py caffe3d/examples/pycaffe/layers/pyloss.py lib/tdcnn/train_retrieval_pairwiserank_adam_caption.py download_image make_net max_pool caffenet conv_relu fc_relu CaffeSolver SimpleTransformer print_info check_params PascalMultilabelDataLayerSync load_pascal_annotation BatchLoader EuclideanLossLayer start_tornado start_from_terminal embed_image_html classify_upload index allowed_file ImagenetClassifier classify_url open_oriented_im apply_orientation main main main parse_args Classifier coord_map UndefinedMapException conv_params coord_map_from_to AxisMismatchException inverse crop_params compose crop Detector get_edge_label draw_net get_layer_label get_pydot_graph choose_color_by_layertype get_pooling_types_dict draw_net_to_file Transformer blobproto_to_array datum_to_array array_to_blobproto array_to_datum resize_image arraylist_to_blobprotovector_str blobprotovector_str_to_arraylist load_image oversample Layers Function Parameters Top NetSpec assign_proto param_name_dict to_proto _Net_blobs _Net_forward_all _Net_set_input_arrays _Net_backward _Net_params _Net_forward _Net_outputs _Net_forward_backward_all _Net_blob_loss_weights _Net_batch _Net_get_id_name _Net_inputs TestCoordMap coord_net_spec TestBlobProtoToArray TestArrayToDatum TestLayerTypeList TestLevels TestStages simple_net_file TestNet TestAllInOne lenet TestNetSpec silent_net anon_lenet exception_net_file parameter_net_file SimpleLayer phase_net_file TestPythonLayer ParameterLayer PhaseLayer python_net_file ExceptionLayer SimpleParamLayer TestLayerWithParam python_param_net_file TestSolver ParseNolintSuppressions CheckVlogArguments CheckSectionSpacing FindNextMultiLineCommentEnd ReplaceAll CheckForFunctionLengths _SetOutputFormat _IsTestFilename _VerboseLevel CheckBraces RemoveMultiLineComments ResetNolintSuppressions CheckForNonStandardConstructs _SetVerboseLevel PrintUsage _NestingState CheckIncludeLine CheckAccess _CppLintState Search CheckInvalidIncrement RemoveMultiLineCommentsFromRange CleansedLines CheckForBadCharacters UpdateIncludeState FindPreviousMatchingAngleBracket CheckEmptyBlockBody FindNextMultiLineCommentStart Match _NamespaceInfo CheckMakePairUsesDeduction CheckCheck IsBlankLine _SetFilters ProcessLine _FunctionState CheckPosixThreading GetLineWidth GetHeaderGuardCPPVariable IsCppString _IncludeState CheckSpacing _ClassInfo CheckForCopyright IsErrorSuppressedByNolint ProcessFileData CheckForMultilineCommentsAndStrings CloseExpression _PreprocessorInfo _OutputFormat CheckForIncludeWhatYouUse CheckSpacingForFunctionCall FindEndOfExpressionInLine FindNextMatchingAngleBracket _SetCountingStyle ProcessFile _IncludeError CleanseRawStrings CheckAltTokens CheckForNewlineAtEOF ParseArguments CheckForNonConstReference PrintCategories _Filters main FilesBelongToSameModule CheckCStyleCast FileInfo _BlockInfo CheckForHeaderGuard CheckCaffeDataLayerSetUp ReverseCloseExpression CleanseComments _DropCommonSuffixes _ClassifyInclude CheckStyle CheckCaffeAlternatives FindStartOfExpressionInLine _ShouldPrintError CheckComment Error _GetTextInside CheckLanguage CheckCaffeRandom GetPreviousNonBlankLine reporthook parse_readme_frontmatter model_checks_out valid_dirname get_start_time extract_seconds extract_datetime_from_line get_log_created_year write_csv parse_log fix_initial_nan_learning_rate save_csv_files main parse_args parse_line_for_net_output main ResizeCropImagesMapper PILResizeCrop OpenCVResizeCrop print_table printed_len summarize_net main read_net format_param parse_args get_test_roidb get_blocked_videos segment_iou interpolated_prec_rec add_path parse_args get_train_roidb add_path parse_args get_test_roidb get_blocked_videos segment_iou interpolated_prec_rec add_path computer_eval _tiou _nms parse_args get_train_roidb add_path find_in_path customize_compiler_for_nvcc custom_build_ext locate_cuda LSTMLastLayer py_cpu_nms RoIDataLayer BlobFetcher RoIDataLayer get_minibatch _get_video_blob get_minibatch _get_video_blob prepare_roidb _compute_targets add_twin_regression_targets _unmap _compute_targets AnchorTargetLayer CaptionLayer _filter_wins _compute_targets CaptionTargetLayer _get_twin_regression_labels _sample_positive_rois generate_anchors _scale_enum _mkanchors _whctrs CaptionTargetLayer ProposalTargetLayer ProposalTargetLayer CaptionTargetLayer CaptionTargetLayer ExtractPairSimLayer _filter_wins ProposalLayer _sample_hard_rois _sample_all_rois _sample_rois _compute_targets ProposalTargetLayer _get_twin_regression_labels cfg_from_file cfg_from_list _merge_a_into_b nms vis_detections test_net generate_sentence _get_video_blob _get_blobs softmax apply_nms random_choice_from_probs video_detect vis_detections test_net generate_sentence _get_video_blob cos dist _get_blobs softmax extract_retrieval_score apply_nms random_choice_from_probs lstm_last_hidden_state video_detect train_net SolverWrapper train_net SolverWrapper clip_wins twin_transform twin_transform_inv prep_im_for_blob video_list_to_blob Caption TopN CaptionGenerator Timer generate_roi generate_roidb generate_roi generate_roidb init_vocabulary split_sentence generate_segment get_fps generate_segment_withRandomQueries ffmpeg ffmpeg_convert_fps line_to_stream mkdir resize dump_vocabulary rm get_length imread urlretrieve Convolution InnerProduct Data SoftmaxWithLoss LRN Accuracy max_pool InnerProduct conv_relu fc_relu Dropout join list getElementsByTagName get_data_from_tag csr_matrix dict zip zeros float range enumerate len print format get read info load_image classify_image StringIO join replace info secure_filename save filename open_oriented_im classify_image fromarray replace astype save resize StringIO items listen HTTPServer format print start WSGIContainer update start_tornado add_option OptionParser debug port parse_args ImagenetClassifier forward run hasattr _getexif astype float32 tile apply_orientation open transpose model_def endswith ArgumentParser save mean_file channel_swap output_file dirname expanduser parse_args input_file predict Classifier set_mode_cpu load time isdir print add_argument set_mode_gpu pretrained_model gpu len DataFrame Detector format to_hdf detect_selective_search mean set_index to_csv detect_windows read_csv add_argument ArgumentParser read NetParameter output_image_file rankdir Merge TRAIN draw_net_to_file TEST get params array get params array crop_params conv_params pop collect_bottoms add fn coord_map compose coord_map_from_to items list DESCRIPTOR batch_size str num_output print join get_pooling_types_dict add_edge get_edge_label list Dot exclude get_layer_label add_node values choose_color_by_layertype Edge Node bottom append type layer include top data array diff shape BlobProto extend flat extend BlobProtoVector ParseFromString BlobProtoVector extend tostring shape Datum flat data len astype float32 tile zoom tuple resize fill empty array range concatenate shape tile empty array LayerParameter list NetParameter _to_proto extend Counter OrderedDict values iteritems hasattr isinstance extend add getattr setattr list OrderedDict _blobs _blob_names zip list _blob_loss_weights OrderedDict _blob_names zip OrderedDict list keys list keys iteritems layers index set outputs _forward len iteritems _backward layers inputs index set len iteritems asarray extend copy next _batch itervalues forward len iteritems asarray backward extend copy next _batch itervalues zip_longest zip forward len ascontiguousarray concatenate itervalues zeros next range len data Pooling pool Convolution NetSpec Deconvolution conv Input NamedTemporaryFile str close write data Pooling pool1 conv2 pool2 ip1 relu1 SoftmaxWithLoss Convolution NetSpec DummyData ip2 ReLU InnerProduct label conv1 Pooling SoftmaxWithLoss Convolution DummyData ReLU InnerProduct data NetSpec DummyData Silence data2 error search add group clear compile compile compile SetOutputFormat SetCountingStyle SetFilters _Filters startswith IsErrorSuppressedByNolint _ShouldPrintError write IncrementErrorCount replace append Match group find startswith endswith range error FindNextMultiLineCommentEnd RemoveMultiLineCommentsFromRange FindNextMultiLineCommentStart rstrip find range len FindEndOfExpressionInLine range len FindStartOfExpressionInLine error min search I range len FileInfo RepositoryName sep sub ParseNolintSuppressions error startswith split GetHeaderGuardCPPVariable enumerate error enumerate error len error replace count error find error find error find error find error Search error match InnermostClass replace error escape Match Search error group Search Check error lines Count End group Begin NumLines Match raw_lines range Search error match group error Match group pop group append Search pop group append Search elided replace CheckSpacingForFunctionCall rfind error len group min CloseExpression NumLines sub find CheckComment Match range Search lines_without_raw_strings error group starting_linenum Match range Search error rfind len group ReverseCloseExpression Search Match CloseExpression find error Match CloseExpression find elided error strip group FindEndOfExpressionInLine find Match range CloseExpression len error Match finditer normalize isinstance GetLineWidth int InnermostClass CheckCheck error CheckAltTokens CheckBraces CheckSpacing CheckSectionSpacing CheckEmptyBlockBody CheckAccess GetHeaderGuardCPPVariable lines_without_raw_strings _DropCommonSuffixes RepositoryName match split CheckNextIncludeOrder CanonicalizeAlphabeticalOrder FileInfo error search group SetLastHeader match _ClassifyInclude Match pop end search set append values M rstrip replace CheckCStyleCast error _GetTextInside CheckIncludeLine search group lstrip startswith Match ResetSection Search split rfind error group ReverseCloseExpression lstrip findall Match range Search ReplaceAll error Match Search endswith replace setdefault group search CleanseComments open list FilesBelongToSameModule error search copy sub NumLines FullName keys range error search CheckPosixThreading ParseNolintSuppressions CheckVlogArguments CheckMakePairUsesDeduction CheckCaffeDataLayerSetUp CheckLanguage CheckInvalidIncrement CheckCaffeRandom CheckForNonConstReference check_fn Update CheckForNonStandardConstructs CheckStyle raw_lines CheckForMultilineCommentsAndStrings CheckCaffeAlternatives CheckForFunctionLengths CleansedLines _NestingState CheckForBadCharacters CheckForNewlineAtEOF _IncludeState RemoveMultiLineComments CheckForCopyright ResetNolintSuppressions CheckForHeaderGuard NumLines CheckCompletedBlocks CheckForIncludeWhatYouUse range ProcessLine _FunctionState Error rstrip endswith len write ProcessFileData _SetVerboseLevel range split write exit join write exit _VerboseLevel int getopt _SetOutputFormat set _SetVerboseLevel PrintCategories _SetFilters _OutputFormat PrintUsage _SetCountingStyle split getreader ParseArguments ResetErrorCounts stderr exit verbose_level PrintErrorCounts StreamReaderWriter ProcessFile getwriter int time write flush load join index int rfind datetime split getctime year strip extract_datetime_from_line get_start_time total_seconds strip write get_log_created_year close extract_datetime_from_line open float get_log_created_year compile fix_initial_nan_learning_rate search group OrderedDict append float join basename write_csv print excel parse_log save_csv_files output_dir logfile_path subplots grid add_subplot open exp ones set_xlabel savefig legend append plot readlines close float convolve set_ylabel figure isfile array set_ylim split NetParameter decay_mult format name lr_mult append print zip len get join str format convolution_param list setdefault param kernel_size map set top bottom append type module layer enumerate print_table filename summarize_net read_net exit print_help load open Request urlopen format sum hstack max minimum clip astype maximum insert load open minimum maximum append maximum minimum items hstack vstack append _tiou empty enumerate pathsep pjoin exists split find_in_path items pjoin pathsep dirname sep append _compile compiler_so append maximum minimum FG_FRACTION BATCH_SIZE _get_video_blob HAS_RPN shape empty randint round zeros len video_list_to_blob tuple prep_im_for_blob waitKey imshow PIXEL_MEANS append zeros imread range CROP_SIZE destroyAllWindows enumerate video_path_at toarray start_frame roidb video_index argmax max range len EPS TWIN_NORMALIZE_STDS NUM_CLASSES _compute_targets print mean sqrt TWIN_NORMALIZE_TARGETS_PRECOMPUTED TWIN_NORMALIZE_MEANS tile TWIN_NORMALIZE_TARGETS zeros array range twin_overlaps ascontiguousarray twin_transform zeros argmax fill empty TWIN_INSIDE_WEIGHTS zeros int shape TWIN_NORMALIZE_STDS TWIN_NORMALIZE_TARGETS_PRECOMPUTED TWIN_NORMALIZE_MEANS array transpose twin_overlaps ascontiguousarray argmax max array _scale_enum hstack _whctrs _mkanchors size min twin_overlaps ascontiguousarray _compute_targets choice append argmax max _get_twin_regression_labels _compute_targets twin_overlaps ascontiguousarray argmax max _get_twin_regression_labels size min twin_overlaps ascontiguousarray _compute_targets choice append argmax max _get_twin_regression_labels items ndarray isinstance type array _merge_a_into_b literal_eval zip split exp isnan isinf sum max random softmax enumerate append random_choice_from_probs forward array print _get_rois_blob reshape forward _get_blobs show minimum format add_patch imshow title Rectangle range cla nms range copy len permutation arange _get_video_blob vstack exists video_detect shape tic LSTM_BATCH_SIZE append range format average_time hstack close HAS_RPN empty toc print makedirs File rmtree create_dataset randint len data sum T ones reshape shape empty tile forward enumerate join astype float32 dot unique HAS_RPN array reshape forward extract_retrieval_score max open squeeze transpose dump enumerate minimum maximum zeros lstm_last_hidden_state print train_model filter_roidb SolverWrapper transpose log dtype exp astype shape zeros minimum maximum transpose shape zeros range len astype float32 resize ones print zfill array len minimum deepcopy max list generate_roi print min maximum set keys append listdir array range len list sorted print strip append keys enumerate len print append strip list split_sentence print line_to_stream append listdir keys range len list split_sentence print line_to_stream append listdir keys range len communicate Popen communicate loads Popen makedirs rmtree communicate Popen communicate Popen imread imwrite | # Multilevel Language and Vision Integration for Text-to-Clip Retrieval Code released by Huijuan Xu (Boston University). ### Introduction We address the problem of text-based activity retrieval in video. Given a sentence describing an activity, our task is to retrieve matching clips from an untrimmed video. Our model learns a fine-grained similarity metric for retrieval and uses visual features to modulate the processing of query sentences at the word level in a recurrent neural network. A multi-task loss is also employed by adding query re-generation as an auxiliary task. ### License | 2,037 |
esennesh/categorical_bpl | ['program induction'] | ['Learning a Deep Generative Model like a Program: the Free Category Prior'] | base/base_type.py model/modules.py utils/util.py base/base_trainer.py train.py logger/visualization.py utils/name_stack.py trainer/__init__.py logger/logger.py data_loader/data_loaders.py model/loss.py utils/__init__.py trainer/trainer.py test.py logger/__init__.py parse_config.py model/metric.py model/model.py base/__init__.py base/base_model.py new_project.py base/base_data_loader.py _set_by_path ConfigParser _get_opt_name _update_config _get_by_path main main BaseTargetBatchDataLoader BaseDataLoader TargetBatchRandomSampler BaseModel TypedModel BaseTrainer tensor_type _label_dtype type_size OmniglotTargetBatchDataLoader FashionMnistDataLoader MnistTargetBatchDataLoader OmniglotTargetTransform FashionMnistTargetBatchDataLoader OmniglotDataLoader MnistDataLoader setup_logging TensorboardWriter nll_loss top_k_acc accuracy VlaeCategoryModel CategoryModel VaeCategoryModel GlimpseCategoryModel LadderPrior DensityEncoder LadderDecoder DiagonalGaussian SpatialTransformerWriter SpatialTransformerReader ContinuousBernoulliModel NullPrior LadderPosterior glimpse_transform DensityNet DensityDecoder inverse_glimpse LadderEncoder StandardNormal CanvasPrior CanvasEncoder Trainer EmCounter ExpectationMaximizationOptim NamePushMessenger _stacked_name_base name_push name_pop NamePopMessenger _count_stacked MetricTracker read_json show_tensor powers_of inf_loop write_json ensure_dir _set_by_path list items startswith split load init_obj update format sampler device DataParallel eval getattr resume load_state_dict info zeros to get_logger len train Trainer split_validation optim search str basicConfig format list items print read_json dictConfig Path is_file to view max split NamePushMessenger NamePopMessenger show int view imshow sqrt numpy ceil range floor log mkdir Path Path Path repeat | ## Categorical Bayesian Program Learning (BPL) This project tests out the [Discopyro](https://github.com/neu-pml/discopyro/) framework for generative modeling of compositional structures on a series of compositional variational autoencoder architectures. ## License This project is licensed under the MIT License. See LICENSE for more details ## Acknowledgements This project was built from the [PyTorch Template Project](https://github.com/victoresque/pytorch-template) by Victor Huang. | 2,038 |
eth-sri/eran | ['adversarial attack'] | ['Beta-CROWN: Efficient Bound Propagation with Per-neuron Split Constraints for Complete and Incomplete Neural Network Robustness Verification'] | tf_verify/optimizer.py tf_verify/eran.py tf_verify/config.py tf_verify/spatial/t_2_norm_transformer.py tf_verify/deeppoly_nodes.py tf_verify/spatial/interpolation.py tf_verify/constraint_utils.py testing/check_models.py tf_verify/read_net_file.py tf_verify/read_zonotope_file.py tf_verify/__main__.py tf_verify/spatial/t_1_norm_transformer.py tf_verify/spatial/t_inf_norm_transformer.py data/create_zonotope.py tf_verify/spatial/t_norm_transformer.py tf_verify/refine_activation.py tf_verify/analyzer.py tf_verify/ai_milp.py tf_verify/spatial/__init__.py tf_verify/refine_gpupoly.py tf_verify/onnx_translator.py tf_verify/tensorflow_translator.py tf_verify/eranlayers.py tf_verify/krelu.py tf_verify/deepzono_nodes.py tf_verify/utils.py get_tests normalize get_out_tensors handle_tanh_sigmoid add_spatial_constraints handle_affine _add_kactivation_constraints handle_sign get_bounds_for_layer_with_milp handle_conv create_model verify_network_with_milp Cache milp_callback lp_callback handle_padding obj_from_is_greater_tuple_old sigmoid solver_call handle_residual handle_relu handle_maxpool Analyzer layers config Device clean_string isfloat get_constraints_for_dominant_label label_index get_constraints_from_file DeeppolySubNode DeeppolyGather calc_bounds DeeppolyConv2dNode DeeppolyNode DeeppolyTanhNode DeeppolyConcat DeeppolyInput DeeppolyPoolNode DeeppolyResidualNode DeeppolyFCNode DeeppolyNonlinearity DeeppolyPaddingNode DeeppolyTile DeeppolyLeakyReluNode DeeppolyMulNode DeeppolyReluNode DeeppolySigmoidNode add_input_output_information_deeppoly DeeppolySignNode DeepzonoAffine DeepzonoMatmul DeepzonoNonlinearity DeepzonoTanh add_bounds DeepzonoSigmoid DeepzonoSub DeepzonoInputZonotope add_dimensions DeepzonoConvbias remove_dimensions DeepzonoResadd DeepzonoRelu add_input_output_information DeepzonoAdd DeepzonoPool DeepzonoDuplicate DeepzonoConv DeepzonoMul DeepzonoGather get_xpp DeepzonoInput ERAN eran_input eran_dense eran_conv2d_without_activation eran_maxpool eran_activation eran_resnet_dense tensorshape_to_intlist eran_affine eran_resnet_conv2d eran_reshape eran_conv2d get_ineqs_zono sparse_heuristic_curve generate_linexpr0 KAct make_kactivation_obj sparse_heuristic_with_cutoff encode_kactivation_cons prepare_model ONNXTranslator reshape_nhwc nchw_to_nhwc_index nchw_to_nhwc_shape nchw_to_nhwc onnxshape_to_intlist Optimizer permutation product numel runRepl parseVec read_tensorflow_net extract_mean myConst read_onnx_net extract_std read_zonotope refine_activation_with_solver_bounds refine_gpupoly_results calculate_padding tensorshape_to_intlist TFTranslator identify_var identify_variables check_numeric parse_vnn_lib_prop extract_terms translate_gurobi_status negate_cstr_or_list_old translate_output_constraints translate_input_to_box normalize_plane isnetworkfile acasxu_recursive get_tests denormalize estimate_grads parse_input_box print_progress model_predict init normalize_poly normalize init_domain str2bool show_ascii_spec compute_grid interpolate T1NormTransformer T2NormTransformer TInfNormTransformer TNormTransformer str read format reader open range len MIP_OBJBST cbGet terminate MIP_OBJBND print cbGet terminate SPX_OBJVAL addTerms str addConstant addVar LinExpr addConstr EQUAL append prod range len str addVar LinExpr addConstr EQUAL append prod range len str int GREATER_EQUAL LESS_EQUAL addVar LinExpr addConstr addGenConstrIndicator EQUAL append float prod range len addTerms str addConstant addVar LinExpr addConstr EQUAL append range len str addConstant addVar LinExpr addConstr EQUAL append range len addTerms cons addConstant LinExpr addConstr varsid enumerate len str sorted list max GREATER_EQUAL LESS_EQUAL addVar ones _add_kactivation_constraints addConstr addGenConstrIndicator EQUAL append array range enumerate len str GREATER_EQUAL LESS_EQUAL addVar addConstr addGenConstrIndicator EQUAL append range len str addVar _add_kactivation_constraints append range len addConstant handle_tanh_sigmoid addVar setParam abs addTerms str handle_affine LinExpr set_printoptions FeasibilityTol Model handle_sign append range handle_conv EQUAL addConstr handle_padding handle_residual handle_relu handle_maxpool len max optimize LinExpr min copy setObjective reset MINIMIZE MAXIMIZE flush update sorted create_model zip print Threads reset TimeLimit numproc setParam append range enumerate len get list addVar addConstr tolist add_norm_constraints zip append range enumerate add_spatial_constraints warn objval INFINITY setParam getConstrByName LESS_EQUAL FeasibilityTol TimeLimit append range create_model numlayer MINIMIZE enumerate time optimize remove print Cutoff addConstr obj_from_is_greater_tuple_old setObjective reset len LinExpr CPU float cpu_count append float int clean_string readlines len label_index startswith append range split calc_layerno elina_interval_array_free get_num_neurons_in_layer box_for_layer append reduce elina_dimchange_alloc elina_dimchange_free elina_abstract0_add_dimensions elina_dimchange_init range elina_dimchange_alloc elina_dimchange_free elina_dimchange_init range elina_abstract0_remove_dimensions reduce elina_interval_array_free realdim elina_abstract0_to_box append intdim elina_abstract0_dimension Variable reshape reduce matmul tensorshape_to_intlist shape Variable int conv2d sigmoid tanh relu append append eran_affine eran_conv2d_without_activation int eran_conv2d_without_activation add eran_conv2d range eran_dense reduce add tensorshape_to_intlist shape eran_affine eran_reshape range ElinaDim coeff cst pointer elina_scalar_set_double ELINA_LINEXPR_SPARSE elina_linexpr0_alloc zip enumerate scalar len element all offset product length dbl man elina_abstract0_assign_linexpr_array generate_linexpr0 tdim elina_abstract0_bound_dimension append sorted generate_sparse_cover sparse_n tuple print min append len generate_sparse_cover sparse_n tuple print min append len pop sparse_heuristic_curve ElinaDim asarray time all product get_upper_bound_for_linexpr0 print debug remove_dimensions generate_linexpr0 elina_linexpr0_array_alloc sparse_heuristic_with_cutoff append add_dimensions len list reshape transpose prod len initializer floor attribute list name multiply node add shape i nchw_to_nhwc_index input append expand_dims ceil prod range onnxshape_to_intlist concatenate ints insert subtract copy take item nchw_to_nhwc_shape float s int divide zeros array len replace float64 search group split zeros range len float64 search group split zeros range len zeros range leaky_relu float64 concat sign myConst bias_add open transpose placeholder matmul add runRepl conv2d readline relu extract_mean extract_std tanh print reshape max_pool sigmoid parseVec load node check_model int read array split abstract_information calc_layerno sigmoid_zono_layerwise handle_relu_layer specLB specUB append tanh_zono_layerwise range get_bounds_for_layer_with_milp handle_sigmoid_layer debug numlayer encode_kactivation_cons handle_tanh_layer update_bounds_for_neuron time print relu_zono get_num_neurons_in_layer relu_zono_layerwise relu_zono_refined sparse_heuristic_with_cutoff setParam max all LinExpr FeasibilityTol specLB _nn TimeLimit specUB append ceil range get_bounds_for_layer_with_milp evalAffineExpr product create_model concatenate debug numlayer test MINIMIZE getOutputSize int time optimize print Cutoff write setObjective reset zeros array len int max int match group match identify_var check_numeric append startswith int group match startswith append identify_variables concatenate endswith group match extract_terms startswith zeros append range nonzero minimum concatenate ones maximum matmul append array append splitext product replace findall append split print range zeros clone range len zeros range len run prepare range model_predict zeros ConfigProto sum array Session len debug write argmax timeout_lp use_default_heuristic verify_network_with_milp negate_cstr_or_list_old timeout_milp print back_propagate_gradient analyze_box print_progress set_last_weights subset mul add shape unsqueeze repeat device to double unsqueeze arange | ERAN <img width="100" alt="portfolio_view" align="right" src="http://safeai.ethz.ch/img/sri-logo.svg"> ========  ETH Robustness Analyzer for Neural Networks (ERAN) is a state-of-the-art sound, precise, scalable, and extensible analyzer based on [abstract interpretation](https://en.wikipedia.org/wiki/Abstract_interpretation) for the complete and incomplete verification of MNIST, CIFAR10, and ACAS Xu based networks. ERAN produces state-of-the-art precision and performance for both complete and incomplete verification and can be tuned to provide best precision and scalability (see recommended configuration settings at the bottom). ERAN is developed at the [SRI Lab, Department of Computer Science, ETH Zurich](https://www.sri.inf.ethz.ch/) as part of the [Safe AI project](http://safeai.ethz.ch/). The goal of ERAN is to automatically verify safety properties of neural networks with feedforward, convolutional, and residual layers against input perturbations (e.g., L∞-norm attacks, geometric transformations, vector field deformations, etc). ERAN combines abstract domains with custom multi-neuron relaxations from PRIMA to support fully-connected, convolutional, and residual networks with ReLU, Sigmoid, Tanh, and Maxpool activations. ERAN is sound under floating point arithmetic with the exception of the (MI)LP solver used in RefineZono and RefinePoly. The employed abstract domains are specifically designed for the setting of neural networks and aim to balance scalability and precision. Specifically, ERAN supports the following analysis: * DeepZ [NIPS'18]: contains specialized abstract Zonotope transformers for handling ReLU, Sigmoid and Tanh activation functions. * DeepPoly [POPL'19]: based on a domain that combines floating point Polyhedra with Intervals. * GPUPoly [MLSys'2021]: leverages an efficient GPU implementation to scale DeepPoly to much larger networks. * RefineZono [ICLR'19]: combines DeepZ analysis with MILP and LP solvers for more precision. * RefinePoly/RefineGPUPoly [NeurIPS'19]: combines DeepPoly/GPUPoly analysis with (MI)LP refinement and PRIMA framework [arXiv'2021] to compute group-wise joint neuron abstractions for state-of-the-art precision and scalability. | 2,039 |
eth-sri/transformation-smoothing | ['adversarial defense'] | ['Certified Defense to Image Transformations via Randomized Smoothing'] | classify_utils.py mnist_net.py smoothing-adversarial/code/core.py smoothing-adversarial/code/analyze.py classify_ksmooth.py distSPTD_certify_attacked.py util.py classify_f.py audio/distributional.py smoothing-adversarial/code/archs/cifar_resnet.py indivSPT_certify_attacked.py audio/inverse_online_error.py train.py getErrSampling.py smoothing-adversarial/code/visualize.py data.py smoothing-adversarial/code/attacks.py transformations.py ds/GTSRB/test/to_folder.py audio/transforms.py smoothing-adversarial/code/predict.py geometrictools/setup.py smoothing-adversarial/code/certify.py distSPTD_certify.py smoothing-adversarial/code/architectures.py verifyE.py smoothing-adversarial/code/datasets.py smoothing-adversarial/code/generate_github_result.py smoothing-adversarial/code/train_pgd.py geometrictools/utils/try.py smoothing-adversarial/code/zipdata.py smoothing-adversarial/code/train_utils.py distSPTx_certify.py classifyHeuristicRadius.py cifar_resnet.py audio/train.py training_attack.py audio/sample_error.py eval.py refinementExp.py distSPT_predict.py classifyHeuristic.py resnet.py getErr.py smoothing-adversarial/code/train.py ResNet Bottleneck conv3x3 resnet BasicBlock getRotations pre worst_of_k getTranslations smooth translate rotate sample classify smooth_pred l2_sample sample_transformation l2_smooth pre setup_args_ksmooth test_predict l2_predict ksmooth VingetteModule run_model kpredict setup_args setup get_basemodel setup_args_preprocessing RobustnessModelInnerWrapper setup_args_getE get_data get_logger VingetteModule NormalizeLayer ImagenetSizeDataset ImagenetSizeLoader get_dataset getRotations getTranslations attack rotate_ translate_ get_err compute_err sample_err _sample_err gen sample_transformation pre parseRefinementExp parseClassifyHeuristicRadius parseClassifyOnline parseGetErr parseClassifyDistributional parseClassifyHeuristic parseClassifyDistributionalJustPrediction _sample_err sample_err run sample_transformation gen attack pre MNISTConvNet conv1x1 resnext50_32x4d wide_resnet50_2 ResNet resnet50 resnext101_32x8d Bottleneck resnet152 wide_resnet101_2 conv3x3 _resnet resnet34 resnet18 BasicBlock resnet101 GaussianL2Step GaussianNoise Filter T Id Blur get_vingette_mask Vingette filterNP resize_crop lpfilter translate rotate get_postprocessing LP getGaussianFilter Wavelet Round get_interpolation str2FloatOrNone torch_image_to_PIL Logger str2bool split sample_transformation to_mfcc Loader RobustnessModelOuterWrapper pre RobustnessModelInnerWrapper run round get_dataset transform spect_tranform run get_dataset CustomGCommandLoader Wrapper ID scale_volume round shift print_nice Filter get_vingette_mask VingetteModule getGaussianFilter get_architecture Attacker DDN PGD_L2 Smooth Filter Vingette get_normalize_layer _imagenet get_vingette_mask _imagenet32 TiTop50KDataset get_num_classes get_input_center_layer _imagenet_on_philly _restricted_imagenet MultiDatasetsDataLoader get_dataset _cifar10 getGaussianFilter ImageNetDS NormalizeLayer InputCenterLayer main train test init_logfile AverageMeter accuracy requires_grad_ copy_code log ZipData ResNet Bottleneck conv3x3 resnet BasicBlock center_crop resize tolist gamma1 getRotations model getTranslations uniform device full to attack_k gamma0 to device normal getRotations print getTranslations device to sample ppf n_gamma most_common batch_size min randn_like tolist extend repeat l2_sample ppf sample_transformation l2_smooth ppf pre Sequential randn_like extend tolist alpha_eps n_eps sigma_eps fl device append trange to to_tensor n_gamma most_common l2_sample sample_transformation alphaI pre Sequential l2_predict Counter fl device append trange to to_tensor n_gamma to fl device add_argument normal sigma_gamma rotate translate center_crop_post_transform resize_post_transform add_argument add_argument add_argument model RobustnessModelInnerWrapper Sequential MNISTConvNet DataParallel VingetteModule device tensor dataset get_architecture ds_class AttackerModel exit make_and_restore_model load_state_dict resnet18 append to module NormalizeLayer normalizer Namespace glob radiusDecrease eval enumerate load join use_cuda isinstance print seed debug set_device device manual_seed gamma stdout join replace print strftime Logger makedirs list min get_dataset sample N range len join RestrictedImageNet gtsrb_path abspath gamma1 getRotations getTranslations uniform run_model full attack_k gamma0 max V norm pre reshape interpolation rotate translate get_postprocessing append expand_dims array range normal center_crop_post_transform Vingette list gamma1 nrBetas resize_post_transform threads shape uniform append sigma_gamma range gamma0 split normal nrBetas print getE center_crop astype extend sigma_gamma resize expand_dims array guess_E_add compute_err guess_E_mult sample_err sum max sampleE range print read_csv print read_csv print to_numeric read_csv print to_numeric read_csv print max read_csv set_visible list tolist ylabel bar savefig gca range replace set_xticklabels tight_layout pivot_table set_label_coords enumerate xlabel text index set_xticks set_facecolor figure array read_csv len _sample_err seed zip resize close sample_err send reshape shape reshape shape append range ResNet exp pi zeros sum range convolve shape getGaussianFilter zeros range tolist ndarray isinstance imresize_single_image fft2 shape real zeros ifft2 range int ellipse Draw ones new astype float32 tile expand_dims eval compose isinstance divmod len transpose numpy astype uint8 int window_stride FloatTensor stft resize hstack window_size magphase log1p zeros to_mfcc make_dataset find_classes load BytesIO write seek minimum sum print min tolist maximum sqrt alpha_min_max logical_or floor linspace ceil round max append enumerate stft int magphase log1p t spect_tranform load normal norm itemgetter extend get_dataset uniform lmap K beta_std transform ravel range round round print around real get_normalize_layer resnet50 to cuda Linear join join join Compose join Compose RestrictedImageNet join init_logfile outdir SGD DataLoader save arch dataset get_architecture cuda log StepLR get_dataset epochs noise_sd range format test join time parameters train step gpu makedirs update time format criterion model backward print size AverageMeter randn_like accuracy zero_grad item step cuda enumerate len eval time AverageMeter close write open close write open parameters join format print set mkdir copy2 walk | Certified Defense to Image Transformations via Randomized Smoothing <a href="https://www.sri.inf.ethz.ch/"><img width="100" alt="portfolio_view" align="right" src="http://safeai.ethz.ch/img/sri-logo.svg"></a> ============================================================================================================= We present extend an extension of randomized smoothing to cover parametrized transformations (here image rotations and translations) and certify robustness in the parameter space (e.g., rotation angle). Here we evaluate three defense with different guarantees based this. For further details, please see [our NeurIPS 2020 paper](https://www.sri.inf.ethz.ch/publications/fischer2020smoothing).  ## Note | 2,040 |
ethz-asl/modular_semantic_segmentation | ['semantic segmentation'] | ['Modular Sensor Fusion for Semantic Segmentation'] | xview/models/uncertainty_dirichlet_mix.py xview/models/simple_fcn.py xview/datasets/cityscapes.py xview/datasets/mixed_data.py xview/datasets/pascalvoc.py xview/models/test_mix_fcn.py xview/models/base_model.py experiments/dirichlet_fusion.py experiments/ibcc_fusion.py xview/models/bayes_mix.py experiments/uncertainty_eval.py experiments/finetuning.py xview/models/test_adapnet.py experiments/bayes_fusion.py xview/datasets/raw_synthia.py xview/datasets/data_baseclass.py xview/models/average_mix.py experiments/rerun.py xview/datasets/cityscapesB.py experiments/timing.py xview/datasets/wrapper.py xview/models/dirichletEstimation.py xview/models/test_fcn.py xview/models/vgg16.py xview/models/custom_layers.py xview/models/dirichlet_mix.py xview/models/basic_fusion_model.py xview/datasets/synthia_rand.py xview/models/test_simple_fcn.py experiments/utils.py xview/datasets/not_cityscapes.py xview/datasets/test_synthia.py experiments/evaluation.py xview/models/bayesian_fcn.py xview/datasets/__init__.py xview/models/utils.py xview/models/fusion_fcn.py xview/models/__init__.py xview/models/variance_mix.py xview/datasets/toydata.py xview/models/adapnet.py xview/models/dirichlet_fastfit.py xview/models/test_progressive_fcn.py setup.py xview/datasets/cityscapesA.py experiments/training.py xview/datasets/augmentation.py experiments/train_and_evaluate_progressive.py xview/datasets/synthia_cityscapes.py xview/models/dirichletDifferentiation.py xview/datasets/synthia.py experiments/different_evaluation_parameters.py evaluate split_test_data average fit_and_evaluate collect_data main grid_search parameter_combinations fit_and_evaluate test_parameters all_synthia evaluate import_weights_into_network also_load_config main evaluate_on_all_synthia_seqs depth_to_rgb rgb_to_depth train get_all_sequence_validation_sets collect_data rerun time_variance_fcn time_bayes_fcn time_fusion_fcn time_bayes_lookup_fcn time_dirichlet_fcn time_depth_fcn time_rgb_fcn time_average_fcn train_and_evaluate main create_directories train_network depth_to_rgb main rgb_to_depth measure_metrics evaluate_temperature_scaling evaluate_uncertainty uncertainty_benchmark measure uncertainty_parameter_search train_ambiguous reverse_convert_datatypes ExperimentData get_observer load_data flip_labels crop_multiple rotate_image crop_around_center augmentate largest_rotated_rect Cityscapes Cityscapes Cityscapes DataBaseclass MixedData AddRandomObjects get_dataset PascalVOC Synthia one_channel_image_reader Synthia crop_resized_image SynthiaCityscapes SynthiaRand test_preprocessing_produces_all_ouputs test_can_open_data test_can_find_data ToyData DataWrapper get_dataset block_a Adapnet adapnet block_b AverageFusion BaseModel transform_inputdata with_graph test_pipeline FusionModel BayesianFCN sampling_uncertainty bayes_fusion bayes_decision_matrix BayesFusion bilinear_filter_initializer entropy log_softmax unpool_2d adap_conv deconv2d conv2d softmax Selection predictStepLogSpace getPredictedStep logProbForMultinomials trigamma sqVectorSize findDirichletPriorsFromMultinomials getPredictedStepAlt getTotalLoss priorHessianConst digamma testTrialPriors priorHessianDiag getGradientForMultinomials findDirichletPriors getSufficientStatistic predictStepUsingHessian predictStepLogSpace getPredictedStep logProbForMultinomials trigamma sqVectorSize findDirichletPriorsFromMultinomials getPredictedStepAlt getTotalLoss priorHessianConst digamma testTrialPriors priorHessianDiag getGradientForMultinomials findDirichletPriors getSufficientStatistic predictStepUsingHessian _ipsi loglikelihood_logp meanprecision_with_sufficient_statistic _piecewise _trigamma pdf test fixedpoint_with_sufficient_statistic _init_a loglikelihood _fixedpoint _meanprecision mle _fit_m _fit_s meanprecision DirichletFusion dirichlet_fusion FusionFCN fusion_fcn fcn encoder decoder SimpleFCN test_can_run_training test_can_build_model test_can_run_training test_can_build_model test_can_export_and_import_weights test_can_build_model test_can_fit_model test_can_run_training test_can_build_model test_can_run_training test_can_build_model UncertaintyMix dirichlet_uncertainty_fusion define_scope cross_entropy doublewrap VarianceFusion variance_fusion progressive_vgg16 vgg16 get_model train_test_split testset join format ExperimentData split_test_data mkdir savez_compressed format print get_dataset get_model flush print format flush get_dataset deepcopy format setdefault print get_dataset get_model flush append deepcopy list items parameter_combinations append_deep_value tqdm evaluation append grid_search get_model list parameter_combinations measureset split_test_data dict zip train_test_split split_test_data get_testset score labelinfo __name__ deepcopy format evaluate print flush import_weights_from_description items list isinstance update get_model ExperimentData print get_model deepcopy get_validation_data load_data load ExperimentData create_directories savez get_artifact tile next _id load ExperimentData create_directories savez mean get_artifact next _id create_directories _id deepcopy get_dataset get_model pop get_record print command getattr flush flush time format ones print mean fusion_fcn append global_variables_initializer range std Session local_variables_initializer run flush get_record time format ones print range mean softmax append global_variables_initializer argmax std Session local_variables_initializer run get_record argmax Session run bayes_decision_matrix ones to_int64 multiply reduce_sum append expand_dims range format one_hot mean softmax local_variables_initializer flush time constant print global_variables_initializer std get_record argmax dirichlet_fusion Session values run list ones append range format Dirichlet astype mean softmax local_variables_initializer flush time print global_variables_initializer std flush get_record time format ones print range mean variance_fusion test_pipeline append global_variables_initializer argmax std Session local_variables_initializer run flush time format ones print range mean reduce_mean stack softmax append global_variables_initializer argmax std Session local_variables_initializer run flush time format ones print range mean softmax append global_variables_initializer argmax std Session local_variables_initializer run flush time format ones print range mean softmax append global_variables_initializer argmax std Session local_variables_initializer run format rmtree mkdir append exists makedirs add_artifact join get_trainset export_weights listdir import_weights_into_network fit evaluate train_network create_directories _id get_dataset format name print out_of_distribution_detection_score __name__ flush misclassification_detection_score prob_distribution nll_score zeros value_distribution mean_diff list ExperimentData force_bson_encodeable grid_search get_dataset update_record get_model get_data_description grid_search get_model get_dataset update list create_directories isinstance print get_dataset choice get_model range get_data_description _id enumerate list get_model get_data_description get_dataset get_model get_dataset isinstance max int warpAffine tuple min vstack matrix abs array int tan cos pi floor sin int flip_labels crop_around_center resize max list Add rotate_image uniform augment_image radians astype Affine choice ContrastNormalization Sometimes float flip keys print min randint largest_rotated_rect vstack map asDirect Reader Synthia Synthia print Synthia format next fcn argmax softmax adapnet reduce_mean range nan_to_num append gather float sum log len argmax list product range nan_to_num zeros float sum array log len ceil zeros abs range _deconv2d batch_normalization activation isinstance get update activation batch_normalization range len sum gammaln sum digamma range len range len sum range len priorHessianConst priorHessianDiag priorHessianConst priorHessianDiag range len predictStepLogSpace exp sqVectorSize print len getTotalLoss testTrialPriors getGradientForMultinomials range predictStepUsingHessian getSufficientStatistic shape mle vstack loglikelihood sum exp gammaln sum mean shape _ipsi mean shape psi _init_a sum range _fit_m ones range mean shape _init_a sum _fit_s len _ipsi psi sum range _fit_m ones range shape sum _fit_s len exp _trigamma psi sum range _ipsi sum range dtype squeeze shape item append zeros asanyarray range len mean _piecewise _trigamma psi asanyarray range reduce_sum stack append range len update deepcopy list items decoder concat conv2d deconv2d add_n vgg16 deepcopy update encoder decoder close Adapnet Synthia FCN Synthia int ones reshape reduce_max astype reduce_sum reduce_mean stack eye append expand_dims range len __name__ stack conv2d max_pooling2d update deepcopy max_pooling2d adap_conv conv2d | Corresponding implementations for the IROS 2018 paper ["Modular Sensor Fusion for Semantic Segmentation"](https://arxiv.org/abs/1807.11249) by Hermann Blum, Abel Gawel, Roland Siegwart and Cesar Cadena. <table> <tr> <td>Fusion with Synthia RAND</td> <td>Fusion with Cityscapes</td> </tr> <tr> <td> <img height="400px" src="https://github.com/ethz-asl/modular_semantic_segmentation/raw/publish/synthia.png"> </td> | 2,041 |
ethz-asl/segmap | ['data compression', 'autonomous driving'] | ['SegMap: Segment-based mapping and localization using data-driven descriptors', 'SegMap: 3D Segment Mapping using Data-Driven Descriptors'] | segmappy/bin/plot_train_progress.py tools/remap_bag.py segmappy/segmappy/tools/roccurve.py segmappy/bin/create_matches_from_run.py segmappy/bin/plot_loss.py segmappy/bin/make_labels.py segmappy/bin/ensure_segmappy_is_installed.py tools/cut.py tools/add_header_time_offset.py segmappy/segmappy/__main__.py tools/remove_tf_from_bag.py segmappy/segmappy/tools/classifiertools.py segmappy/segmappy/core/generator.py segmappy/segmappy/core/preprocessor.py segmappy/bin/plot_reconstructions.py segmappy/bin/verify_matches_extracted_from_run.py segmappy/bin/visualise.py segmappy/segmappy/core/dataset.py segmappy/segmappy/__init__.py segmappy/segmappy/models/model_groups_tf.py segmappy/bin/analyse_database.py segmappy/segmappy/tools/hull.py segmappy/setup.py segmappy/segmappy/tools/import_export.py segmappy/segmappy/tools/voxeltools.py segmappy/segmappy/core/config.py print_class_info get_class Config get_config_dir get_segmap_home_dir get_default_model_dir get_default_dataset_dir Dataset Generator GeneratorFeatures Preprocessor init_model get_default_dataset to_onehot visualize_side_by_side get_default_preprocessor visualize point_in_hull n_points_in_hull are_in_hull load_trajectory write_matches load_segments load_segments_no_duplicates load_playground_segments load_merges load_positions load_classes write_segments load_matches load_features write_features load_labels convert_strings_to_floats_in_list_of_lists write_classes write_list_of_lists load_list_of_lists get_roc_curve get_roc_pairs voxel_grid_to_cloud_by_n_points compute_jaccard_index voxel_grid_to_cloud compute_accuracies keep_voxels_above_threshold keep_n_most_probable_voxels compute_jaccard_indices_on_voxels_lists compute_accuracies_on_voxel_lists voxel_grid_to_cloud_by_probability add_offset cut remove_tf print_class_info get_class Config get_config_dir get_segmap_home_dir get_default_model_dir get_default_dataset_dir Dataset Generator GeneratorFeatures Preprocessor init_model get_default_dataset to_onehot visualize_side_by_side get_default_preprocessor visualize point_in_hull n_points_in_hull are_in_hull load_trajectory write_matches load_segments load_segments_no_duplicates load_playground_segments load_merges load_positions load_classes write_segments load_matches load_features write_features load_labels convert_strings_to_floats_in_list_of_lists write_classes write_list_of_lists load_list_of_lists get_roc_curve get_roc_pairs voxel_grid_to_cloud_by_n_points compute_jaccard_index voxel_grid_to_cloud compute_accuracies keep_voxels_above_threshold keep_n_most_probable_voxels compute_jaccard_indices_on_voxels_lists compute_accuracies_on_voxel_lists voxel_grid_to_cloud_by_probability add_offset cut remove_tf print unique join copy mkdir abspath dirname expanduser conv3d concat flatten assign argmax softmax_cross_entropy_with_logits_v2 get_collection identity placeholder add batch_normalization cast max_pooling3d conv3d_transpose dropout softmax equal dense constant Variable reshape float32 AdamOptimizer reduce_mean UPDATE_OPS placeholder_with_default add_subplot clf get_zmajorticklabels ion max str scatter setp input get_xmajorticklabels get_ymajorticklabels set_xlim set_zticks close eval set_zlim flush_events enumerate minimum ioff set_yticks min write maximum set_xticks figure array set_ylim zeros enumerate grid add_subplot clf get_zmajorticklabels max set_aspect scatter setp append get_xmajorticklabels input range get_ymajorticklabels set_xticklabels set_xlim set_zticks eval set_zlim enumerate minimum int print set_yticks min set_axis_off maximum subplots_adjust draw pause set_xticks set_facecolor figure jet array set_ylim len Dataset Preprocessor point_in_hull range point_in_hull range append join list str print size len astype split zip append values enumerate join str print len astype split append values enumerate join str print isfile len join str print isfile len join str print isfile len join str print isfile len print join len str join load_list_of_lists print str values join str print write_list_of_lists len join str print write_list_of_lists len join join str print write_list_of_lists len append is_integer float int str print loadtxt append listdir range len seed list norm print unique append array range enumerate len norm roc_curve append range auc append range append print range len shape append zeros range len zeros range shape append range float sum transforms read_messages write close Bag loginfo _has_header from_sec stamp read_messages write close Bag loginfo from_sec print close | ## *SegMap* SegMap is a map representation based on 3D segments allowing for robot localization, environment reconstruction, and semantics extraction. The SegMap code is open-source (BSD License) and has been tested under Ubuntu 14.04, 16.04 and ROS Indigo, Kinetic. Please remember that this is on-going research code which is subject to changes in the future. ### Related Publications R. Dubé, A. Cramariuc, D. Dugas, H. Sommer, M. Dymczyk, J. Nieto, R. Siegwart, and C. Cadena. **SegMap: Segment-based mapping and localization using data-driven descriptors** *The International Journal of Robotics Research (IJRR), 2019* [pdf](https://arxiv.org/pdf/1909.12837.pdf) R. Dubé, A. Cramariuc, D. Dugas, J. Nieto, R. Siegwart, and C. Cadena. **"SegMap: 3D Segment Mapping using Data-Driven Descriptors."** *Robotics: Science and Systems (RSS), 2018.* [pdf](http://www.roboticsproceedings.org/rss14/p03.pdf) - [video](https://youtu.be/CMk4w4eRobg) R. Dubé, MG. Gollub, H. Sommer, I. Gilitschenski, R. Siegwart, C. Cadena and , J. Nieto. **"Incremental Segment-Based Localization in 3D Point Clouds."** *IEEE Robotics and Automation Letters, 2018.* [pdf](http://n.ethz.ch/~cesarc/files/RAL2018_rdube.pdf) R. Dubé, D. Dugas, E. Stumm, J. Nieto, R. Siegwart, and C. Cadena. **"SegMatch: Segment Based Place Recognition in 3D Point Clouds."** *IEEE International Conference on Robotics and Automation, 2017.* [pdf](https://arxiv.org/pdf/1609.07720.pdf) - [video](https://youtu.be/iddCgYbgpjE) ## Features - 3D CNN encoder – decoder - ICP based LiDAR odometry | 2,042 |
ethz-asl/segmatch | ['autonomous driving'] | ['SegMap: Segment-based mapping and localization using data-driven descriptors'] | segmappy/bin/plot_train_progress.py tools/remap_bag.py segmappy/segmappy/tools/roccurve.py segmappy/bin/create_matches_from_run.py segmappy/bin/plot_loss.py segmappy/bin/make_labels.py segmappy/bin/ensure_segmappy_is_installed.py tools/cut.py tools/add_header_time_offset.py segmappy/segmappy/__main__.py tools/remove_tf_from_bag.py segmappy/segmappy/tools/classifiertools.py segmappy/segmappy/core/generator.py segmappy/segmappy/core/preprocessor.py segmappy/bin/plot_reconstructions.py segmappy/bin/verify_matches_extracted_from_run.py segmappy/bin/visualise.py segmappy/segmappy/core/dataset.py segmappy/segmappy/__init__.py segmappy/segmappy/models/model_groups_tf.py segmappy/bin/analyse_database.py segmappy/segmappy/tools/hull.py segmappy/setup.py segmappy/segmappy/tools/import_export.py segmappy/segmappy/tools/voxeltools.py segmappy/segmappy/core/config.py print_class_info get_class Config get_config_dir get_segmap_home_dir get_default_model_dir get_default_dataset_dir Dataset Generator GeneratorFeatures Preprocessor init_model get_default_dataset to_onehot visualize_side_by_side get_default_preprocessor visualize point_in_hull n_points_in_hull are_in_hull load_trajectory write_matches load_segments load_segments_no_duplicates load_playground_segments load_merges load_positions load_classes write_segments load_matches load_features write_features load_labels convert_strings_to_floats_in_list_of_lists write_classes write_list_of_lists load_list_of_lists get_roc_curve get_roc_pairs voxel_grid_to_cloud_by_n_points compute_jaccard_index voxel_grid_to_cloud compute_accuracies keep_voxels_above_threshold keep_n_most_probable_voxels compute_jaccard_indices_on_voxels_lists compute_accuracies_on_voxel_lists voxel_grid_to_cloud_by_probability add_offset cut remove_tf print unique join copy mkdir abspath dirname expanduser conv3d concat flatten assign argmax softmax_cross_entropy_with_logits_v2 get_collection identity placeholder add batch_normalization cast max_pooling3d conv3d_transpose dropout softmax equal dense constant Variable reshape float32 AdamOptimizer reduce_mean UPDATE_OPS placeholder_with_default add_subplot clf get_zmajorticklabels ion max str scatter setp input get_xmajorticklabels get_ymajorticklabels set_xlim set_zticks close eval set_zlim flush_events enumerate minimum ioff set_yticks min write maximum set_xticks figure array set_ylim zeros enumerate grid add_subplot clf get_zmajorticklabels max set_aspect scatter setp append get_xmajorticklabels input range get_ymajorticklabels set_xticklabels set_xlim set_zticks eval set_zlim enumerate minimum int print set_yticks min set_axis_off maximum subplots_adjust draw pause set_xticks set_facecolor figure jet array set_ylim len Dataset Preprocessor point_in_hull range point_in_hull range append join list str print size len astype split zip append values enumerate join str print len astype split append values enumerate join str print isfile len join str print isfile len join str print isfile len join str print isfile len print join len str join load_list_of_lists print str values join str print write_list_of_lists len join str print write_list_of_lists len join join str print write_list_of_lists len append is_integer float int str print loadtxt append listdir range len seed list norm print unique append array range enumerate len norm roc_curve append range auc append range append print range len shape append zeros range len zeros range shape append range float sum transforms read_messages write close Bag loginfo _has_header from_sec stamp read_messages write close Bag loginfo from_sec print close | ## *SegMap* SegMap is a map representation based on 3D segments allowing for robot localization, environment reconstruction, and semantics extraction. The SegMap code is open-source (BSD License) and has been tested under Ubuntu 14.04, 16.04 and ROS Indigo, Kinetic. Please remember that this is on-going research code which is subject to changes in the future. ### Related Publications R. Dubé, A. Cramariuc, D. Dugas, H. Sommer, M. Dymczyk, J. Nieto, R. Siegwart, and C. Cadena. **SegMap: Segment-based mapping and localization using data-driven descriptors** *The International Journal of Robotics Research (IJRR), 2019* [pdf](https://arxiv.org/pdf/1909.12837.pdf) R. Dubé, A. Cramariuc, D. Dugas, J. Nieto, R. Siegwart, and C. Cadena. **"SegMap: 3D Segment Mapping using Data-Driven Descriptors."** *Robotics: Science and Systems (RSS), 2018.* [pdf](http://www.roboticsproceedings.org/rss14/p03.pdf) - [video](https://youtu.be/CMk4w4eRobg) R. Dubé, MG. Gollub, H. Sommer, I. Gilitschenski, R. Siegwart, C. Cadena and , J. Nieto. **"Incremental Segment-Based Localization in 3D Point Clouds."** *IEEE Robotics and Automation Letters, 2018.* [pdf](http://n.ethz.ch/~cesarc/files/RAL2018_rdube.pdf) R. Dubé, D. Dugas, E. Stumm, J. Nieto, R. Siegwart, and C. Cadena. **"SegMatch: Segment Based Place Recognition in 3D Point Clouds."** *IEEE International Conference on Robotics and Automation, 2017.* [pdf](https://arxiv.org/pdf/1609.07720.pdf) - [video](https://youtu.be/iddCgYbgpjE) ## Features - 3D CNN encoder – decoder - ICP based LiDAR odometry | 2,043 |
etjoa003/gpt | ['explainable artificial intelligence', 'semantic segmentation'] | ['Convolutional Neural Network Interpretability with General Pattern Theory'] | src/gpt_genmod/model.py src/gpt_genmod/evaluation_modules/pytorch_fid/fid_score.py src/APL_to_python/simple_algo_utils.py main_gpt_genmod.py src/gpt_mnist/finetuner/finetune_Simple0001.py src/gpt_mnist/finetuner/finetune_NSCC_NICE1.py src/utils/utils.py src/gpt_mnist/data.py src/gpt_mnist/pipeline.py main_gpt_mnist_eval.py src/gpt_mnist/evaluation.py src/gpt_genmod/pipeline.py src/gpt_mnist/finetuner/finetune.py src/APL_to_python/simple_algo_section4.2.2.py src/gpt_genmod/evaluation_modules/pytorch_fid/__init__.py src/gpt_mnist/sampler.py src/APL_to_python/simple_algo_section4.2.3.py src/gpt_genmod/evaluation_modules/pytorch_msssim/__init__.py src/gpt_genmod/utils.py src/APL_to_python/simple_algo_utils.backup.py src/gpt_genmod/printing_manager.py src/gpt_genmod/data.py main_gpt_mnist.py src/gpt_genmod/evaluation_modules/pytorch_fid/__main__.py src/gpt_mnist/utils.py src/gpt_mnist/model.py src/gpt_mnist/finetuner/finetune_NSCC_NICE2.py src/gpt_genmod/evaluation_modules/pytorch_msssim/ssim.py src/gpt_genmod/evaluation_modules/pytorch_fid/inception.py compenv develop print_in_alphabet chararray_letterwise_print opposite growth1 replace_index_array_with_values_from setEXTG0 decode_generator setBSG code_generator compnonzero growth1 code_generator move1 develop chararray_letterwise_print opposite tau1 replicate growth3 setBSG compnonzero print_in_alphabet test growth2 index_col_and_row_by_list compenv newg2 split1 split2 replace_index_array_with_values_from setEXTG0 decode_generator changeI prepare_celebaloader CelebADataset CIFAR10Dataset prepare_cifarloader bytes_to_ndarray PatternEncode PatternGrowth get_one_funnel_conv SimpleGrowth Reconstructor manage_directories2 compute_ssim prepare_data_for_FID plot_results training Trainer save_loss_image entry manage_dir_for_FID print_settings print_result_dict sample save_reconstructed_images save_sixteen_images set_figure_setting ShortPrint PrintingManager print_shape_min_max FastPickleClient count_parameters parse_bool_from_string numpy_batch_CHW_resize average_every_n manage_directories calculate_activation_statistics ImagesPathDataset get_activations main calculate_fid_given_paths calculate_frechet_distance _compute_statistics_of_path FIDInceptionE_1 InceptionV3 _inception_v3 FIDInceptionE_2 fid_inception_v3 FIDInceptionC FIDInceptionA MS_SSIM _ssim ms_ssim SSIM ssim _fspecial_gauss_1d gaussian_filter CustomBoardPrinter SimpleGen chararray_letterwise_print opposite replace_index_array_with_values_from decode_generator code_generator setup_xai_method compute_correlations compute_one_sample_basic_metrics eval_sanity_weight_randomization filter_spearman_values view_AOPC get_cascade_randomized_net randomizing_one_resnet_basic_block setup_AOPC_dicts eval_AOPC setup_xai_methods_with_cascaded_randomization compute_correlationsGC entry compute_one_sample_AOPC_for_y_pred pad_dict_for_dataframe compute_AOPC_for_y_pred eval_basic_metrics get_uniform_random_initialization prepare_results_dictionary get_cascade_randomized_net_for_GC randomizing_one_resnet_sequential_of_basic_blocks view_sanity_weight_randomization ExpansionBlock ConvPipe ResGPTNet34 compute_attr_non_emtpy_gens compute_attr_one_pixel_target compute_attr_transformations_of_non_emtpy_gens generate_generator_dist view_losses evaluation_visual_results gen_distribution training manually_parse_boolean plot_setting plot_setting2 entry generate_samples heatmaps_of_gen_config apply_to_mnist arrange_heatmaps arrange_heatmapsGC WrapperNet heatmaps Pytorch_GPT_MNIST_Sampler GPT_MNIST_Sampler save_ckpt_dir folder_check Logger redirect LossMonitor setup_main_components do_finetune do_finetune save_samples_for_display_and_adjustments LossMonitor setup_main_components do_finetune ShortPrint Logger FastPickleClient PrintingManager floor list astype roll zeros array range zeros range zeros range roll str compenv print print_in_alphabet growth1 range astype replace_index_array_with_values_from chararray_letterwise_print compnonzero list print logical_and astype code_generator sum array range tensordot list astype append array range str chararray reshape astype set print growth2 growth3 growth1 astype int move1 list all split1 index split2 ceil range list ones diag index_col_and_row_by_list astype test get_last copy zeros array range tensordot list replicate array range len ones astype tensordot remainder ones tau1 code_generator tensordot ones newg2 zeros range changeI zeros logical_or astype sign zip meshgrid T DataLoader CIFAR10Dataset demo_some_images BytesIO bytearray open DataLoader CelebADataset Sequential str print plot_results training sample print Reconstructor print run_training print_settings Trainer join clip transpose add_subplot tight_layout range close imshow eval savefig figure zeros numpy net set_figure_setting set_yticks set_xticks set_title getcwd join mkdir getcwd join mkdir manage_directories print items close savefig figure plot save tensor __getitem__ fromarray str getcwd manage_dir_for_FID CIFAR10Dataset range concatenate astype shuffle CelebADataset type join uint8 isinstance print __len__ numpy array ms_ssim_module MS_SSIM ssim_module SSIM print str items set_title add_subplot close imshow savefig figure range set_figure_setting show items defaultdict plot print set_xlabel init_or_load_model total_iter tick_params twinx print_result_dict figure legend Reconstructor manage_directories2 load_pickled_data append print print getcwd join mkdir transpose resize zeros tensor numpy range int list reshape mean append array range len adaptive_avg_pool2d print ImagesPathDataset DataLoader eval numpy to empty enumerate len atleast_2d print iscomplexobj atleast_1d dot sqrtm trace eye real abs max imag mean cov get_activations load list endswith glob calculate_activation_statistics close Path to calculate_frechet_distance _compute_statistics_of_path batch_size print dims path device parse_args calculate_fid_given_paths tuple map FIDInceptionE_1 _inception_v3 load_state_dict FIDInceptionE_2 load_state_dict_from_url FIDInceptionC FIDInceptionA to exp conv3d warn conv2d conv enumerate mean pow device to gaussian_filter relu squeeze _ssim shape repeat _fspecial_gauss_1d range len view relu squeeze min _ssim avg_pool avg_pool2d shape stack repeat prod device append avg_pool3d to _fspecial_gauss_1d range len eval_basic_metrics eval_AOPC eval_sanity_weight_randomization view_AOPC view_sanity_weight_randomization load join str print ResGPTNet34 Pytorch_GPT_MNIST_Sampler to_csv eval folder_check mkdir to DataFrame mean sum astype join time print Pytorch_GPT_MNIST_Sampler compute_correlationsGC folder_check mkdir range compute_correlations GuidedBackprop LayerGradCam Deconvolution channel_adj spearmanr abs max str get_cascade_randomized_net append range normal get_sample_batch setup_xai_methods_with_cascaded_randomization enumerate join print reshape min prepare_results_dictionary to_csv isnan len load data layer1 ResGPTNet34 layer3 layer4 randomizing_one_resnet_sequential_of_basic_blocks eval to layer2 where compute_attr_one_pixel_target argmax abs spearmanr max str append range normal get_sample_batch net0 zip net enumerate join print min get_cascade_randomized_net_for_GC prepare_results_dictionary to_csv isnan numpy len items zeros len load data layer1 get_uniform_random_initialization ResGPTNet34 shape eval randomizing_one_resnet_sequential_of_basic_blocks to layer2 max str randomizing_one_resnet_basic_block getattr sum range data shape to load join time print ResGPTNet34 Pytorch_GPT_MNIST_Sampler to_csv eval folder_check mkdir setup_AOPC_dicts to array range compute_AOPC_for_y_pred str print mean compute_one_sample_AOPC_for_y_pred append get_sample_batch_uniform_random range net to where uniform numpy setup_xai_method abs max range GuidedBackprop LayerGradCam Deconvolution channel_adj show join plot print set_xlabel add_subplot folder_check set_ylabel figure legend array read_csv isnan float append plot_sanity_check_cascade_randomization join show print add_subplot folder_check figure generate_samples heatmaps_of_gen_config apply_to_mnist gen_distribution heatmaps redirect view_losses MNIST load evaluation_visual_results print float ResGPTNet34 Pytorch_GPT_MNIST_Sampler eval folder_check append to array range __getitem__ net show get_sample_batch set_title print add_subplot Pytorch_GPT_MNIST_Sampler colorbar imshow figure range Pytorch_GPT_MNIST_Sampler zero_grad criterion_gt Logger save pickle_data round str ones ResGPTNet34 Adam manually_parse_boolean append load_pickled_data to get_sample_batch_uniform_random CrossEntropyLoss range item float net load items time criterion backward criterion_gc save_ckpt_dir parameters folder_check train step argmax show set_title ones develop add_subplot imshow nBSG array figure savefig plot_setting numpy code_generator range set_yticks set_xticks show items plot print reshape xlabel ylabel mean loss_array folder_check Logger legend iter_array load_pickled_data load str get_sample_batch evaluation_visual_results print ResGPTNet34 Pytorch_GPT_MNIST_Sampler manually_parse_boolean folder_check Logger load_pickled_data to get_sample_batch_uniform_random net range print load join LayerGradCam get_sample_batch print arrange_heatmaps ResGPTNet34 Pytorch_GPT_MNIST_Sampler Deconvolution GuidedBackprop eval folder_check mkdir to numpy channel_adj show str items set_title abs print add_subplot imshow savefig figure plot_setting max range enumerate len compute_attr_non_emtpy_gens Pytorch_GPT_MNIST_Sampler compute_attr_transformations_of_non_emtpy_gens argmax str ResGPTNet34 to range get_sample_batch eval mkdir zip net load join print arrange_heatmapsGC folder_check numpy zip compute_attr_one_pixel_target where len zip compute_attr_one_pixel_target where len attribute LayerGradCam Deconvolution WrapperNet GuidedBackprop numpy channel_adj show str plot_setting2 items set_title abs print add_subplot close imshow savefig figure zeros plot_setting max enumerate len set_yticks set_xticks print join mkdir exists int join mkdir print do_finetune zero_grad criterion_gt save pickle_data setup_main_components round str all Adam manually_parse_boolean append get_sample_batch_uniform_random CrossEntropyLoss range net items time criterion backward print LossMonitor criterion_gc compute_current_running_avg save_ckpt_dir parameters folder_check train step load str print ResGPTNet34 Pytorch_GPT_MNIST_Sampler Logger load_pickled_data to sum get_sample_batch choice eval save_samples_for_display_and_adjustments mkdir load join get_sample_batch evaluation_visual_results print choice net range | # GPT This repository contains codes related to the concepts General Pattern Theory (GPT) from: _Grenander, U. General Pattern Theory: A Mathematical Study of Regular Structures. Oxford Mathematical Monographs. Clarendon Press, 1993. ISBN 9780198536710_. We refer to this as the _textbook_. See "Get Started.docx" in notebooks folder. Currently available: **1. src/APL_to_python**<br> The folder contains python version of APL codes in the textbook. See "Get Started.docx" to read about step by step translations plus simple testings. Or, jump directly to the results using the following (also try it with different example_n): ``` python simple_algo_section4.2.2.py --example_n 1 ``` | 2,044 |
eugene/spngp | ['gaussian processes', 'time series'] | ['Learning Deep Mixtures of Gaussian Process Experts Using Sum-Product Networks'] | learnspngp.py concrete-spngp.py spngp.py energy-spngp.py cccp-spngp.py Mixture Separator build_bins GPMixture query build _cached_gp get_splits Color Split ExactGPModel dump_tensors pretty_size GP Sum structure get GPMixture range Mixture arange query maxs depth values list Separator Counter _cached_gp append get mins copy mean pop print divmod extend dimension len get T print zeros sum array enumerate Mixture query maxs get_splits depth argmax values list all Separator logical_and Counter _cached_gp append get copy mins choice mean pop print min extend dimension any len get pop children list ones extend Split GP zip Sum append len print is_tensor get_objects | eugene/spngp | 2,045 |
eunh/low_dose_CT | ['denoising'] | ['Deep Convolutional Framelet Denosing for Low-Dose CT via Wavelet Residual Network', 'Wavelet Domain Residual Network (WavResNet) for Low-Dose X-ray CT Reconstruction'] | matconvnet-1.0-beta24/doc/matdoc.py matconvnet-1.0-beta24/utils/proto/caffe_b590f1d_pb2.py matconvnet-1.0-beta24/utils/layers.py matconvnet-1.0-beta24/utils/proto/caffe_old_pb2.py matconvnet-1.0-beta24/utils/proto/caffe_fastrcnn_pb2.py matconvnet-1.0-beta24/utils/proto/caffe_6e3916_pb2.py matconvnet-1.0-beta24/utils/import-caffe.py matconvnet-1.0-beta24/utils/proto/vgg_caffe_pb2.py matconvnet-1.0-beta24/utils/proto/caffe_0115_pb2.py matconvnet-1.0-beta24/utils/proto/caffe_pb2.py matconvnet-1.0-beta24/doc/matdocparser.py extract render_L render_V render_P render_DIVL render_DH render_BL render_L_from_indent render_SL render_S render Context render_DI Frame render_B findNextFunction getFunctionDoc readText MatlabFunction render_DL clean Lexer P PL Parser BH EOF DI L DL SL BL DH DIVL Terminal S DIV NonTerminal B V Symbol blobproto_to_array versiontuple dict_to_struct_array keyboard tolist escape bilinear_interpolate getopts find rowcell CaffeInnerProduct ConversionError CaffeScale CaffeBatchNorm CaffeLayer CaffeCrop CaffeConcat CaffeConv CaffePooling CaffeData reorder CaffeElementWise getFilterOutputSize CaffeROIPooling CaffeSoftMaxLoss CaffeReLU getFilterTransform CaffeModel CaffeTransform CaffeDeconvolution row dictToMatlabStruct rowarray CaffeBlob transposeTransform CaffeDropout CaffeLRN CaffeEltWise composeTransforms CaffeSoftMax HingeLossParameter BlobProto BlobProtoVector NetStateRule LayerParameter PowerParameter FillerParameter ArgMaxParameter V0LayerParameter InnerProductParameter ConvolutionParameter SolverState EltwiseParameter SliceParameter WindowDataParameter DummyDataParameter HDF5OutputParameter TanHParameter TransformationParameter SoftmaxParameter ConcatParameter DataParameter SolverParameter MVNParameter ContrastiveLossParameter NetState NetParameter PoolingParameter DropoutParameter Datum SigmoidParameter AccuracyParameter MemoryDataParameter LRNParameter ReLUParameter ImageDataParameter InfogainLossParameter HDF5DataParameter ThresholdParameter ReductionParameter HingeLossParameter BlobProto BlobProtoVector NetStateRule LayerParameter PowerParameter FillerParameter ArgMaxParameter V0LayerParameter InnerProductParameter ConvolutionParameter SolverState EltwiseParameter LossParameter SliceParameter WindowDataParameter DummyDataParameter HDF5OutputParameter TanHParameter TransformationParameter SoftmaxParameter ConcatParameter DataParameter SPPParameter ParamSpec EmbedParameter SolverParameter MVNParameter ContrastiveLossParameter NetState NetParameter PoolingParameter DropoutParameter Datum SigmoidParameter BlobShape ExpParameter AccuracyParameter LogParameter ThresholdParameter TileParameter MemoryDataParameter LRNParameter ReLUParameter ImageDataParameter ReshapeParameter InfogainLossParameter V1LayerParameter HDF5DataParameter PReLUParameter FlattenParameter PythonParameter ReductionParameter HingeLossParameter BlobProto BlobProtoVector NetStateRule LayerParameter PowerParameter FillerParameter ArgMaxParameter V0LayerParameter InnerProductParameter ConvolutionParameter SolverState EltwiseParameter LossParameter SliceParameter BatchNormParameter WindowDataParameter DummyDataParameter HDF5OutputParameter TanHParameter TransformationParameter SoftmaxParameter ConcatParameter DataParameter SPPParameter ParamSpec EmbedParameter SolverParameter MVNParameter ContrastiveLossParameter NetState NetParameter BiasParameter PoolingParameter DropoutParameter Datum SigmoidParameter BlobShape ExpParameter AccuracyParameter LogParameter ThresholdParameter TileParameter MemoryDataParameter LRNParameter ReLUParameter ImageDataParameter ELUParameter ReshapeParameter InfogainLossParameter ScaleParameter V1LayerParameter HDF5DataParameter PReLUParameter FlattenParameter PythonParameter ROIPoolingParameter HingeLossParameter BlobProto BlobProtoVector NetStateRule LayerParameter PowerParameter FillerParameter ArgMaxParameter V0LayerParameter InnerProductParameter ConvolutionParameter SolverState EltwiseParameter LossParameter SliceParameter WindowDataParameter DummyDataParameter HDF5OutputParameter TanHParameter TransformationParameter SoftmaxParameter ConcatParameter DataParameter ParamSpec SolverParameter MVNParameter ContrastiveLossParameter NetState NetParameter PoolingParameter DropoutParameter Datum SigmoidParameter BlobShape ExpParameter AccuracyParameter ThresholdParameter MemoryDataParameter LRNParameter ReLUParameter ImageDataParameter InfogainLossParameter V1LayerParameter HDF5DataParameter PReLUParameter PythonParameter NetParameter LayerConnection BlobProto BlobProtoVector LayerParameter FillerParameter Datum SolverParameter SolverState BlobProto BlobProtoVector PowerParameter LayerParameter FillerParameter V0LayerParameter InnerProductParameter ConvolutionParameter SolverState WindowDataParameter HDF5OutputParameter ConcatParameter DataParameter SolverParameter NetParameter PoolingParameter DropoutParameter Datum MemoryDataParameter LRNParameter ImageDataParameter InfogainLossParameter HDF5DataParameter NetParameter EvalHistoryIter LayerConnection BlobProto BlobProtoVector LayerParameter FillerParameter Datum EvalHistory SolverParameter SolverState search clean match strip group append getFunctionDoc findNextFunction print print print children render_SL print pop render_DH print Frame push render_DIVL children render_DI print children render_L print pop children render_L_from_indent isa Frame render_B push indent pop children Frame push render_DIVL children render_BL render_V render_S render_DL isa render_P print Context render_DIVL dim tolist hasattr list empty keys RepeatedScalarFieldContainer isinstance update print f_locals interact copy reshape asarray astype clip hasattr list ndarray isinstance append empty keys CaffeTransform CaffeTransform | Paper =============== * A deep convolutional neural network using directional wavelets for low-dose X-ray CT reconstruction * published in Medical Physics (2017): [http://onlinelibrary.wiley.com/doi/10.1002/mp.12344/full] * *2nd winner* of '**2016 Low-Dose CT Grand Challenge**' * Wavelet Domain Residual Network (WavResNet) for Low-Dose X-ray CT Reconstruction * Accepted at Fully3D 2017: [https://arxiv.org/abs/1703.01383] * Deep Convolutional Framelet Denoising for Low-Dose CT via Wavelet Residual Network * published in IEEE Transactions on Medical Imaging (2018): [https://ieeexplore.ieee.org/abstract/document/8332971] Implementation | 2,046 |
evanmiltenburg/LabelingPeople | ['text generation'] | ['Talking about other people: an endless range of possibilities'] | VisualGenome/account_for_all.py Other/length_stats.py VisualGenome/select_gendered_attributes.py Flickr30k/update_grammar.py Flickr30k/check_labels.py Flickr30k/flickr_stats.py Flickr30k/label_parser.py gendered_labels remove_stopwords grammatical clear_stopwords correct_typos load_json normalize normalize_label gendered_labels all_labels analyze_label write_grammar create_lexical_rule load_lexical_items extend_grammar load_json load_face2text gendered_labels get_lengths load_json load_zipped_json get_attribute_dict select_attributes enumerate split replace IGNORECASE compile remove_stopwords join parse replace Counter Counter print list parse split join load_lexical_items create_lexical_rule extend append load_json load close ZipFile open strip defaultdict load_zipped_json update | evanmiltenburg/LabelingPeople | 2,047 |
ezhan94/calibratable-style-consistency | ['imitation learning'] | ['Learning Calibratable Policies using Programmatic Style-Consistency'] | lib/distributions/multinomial.py util/datasets/bball/core.py lib/distributions/normal.py lib/models/ctvae_info.py scripts/check_dynamics_loss.py train.py run.py lib/distributions/bernoulli.py util/datasets/__init__.py util/datasets/bball/label_functions/__init__.py util/datasets/bball/label_functions/heuristics.py lib/distributions/__init__.py util/environments/bball.py lib/models/ctvae_style.py lib/distributions/core.py lib/models/ctvae_mi.py scripts/visualize_samples_ctvae.py util/logging/core.py util/logging/__init__.py lib/models/__init__.py run_single.py scripts/compute_stylecon_ctvae.py util/environments/core.py util/environments/__init__.py lib/models/core.py util/datasets/core.py util/datasets/bball/__init__.py lib/distributions/dirac.py lib/models/ctvae.py run_config run_epoch start_training Bernoulli Distribution Dirac Multinomial Normal BaseSequentialModel CTVAE CTVAE_info CTVAE_mi CTVAE_style get_model_class check_selfcon_tvaep_dm compute_stylecon_ctvae visualize_samples_ctvae LabelFunction TrajectoryDataset load_dataset normalize BBallDataset _set_figax unnormalize Displacement AverageSpeed Curvature Destination Direction BBallEnv BaseEnvironment generate_rollout load_environment LogEntry join format start_training print copy save_dir makedirs str optimize model losses itemize train print absorb requires_environment average eval LogEntry requires_labels to enumerate get_model_class label_dim requires_environment DataLoader save round seed filter_and_load_state_dict active_label_functions dirname action_dim load_dataset append load_environment to sum run_epoch range state_dict format num_parameters prepare_stage lower manual_seed requires_labels stage load join time pop isinstance print state_dim summary randint get_model_class DataLoader clf set_major_formatter abs filter_and_load_state_dict model_class active_label_functions view transpose ylabel title savefig load_dataset append format PercentFormatter mean lower xlim enumerate load join print xlabel hist median train numpy len get_model_class categorical DataLoader linspace values filter_and_load_state_dict model_class active_label_functions list num_values ones name transpose tolist ylabel randperm title savefig iter load_dataset load_environment next sum cat format plot size close eval lower item label enumerate load join items int print xlabel num_samples zeros fill_between get_model_class arange categorical DataLoader unsqueeze around linspace save max output_dim filter_and_load_state_dict model_class active_label_functions ones transpose iter load_dataset load_environment next format eval lower load join print min repeat zeros lower int int join set_xlim add_subplot imshow set_visible figure resize imread set_ylim get_obs is_recurrent act update_hidden unsqueeze append step range | # Learning Calibratable Policies using Programmatic Style-Consistency [(arXiv)](https://arxiv.org/abs/1910.01179) ## Demo The demo will be live during ICML 2020 [here](http://basketball-ai.com/). ## Code Code is written in Python 3.7.4 and [PyTorch](https://pytorch.org/) v.1.0.1. Will be updated for PyTorch 1.3 in the future. ## Usage Train models with: `$ python run_single.py -d <device id> --config_dir <config folder name>` Not specifying a device will use CPU by default. See JSONs in `configs\` to see examples of config files. ### Test Run | 2,048 |
ezhan94/gen-MA-BC | ['imitation learning'] | ['Generating Multi-Agent Trajectories using Programmatic Weak Supervision'] | scripts/compute_bball_stats.py scripts/print_params.py models/macro_shared_vrnn.py models/vrnn_mixed.py train.py datasets/boids/__init__.py datasets/boids/generate_data.py models/vrnn_single.py models/__init__.py datasets/boids/boid.py models/utils.py datasets/boids/core.py models/macro_vrnn.py sample.py datasets/bball/__init__.py datasets/boids/visualize.py datasets/__init__.py datasets/bball/cfg.py datasets/boids/cfg.py models/vrnn_indep.py scripts/show_groundtruth.py models/vrae_mi.py datasets/bball/label_macro_intents.py datasets/bball/visualize.py models/rnn_gauss.py datasets/bball/core.py hyperparams_str run_epoch printlog loss_str GeneralDataset normalize preprocess unnormalize fetch bound get_macro_intent compute_macro_intents_fixed label_macro_intents compute_macro_intents_stationary animate _get_cmap _set_figax display BoidList Boid clip_norm normalize preprocess unnormalize fetch animate _set_figax display MACRO_SHARED_VRNN MACRO_VRNN RNN_GAUSS index_by_agent kld_gauss num_trainable_params sample_multinomial sample_gauss get_params_str entropy_gauss one_hot_encode cudafy_list nll_gauss parse_model_params get_macro_ohe VRAE_MI VRNN_INDEP VRNN_MIXED VRNN_SINGLE load_model oob_rate print format model backward transpose clip_grad_norm_ zero_grad step parameters item sum cuda values enumerate label_macro_intents join format DATAPATH int bound norm list get_macro_intent reversed append zeros range len list get_macro_intent reversed zeros range len int SEQUENCE_DIMENSION join format savez DATAPATH compute_macro_intents_fixed compute_macro_intents_stationary zeros range makedirs set_xlim add_subplot DATAPATH imshow set_visible figure resize imread set_ylim CMAP_OFFENSE show int _set_figax plot _get_cmap N_MACRO_Y add_patch tight_layout savefig Rectangle unnormalize range len FuncAnimation int show _set_figax _get_cmap tight_layout save unnormalize len norm C_SEP R_LOCAL C_ALI BOOST_MAX C_COH R_CLOSE BOOST_MIN BOUND C_ORI Circle add_patch BOUND parameters size parse_known_args add_argument getattr cuda range len transpose clone data one_hot_encode size is_cuda zeros cuda range normal_ cuda is_cuda FloatTensor log pow cuda is_cuda pow log is_cuda append zeros long scatter_ one_hot_encode size squeeze cuda is_cuda lower zeros logical_or sum shape | # Generating Multi-Agent Trajectories using Programmatic Weak Supervision Code for paper titled [Generating Multi-Agent Trajectories using Programmatic Weak Supervision](https://arxiv.org/abs/1803.07612) by Zhan et al., ICLR 2019. ## Installation & Setup Code is written using [PyTorch](https://pytorch.org/) version `1.0.0`. After cloning the repository, you need to download the data. **[Update 11/25/20]** The basketball dataset is now available on [AWS Data Exchange](https://aws.amazon.com/marketplace/pp/prodview-7kigo63d3iln2?qid=1606330770194&sr=0-1&ref_=srh_res_product_title#offers). Please make sure to acknowledge Stats Perform if you use the data for your research. The Boids dataset can be generated by running: ``` $ python datasets/boids/generate_data.py ``` | 2,049 |
ezhan94/multiagent-programmatic-supervision | ['imitation learning'] | ['Generating Multi-Agent Trajectories using Programmatic Weak Supervision'] | scripts/compute_bball_stats.py scripts/print_params.py models/macro_shared_vrnn.py models/vrnn_mixed.py train.py datasets/boids/__init__.py datasets/boids/generate_data.py models/vrnn_single.py models/__init__.py datasets/boids/boid.py models/utils.py datasets/boids/core.py models/macro_vrnn.py sample.py datasets/bball/__init__.py datasets/boids/visualize.py datasets/__init__.py datasets/bball/cfg.py datasets/boids/cfg.py models/vrnn_indep.py scripts/show_groundtruth.py models/vrae_mi.py datasets/bball/label_macro_intents.py datasets/bball/visualize.py models/rnn_gauss.py datasets/bball/core.py hyperparams_str run_epoch printlog loss_str GeneralDataset normalize preprocess unnormalize fetch bound get_macro_intent compute_macro_intents_fixed label_macro_intents compute_macro_intents_stationary animate _get_cmap _set_figax display BoidList Boid clip_norm normalize preprocess unnormalize fetch animate _set_figax display MACRO_SHARED_VRNN MACRO_VRNN RNN_GAUSS index_by_agent kld_gauss num_trainable_params sample_multinomial sample_gauss get_params_str entropy_gauss one_hot_encode cudafy_list nll_gauss parse_model_params get_macro_ohe VRAE_MI VRNN_INDEP VRNN_MIXED VRNN_SINGLE load_model oob_rate print format model backward transpose clip_grad_norm_ zero_grad step parameters item sum cuda values enumerate label_macro_intents join format DATAPATH int bound norm list get_macro_intent reversed append zeros range len list get_macro_intent reversed zeros range len int SEQUENCE_DIMENSION join format savez DATAPATH compute_macro_intents_fixed compute_macro_intents_stationary zeros range makedirs set_xlim add_subplot DATAPATH imshow set_visible figure resize imread set_ylim CMAP_OFFENSE show int _set_figax plot _get_cmap N_MACRO_Y add_patch tight_layout savefig Rectangle unnormalize range len FuncAnimation int show _set_figax _get_cmap tight_layout save unnormalize len norm C_SEP R_LOCAL C_ALI BOOST_MAX C_COH R_CLOSE BOOST_MIN BOUND C_ORI Circle add_patch BOUND parameters size parse_known_args add_argument getattr cuda range len transpose clone data one_hot_encode size is_cuda zeros cuda range normal_ cuda is_cuda FloatTensor log pow cuda is_cuda pow log is_cuda append zeros long scatter_ one_hot_encode size squeeze cuda is_cuda lower zeros logical_or sum shape | # Generating Multi-Agent Trajectories using Programmatic Weak Supervision Code for paper titled [Generating Multi-Agent Trajectories using Programmatic Weak Supervision](https://arxiv.org/abs/1803.07612) by Zhan et al., ICLR 2019. ## Installation & Setup Code is written using [PyTorch](https://pytorch.org/) version `1.0.0`. After cloning the repository, you need to download the data. **[Update 11/25/20]** The basketball dataset is now available on [AWS Data Exchange](https://aws.amazon.com/marketplace/pp/prodview-7kigo63d3iln2?qid=1606330770194&sr=0-1&ref_=srh_res_product_title#offers). Please make sure to acknowledge Stats Perform if you use the data for your research. The Boids dataset can be generated by running: ``` $ python datasets/boids/generate_data.py ``` | 2,050 |
fGAIL3456/fGAIL | ['imitation learning'] | ['$f$-GAIL: Learning $f$-Divergence for Generative Adversarial Imitation Learning'] | a2c_ppo_acktr/arguments.py a2c_ppo_acktr/envs.py a2c_ppo_acktr/algo/ppo.py a2c_ppo_acktr/algo/kfac.py a2c_ppo_acktr/algo/__init__.py a2c_ppo_acktr/matrix_utils.py a2c_ppo_acktr/algo/bc.py a2c_ppo_acktr/algo/gail.py a2c_ppo_acktr/algo/fgail.py save_expert_traj.py a2c_ppo_acktr/algo/a2c_acktr.py a2c_ppo_acktr/model_utils.py main.py main get_args TransposeImage VecPyTorch make_env MaskGoal TransposeObs VecNormalize VecPyTorchFrameStack make_vec_envs TimeLimitMask log_det_2x2 compute_log_det exact_matrix_logarithm_trace power_series_matrix_logarithm_trace power_series_full_jac_exact_trace log_det_other weak_bound MaxMinGroup ActNorm2D ActNorm maxout_by_group minout_by_group get_all_params squeeze injective_pad Split ListModule Layer process_maxmin_groupsize split merge A2C_ACKTR BC T weightConstraint f_conjugate Discriminator ExpertDataset Discriminator _extract_patches compute_cov_g compute_cov_a KFACOptimizer SplitBias update_running_stat PPO ppo_epoch update_linear_schedule recurrent_hidden_state_size save seed str _obfilt algo fnum env_name manual_seed_all clip_param use_proper_time_limits mean manual_seed join time evaluate makedirs fgail_epoch predict_reward use_gae median step num_processes DataLoader save_dir max append expanduser to experts_dir after_update num_steps insert fgail_batch_size eval optimizer value_loss_coef min ExpertDataset ob_rms shape use_linear_lr_decay Discriminator range update dump get_args PPO cleanup_log_dir num_env_steps gamma num_mini_batch log_dir print set_num_threads action_space reset gae_lambda fgail_T_activation copy_ device fgail open gail FloatTensor make_vec_envs format compute_returns lr deque A2C_ACKTR int entropy_coef RolloutStorage Policy len parse_args recurrent_policy add_argument ArgumentParser VecPyTorch SubprocVecEnv VecNormalize DummyVecEnv VecPyTorchFrameStack sum view size stack numpy eye empty range zero_gradients zeros_like compute_log_det matmul sum range list view insert size squeeze matmul shape stack device to sum range size stack view range int size contiguous hasattr isinstance previous_functions nelement insert size list process_maxmin_groupsize process_maxmin_groupsize data view size contiguous unfold _extract_patches view ones size mean div_ cuda is_cuda view size contiguous mul_ sum | # fGAIL This repository contains codes for the project of f-GAIL, which aims at automatically learn an f-divergence for GAIL. Please find the paper at (https://arxiv.org/abs/2010.01207). The f-GAIL implementation can be found in ``` a2c_ppo_acktr/algo/fgail.py ``` ## Running the code Train expert with predefined rewards: ``` python main.py --env-name HalfCheetah-v2 --algo ppo --use-gae --lr 3e-4 --clip-param 0.1 --value-loss-coef 0.5 --num-processes 8 --num-steps 1000 --num-mini-batch 4 --log-interval 1 --use-linear-lr-decay --entropy-coef 0.01 ``` | 2,051 |
fabianbalsiger/point-cloud-segmentation-miccai2019 | ['semantic segmentation'] | ['Learning Shape Representation on Sparse Point Clouds for Volumetric Image Segmentation'] | pc/utilities/testing.py pc/data/data.py pc/utilities/evaluation.py pc/model/sampling/tf_sampling.py snippets/create_dataset.py pc/model/base.py pc/data/split.py pc/utilities/assembler.py pc/utilities/plotting_qualitative.py bin/main.py pc/model/point_cnn.py pc/model/sampling/sampling.py pc/utilities/filesystem.py pc/model/point_cnn_util.py pc/utilities/seeding.py pc/model/point_cnn_patch.py snippets/create_split.py pc/utilities/training.py setup.py pc/utilities/augmentation.py pc/data/handler.py pc/configuration/config.py main load Configuration ImageInformationConfiguration point_cloud_to_image save_to_csv FileTypes PointCloudDataHandler PointCloudIndexing ConcatenateCoordinatesAndPointFeatures PointCloudSizeExtractor split_subject_by_amount load_split save_split BaseModel PointCNN xconv Base Encoder find_duplicate_columns knn_indices_general compute_curvature curvature_based_sample get_indices sort_points batch_distance_matrix_general random_choice_2d depthwise_conv2d compute_determinant batch_normalization conv2d knn_indices compute_eigenvals inverse_density_sampling distance_matrix dense batch_distance_matrix prepare_for_unique_top_k separable_conv2d farthest_point_sample gather_point _gather_point_grad prob_sample init_shape init_subject_assembler on_sample_fn ShuffledDataExtractor PointCloudRotate PointCloudShuffler PointCloudJitter init_evaluator EvaluatorAggregator AggregatedResult aggregate_results AggregatedResultWriter prepare_epoch_result_directory get_directory_name prepare_directories PointCloudPlotter set_seed process_predictions TensorFlowTester validate_on_subject AssemblingTester SegmentationTrainer normalize_unit_cube Collector concat LoadData main create_sample_data main create_split check_for_duplicates load seed format set_seed print prepare_directories cudnn_determinism split_file Configuration PointCloudDataHandler load_split range to_csv apply_along_axis concatenate append remove exists dense subtract knn_indices_general concat reshape depthwise_conv2d matmul squeeze conv2d gather_nd expand_dims separable_conv2d isinstance concatenate min choice randrange append expand_dims full range transpose matmul reduce_sum transpose matmul reduce_sum transpose matmul reduce_sum fill range unique int32 py_func batch_distance_matrix reshape concat prepare_for_unique_top_k shape top_k tile range reshape batch_distance_matrix_general concat prepare_for_unique_top_k shape top_k tile range norm constant print subtract concat reduce_max exit reshape reduce_sum shape reduce_mean gather_nd top_k startswith tile expand_dims reduce_min range reshape cos square pi sqrt compute_determinant trace clip_by_value eye acos compute_eigenvals matmul reduce_sum reduce_mean expand_dims reduce_min reshape concat compute_curvature shape top_k tile range ones range choice batch_distance_matrix reshape concat reduce_sum reduce_mean top_k int32 set_shape tile abs range py_func separable_conv2d numpy FOREGROUND_NAME CSVEvaluatorWriter EvaluatorAggregator Evaluator add_writer ConsoleEvaluatorWriter add_label clear items list isinstance keys mean AggregatedResult writers append float std join format makedirs load join basename result_dir directory_name_fn copyfile model_dir split_file exists makedirs experiment model seed manual_seed_all manual_seed set_random_seed join init_evaluator format list plot extractor_test evaluate squeeze convert WriteImage direct_extract get_assembled_subject argmax keys point_cloud_to_image makedirs save_to_csv extractor_test prepare_epoch_result_directory aggregate_results direct_extract get_assembled_subject argmax init_evaluator basename list squeeze metric format current_epoch plot WriteImage mean keys join evaluate result_dir print convert log_scalar write epoch_duration std point_cloud_to_image makedirs min max join format GetSize CopyInformation rand WriteImage GetImageFromArray ReadImage range makedirs Collector get_subject_files len makedirs dirname subject exists create_sample_data create_split shuffle join format check_for_duplicates save_split split_subject_by_amount | # Learning Shape Representation on Sparse Point Clouds for Volumetric Image Segmentation This repository contains code for the [MICCAI 2019](https://www.miccai2019.org) paper "Learning Shape Representation on Sparse Point Clouds for Volumetric Image Segmentation", which can be found at [https://doi.org/10.1007/978-3-030-32245-8_31](https://doi.org/10.1007/978-3-030-32245-8_31). ## Installation The installation has been tested with Ubuntu 16.04, Python 3.6, TensorFlow 1.10, and CUDA 9.0. The ``setup.py`` file lists all other dependencies. First, create a virtual environment named `pc` with Python 3.6: $ virtualenv --python=python3.6 pc $ source ./pc/bin/activate Second, copy the code: $ git clone https://github.com/fabianbalsiger/point-cloud-segmentation-miccai2019 $ cd point-cloud-segmentation-miccai219 | 2,052 |
fabon/chatbot-intents | ['text classification', 'intent detection'] | ['“Where is My Parcel?” Fast and Efficient Classifiers to Detect User Intent in Natural Language'] | chatbot-intents/utils/path_utils.py chatbot-intents/ngrams/model_ngrams.py chatbot-intents/lstm/model_lstm.py chatbot-intents/utils/nlp_utils.py chatbot-intents/ngrams/train_ngrams.py chatbot-intents/utils/dataset_utils.py chatbot-intents/lstm/train_lstm.py load_model_lstm perfs_binary main train_model perfs_multiclass load_model_ngrams perfs_binary train_model load_english_model test_model main perfs_multiclass load_datasets import_training_sets load_datasets_sequence load_datasets_binary import_clean_labels clean_index token_is_1char clean_token encode_messages_sequence extract_index encode_messages_binary load_data token_is_digit tokenize_ngrams import_stopwords rebuild_index ngrams Input f1_score recall_score precision_score f1_score recall_score precision_score to_categorical argmax open import_clean_labels Adam sum range predict perfs_multiclass dump set load_datasets_sequence load_model_lstm float compile load print fit summary numpy array len train_model iqr print min average append median max range Input print numpy argmax predict perfs_multiclass load array import_clean_labels open SGD tolist shape load_datasets_binary load_model_ngrams import_training_sets open sorted list test_model OrderedDict close keys flush items join write load open print len load clean_index time print tolist len extract_index encode_func rebuild_index __name__ open set append append join append tokenize ngrams tokenize_ngrams lil_matrix print shape tokenize_ngrams max zeros tokenize_ngrams max list keys list keys | # Chatbot-intents Classifiers detecting user intent in Natural Language * `bash configure` - initialise the chatbot-intents package for multilingual intent detection. * `make setup` - synchronize your conda environement. * `make model-language` - train a new intent classifier where `model` is in ngrams or lstm and `language` in {english, french, russian, german, italian, spanish} Full paper: [IEEE Xplore](https://ieeexplore.ieee.org/document/8931717) [Read the pdf here!](where_is_my_parcel.pdf) | 2,053 |
facebookresearch/CoDraw | ['imitation learning'] | ['CoDraw: Collaborative Drawing as a Testbed for Grounded Goal-driven Communication'] | script/preprocess.py json_save makePartialJsons os_mkdir mkdir json_save join os_mkdir print | # CoDraw Dataset Collaborative Drawing game (CoDraw) involves perception, communication, and actions in a partially observable virtual environment. Our game is grounded in a virtual world constructed by clip art objects [[Zitnick et al. (2013)](http://ieeexplore.ieee.org/document/6751319/) & [Zitnick and Parikh (2013)](https://www.cv-foundation.org/openaccess/content_cvpr_2013/papers/Zitnick_Bringing_Semantics_into_2013_CVPR_paper.pdf)], and a drag-and-drop interface based on the Visual Dialog chat interface [[Das et al. (2017)](https://visualdialog.org/)] as shown below: <p align="center"><img alt="Overview of the proposed Collaborative Drawing task" src="imgs/codraw_schema.png" width="400px"></p> Two players, **(a)** Teller and **(b)** Drawer, play the game. The Teller sees an abstract scene made from clip art objects, with a semantically meaningful configuration, while the Drawer sees a drawing canvas initialized with an empty canvas. Both players need to collaborate via natural language communication so that the Drawer can reconstruct the image of the Teller. ### Peek Chance While language-based communication is effective when exchanging high-level ideas, to get the details right, direct input of visual information could be constructive. For this, we give one chance for the Teller to peek at the Drawer's canvas using the `peek` button in the interface. Communication is only allowed after the peek window is closed. ## Download * [CoDraw v1.0 (149MB)](https://drive.google.com/file/d/0B-u9nH58139bTy1XRFdqaVEzUGs/view?usp=sharing) ### Links * [CoDraw Models](https://github.com/facebookresearch/codraw-models) | 2,054 |
facebookresearch/DME | ['word embeddings'] | ['Dynamic Meta-Embeddings for Improved Sentence Representations'] | dme/tasks/__init__.py dme/tasks/img_cap_retrieval.py dme/args.py train.py get_flickr30k.py dme/__init__.py dme/logger.py dme/encoders.py dme/embeddings.py dme/tasks/nli.py get_multinli.py dme/tasks/text_classification.py get_sst2.py dme/datasets/__init__.py get_embeddings.py dme/datasets/img_cap_retrieval.py dme/datasets/nli.py dme/datasets/sst.py dme/embedders.py dme/tasks/base.py get_snli.py main normf main get_pil_img MyImageFolder main main main EarlyStoppingCriterion get_data_iters resume_from_checkpoint save_checkpoint get_data_splits main preprocess get_available_embeddings CatEmbedder get_emb_key load_cached_pretrained_embeddings match_pretrained_embeddings SingleEmbedder ProjSumEmbedder get_embeds_vocab normalize_embeddings get_embedder SentEncoder create_logger LogFormatter ImageCaptionDataset NLIDataset MultiNLIDataset SNLIDataset AllNLIDataset SST2Dataset nn_init get_optimizer_scheduler init_weight BaseTask normf count_param_num get_lr ImageCaptionRetrievalRanker ImageCaptionRetrievalModel ImageFeatureEncoder ImageCaptionRetrievalTask VseLoss CaptionEncoder ImageEncoder LanguageInferenceModel NaturalLanguageInferenceTask TextClassificationTask TextClassificationModel join chdir print endswith system makedirs join convert open load flickr30k_root model Variable Sequential tqdm normf DataLoader MyImageFolder resnet152 numpy cuda enumerate download save load load_state_dict splits info datasets_root Field get_embeds_vocab build_vocab get_dataloader splits batch_sz info itos EarlyStoppingCriterion resume_from_checkpoint checkpoint_path clip_grad_norm_ zero_grad is_improved n_classes create_logger TextClassificationTask save_checkpoint stoi NaturalLanguageInferenceTask cache_path init_epoch hasattr ImageCaptionRetrievalTask load_state_dict ceil grad_clip get_lr batch_forward range early_stop_patience get_train_stats get_data_iters report_train_stats state_dict max_epochs resume_from info optimizer int deepcopy scheduler evaluate backward dict parameters get_data_splits train step len cache_path join savedir checkpoint_path name min resume_from get_available_embeddings exists makedirs info split Tensor exists open items list copy_ get_emb_key info append mean info join sorted embeds len set add info open exists split setFormatter getLogger addHandler LogFormatter StreamHandler DEBUG setLevel INFO FileHandler ReduceLROnPlateau Adam SGD param_groups xavier_uniform_ kaiming_uniform_ orthogonal_ isinstance Sequential init_weight named_parameters getattr uniform_ split | # Dynamic Meta-Embeddings for Improved Sentence Representations Code and models for the paper [Dynamic Meta-Embeddings for Improved Sentence Representations](https://arxiv.org/abs/1804.07983). ## Requirements * Python 2.7 or 3.6+ * PyTorch >= 0.4.1 * torchtext >= 0.2.3 * torchvision >= 0.2.1 * Spacy >= 2.0.11 * NumPy >= 1.14.0 * jsonlines | 2,055 |
facebookresearch/codraw-models | ['imitation learning'] | ['CoDraw: Collaborative Drawing as a Testbed for Grounded Goal-driven Communication'] | baseline3_models.py baseline3_eval.py exp28_scenenn.py abs_metric.py baseline2_eval.py interactivity.py baseline1_train.py baseline1_eval.py episode.py baseline4_eval.py abs_render.py saved_models.py baseline1_models.py attention.py baseline2_train.py abs_util_orig.py eval_run_bots.py datagen.py codraw_data.py eval_automatic.py model.py nkfb_util.py baseline3_train.py eval_transcripts.py example.eval_server_common.py baseline4_train.py packer.py baseline4_models.py baseline2_models.py scene_similarity_orig scene_similarity clipart_similarity scene_similarity_v1 clipart_similarity_v1 scene_similarity_v2 svg_from_cliparts display_cliparts snippet_from_clipart get_image_name AbsUtil AttentionSeqToMasked BOWtoClipartDrawer NearestNeighborTeller BOWtoMultiBCEDrawer load_baseline1 BOWNeighborDrawer ClipartToSeqTeller BOWplusCanvasDrawer CharNeighborDrawer BaseAddOnlyDrawer BOWAddOnlyDrawer PragmaticNearestNeighborTeller LSTMAddOnlyDrawer load_baseline2 SceneToSeqTeller load_baseline3 train_teller examples_from_episodes load_baseline4 process_episode collect_episodes RLSceneToSeqTeller discount_rewards train_teller SelectClipart get_truth_and_human_scenes_pre_peek events_from_datum_set_clipart_pre_peek get_contextual_place_many get_scenes_and_scripts Agent get_set_clipart_pre_peek Clipart ObserveTruth DrawClipart get_set_clipart cached_split_wrapper ReplyGroup Event TellGroup AbstractScene events_from_datum_set_clipart ObserveCanvas get_scenes_and_scripts_with_peek events_from_datum_contextual_place_many events_from_datum_place_many TellerIntention get_truth_and_human_scenes events_from_datum_place_one get_place_one DrawGroup get_truth_and_human_scenes_with_js_scores get_scenes Peek data_for_splits SetDrawing TellerObserveCanvas get_place_many vocabulary_for_split NearestNeighborData BOWplusCanvasToMultiData BOWAddUpdateData SceneToSeqData MessageSimilarityData BOWtoClipartData ClipartToSeqData vocabulary_counter_for_split Datagen Transcriber respond_to response_partial Episode print_components_script print_pairwise print_script print_human print_components_pairwise print_eval run_loop make_script_teller_class run_model_pairs Bot model_to_bot_class get_transcript_results connect_to_redis SceneNearestNeighborTeller SceneNearestNeighborData try_cd try_magic select_clipart_to_tell calc_perplexity episodes_from_fns make_fns Model scripted_tell_before_peek scripted_tell drawer_observe_canvas scripted_tell_after_peek ComponentEvaluator draw_nothing eval_fns torch_load logsumexp Packer load_models make_pairs list maximum set zip zeros array range len list maximum set zip zeros array range len list maximum set zip zeros array range len list maximum set zip zeros array range len decode y get_image_name x open decode sorted SVG svg_from_cliparts display items list torch_load print Path items list torch_load NearestNeighborData print PragmaticNearestNeighborTeller Path items list torch_load print eval Path backward print zero_grad Adam teller parameters get_examples_batch make_fns train step range eval_fns enumerate TellGroup list DrawGroup isinstance scene_similarity ObserveTruth cliparts discount_rewards extend scene where split SetDrawing zeros array append len list asarray zeros_like size reversed range tensor tensors_from_episodes eval torch_load RLSceneToSeqTeller validate clip_grad_norm_ Path save std calc_rl_loss spec datagen disable_dropout TellGroup mean eval item float collect_episodes isinstance array sorted isinstance append keys len TellGroup AbstractScene SelectClipart isinstance ObserveTruth set DrawClipart append ReplyGroup list values TellGroup AbstractScene sorted DrawGroup isinstance ObserveTruth set append ReplyGroup list values TellGroup AbstractScene sorted DrawGroup isinstance ObserveCanvas ObserveTruth set append ReplyGroup list values TellGroup AbstractScene isinstance ObserveCanvas ObserveTruth SetDrawing append ReplyGroup list values get AbstractScene TellGroup isinstance ObserveCanvas ObserveTruth SetDrawing append ReplyGroup list values list values get AbstractScene TellGroup list Peek append TellerObserveCanvas values get AbstractScene TellGroup list Peek append TellerObserveCanvas values AbstractScene list values get AbstractScene list values AbstractScene list values TellGroup event_getter isinstance set msg iter TellGroup event_getter update isinstance Counter msg iter split NUM_IDX NUM_DEPTH cumsum NUM_FLIP NUM_SUBTYPE NUM_IDX NUM_DEPTH NUM_FLIP NUM_SUBTYPE NUM_IDX NUM_DEPTH NUM_FLIP NUM_SUBTYPE NUM_IDX NUM_DEPTH NUM_FLIP NUM_SUBTYPE set partial _trigger_types print array print make_fns eval_fns print make_fns scripted_tell eval_fns print make_fns eval_fns print make_fns scripted_tell eval_fns print_components_script print print_pairwise print_script print_human print_components_pairwise get listen pubsub print loads encode connect_to_redis subscribe get_scenes_and_scripts replace __name__ run_loop DRAWER make_script_teller_class append model_to_bot_class TELLER AbstractScene list values StrictRedis sorted list SelectClipart scene set append append Peek isinstance append Peek isinstance append DrawGroup scene append ReplyGroup DrawGroup append ObserveCanvas reconstruct hasattr Module isinstance extend eval get_action_fns append get_scenes get_scenes_and_scripts_with_peek get_scenes_and_scripts sum datagen_cls exp where sum max log update load_baseline3 load_baseline4 load_baseline1 load_baseline2 append load_models | # CoDraw Models This repository contains models for the Collaborative Drawing (CoDraw) task. ## Installation Dependencies: * Python 3.6 or later * PyTorch 0.4.0 * **IMPORTANT**: Our pre-trained models are **not** compatible with PyTorch 0.4.1. Please use version 0.4.0 exactly. * `Pillow` 5.1.0 (or compatible) * `editdistance` 0.4 (or compatible) You will need to clone both this repository and the [CoDraw dataset repository](https://github.com/facebookresearch/CoDraw) into side-by-side folders. The README for the dataset repository contains additional instructions for downloading required dataset files. | 2,056 |
facebookresearch/consistent_depth | ['depth estimation'] | ['Consistent Video Depth Estimation'] | monodepth/depth_model_registry.py utils/frame_range.py monodepth/mannequin_challenge_model.py utils/consistency.py monodepth/midas_v2_model.py utils/load_colmap.py utils/geometry_np.py params.py optimizer/__init__.py utils/torch_helpers.py video.py depth_fine_tuning.py flow.py optical_flow_flownet2_homography.py process.py utils/image_io.py scale_calibration.py utils/url_helpers.py monodepth/depth_model.py utils/helpers.py main.py loss/parameter_loss.py utils/frame_sampling.py utils/calibration.py tools/make_video.py loss/joint_loss.py loss/loss_params.py monodepth/monodepth2_model.py loss/consistency_loss.py utils/geometry.py utils/calibrate.py loaders/video_dataset.py utils/visualization.py tools/colmap_processor.py DepthFineTuner write_summary DepthFineTuningParams log_loss log_loss_stats make_tag Flow warp_by_flow resize_flow matchKeypoints getimage infer process FlowInfer parse_args detectAndDescribe Video3dParamsParser DatasetProcessor ScaleCalibrationParams make_camera_params_from_colmap visualize_all_calibration calibrate_scale visualize_calibration_pair prepare_colmap_color check_frames sample_pairs Video VideoFrameDataset load_mask VideoDataset load_color load_image load_flow ConsistencyLoss weighted_mse_loss select_tensors weighted_mean_loss weighted_rmse_loss JointLoss LossParams ParameterLoss DepthModel get_depth_model create_depth_model get_depth_model_list MannequinChallengeModel MidasV2Model Monodepth2Model create COLMAPParams COLMAPProcessor main parse_args run augment_args make_video make_resized_filename stack_videos frame_size num_frames make_resized_filename_if_exists MakeVideoParams main parse_args make_depth_videos make_overlay calibrate_scale_shift resize_small calibrate_scale calibrate_scale_shift_RANSAC calibrate cvt_by_scale_shift vote_scale store_visible_points_per_image calibrate_frame_w_sparse_points calibrate_w_sparse_colmap sample consistency_mask sse consistent_flow_masks parse_frame_range OptionalSet FrameRange SamplePairs SamplePairsMode SamplePairsOptions to_in_range principal_point depth_to_points warping_field calibrate_scale pixel_grid pixels_to_points warp_image sample focal_length pixels_to_rays project reproject_points principal_point reproject sample focal_length project print_title mkdir_ifnotexists print_banner dotdict SuppressedStdout ResourceWarning save_raw_float32_image load_image_angle save_depth_map_colored load_image save_image load_raw_float32_image resize_to_target ordered_image_ids to_colmap convert_points3D cameras_to_intrinsics intrinsics_to_camera images_to_extrinsics convert_calibration extrinsics_to_images save_colmap to_device get_model_from_url apply_mask create_video visualize_depth_dir visualize_depth items list add_histogram min mean max add_scalar make_grid to_vis stack unsqueeze add_image log_loss_stats add_scalar sample shape to_tensor pixel_grid SURF_create float32 detectAndCompute DescriptorMatcher_create RANSAC float32 findHomography append knnMatch add_argument ArgumentParser matchKeypoints reshape float inv shape resize imread type warpPerspective detectAndDescribe deepcopy view concatenate size getimage astype float32 dot shape to numpy cat reshape resize resize_flow imwrite device flow_to_image visualize im1 len load_state_dict dirname to im2 pretrained_model_flownet2 save_raw_float32_image size eval zip load print infer FlowNet2 isfile out makedirs format imwrite frame_count print path pjoin check_frames imread range makedirs convert_calibration read_model select_tensor imwrite format vis unsqueeze any warp_image pjoin zip to makedirs visualize_calibration_pair list make_camera_params_from_colmap frames dense_depth_suffix warning resize sorted visualize_all_calibration savetxt print_banner pjoin prepare_colmap_color dirname parse_args process read_array format dense_dir save_raw_float32_image sparse_dir isfinite mean stack isinf float load_raw_float32_image keys colmap_bin_path isdir savez print loadtxt COLMAPProcessor extend path isfile median makedirs sample print name_mode_map reshape transpose post_proc_raw post_proc_other IMREAD_UNCHANGED imread load_raw_float32_image load_image load_image clamp sum view clamp norm view clamp view get_depth_model print join process COLMAPProcessor colmap_bin video3d_dir warning num_frames pjoin out_dir color_dir depth_dirs listdir append split make_resized_filename isfile print make_resized_filename frame_size isfile run imwrite COLOR_BGR2GRAY reshape num_frames dirname resize imread range cvtColor join print extend pjoin isfile append run basename make_overlay make_video stack_videos rmtree pjoin dirname isfile makedirs augment_args zip frame_fmt ffmpeg make_video stack_videos extend frame_size pjoin depth_dirs out_dir color_dir make_depth_videos makedirs resize T isfinite dot sum array reshape RANSACRegressor isfinite fit items list T image_ids convert_points3D append xyz percentile mean logical_and sample reproject project ordered_image_ids print name convert_calibration read_model calibrate_frame_w_sparse_points mean nan pjoin store_visible_points_per_image empty load_raw_float32_image enumerate dtype view grid_sample reshape shape to all reshape logical_and sample stack meshgrid update join sorted parse_sub_range set add_range iter next split auto meshgrid linspace cat shape principal_point view shape focal_length principal_point view pixels_to_rays bmm view transpose baddbmm shape shape pixels_to_points pixel_grid bmm view depth_to_points shape dot depth_to_points project reproject_points permute warping_field sample shape reshape astype all get __delitem__ __setitem__ mkdir exists print len print len int print min resize float round max load_image_angle fromarray save_raw_float32_image astype lower save call save_image Camera astype array dot T enumerate BaseImage extrinsics_to_images intrinsics_to_camera to_colmap write_model append params array concatenate dot T append concatenate ordered_image_ids images_to_extrinsics cameras_to_intrinsics items list isinstance Tensor to enumerate remove rstrip print pjoin dirname isfile sep download exists makedirs amin uint8 amax percentile sorted imwrite print min len isfinite visualize_depth warning splitext append imread listdir max range load_raw_float32_image call reshape array | # [SIGGRAPH 2020] Consistent Video Depth Estimation [](https://colab.research.google.com/drive/1i5_uVHWOJlh2adRFT5BuDhoRftq9Oosx#scrollTo=lNc6HHfHDfnE) ### [[Paper](https://arxiv.org/abs/2004.15021)] [[Project Website](https://roxanneluo.github.io/Consistent-Video-Depth-Estimation/)] [[Google Colab](https://colab.research.google.com/drive/1i5_uVHWOJlh2adRFT5BuDhoRftq9Oosx#scrollTo=lNc6HHfHDfnE)] <p align='center'> <img src="thumbnail.gif" width='100%'/> </p> We present an algorithm for reconstructing dense, geometrically consistent depth for all pixels in a monocular video. We leverage a conventional structure-from-motion reconstruction to establish geometric constraints on pixels in the video. Unlike the ad-hoc priors in classical reconstruction, we use a learning-based prior, i.e., a convolutional neural network trained for single-image depth estimation. At test time, we fine-tune this network to satisfy the geometric constraints of a particular input video, while retaining its ability to synthesize plausible depth details in parts of the video that are less constrained. We show through quantitative validation that our method achieves higher accuracy and a higher degree of geometric consistency than previous monocular reconstruction methods. Visually, our results appear more stable. Our algorithm is able to handle challenging hand-held captured input videos with a moderate degree of dynamic motion. The improved quality of the reconstruction enables several applications, such as scene reconstruction and advanced video-based visual effects. <br/> | 2,057 |
facebookresearch/image-to-set | ['multi label classification'] | ['Elucidating image-to-set prediction: An analysis of models, losses and datasets'] | src/data_loader.py src/utils/voc_utils.py src/utils/tb_visualizer.py src/utils/recipe1m_utils.py src/modules/multihead_attention.py src/args.py src/train.py src/utils/metrics.py src/model.py src/eval.py src/modules/layers.py src/modules/rnn_decoder.py src/modules/encoder.py src/modules/ff_decoder.py src/modules/transformer_decoder.py src/modules/utils.py get_parser set_default_model_flags get_args_from_json set_flags_for_model VOC NUSWIDE RandomSamplerWithState Recipe1M COCO collate_fn ADE20K RandomSamplerWithStateIterator get_loader increase_loader_epoch main SetPred mask_from_eos get_model label2_k_hots predictions_to_idxs set_lr StreamToLogger count_parameters make_dir save_checkpoint main EncoderCNN FFDecoder PositionalEmbedding ConvTBC LinearizedConv1d Embedding LearnedPositionalEmbedding LinearizedConvolution make_positions Linear MultiheadAttention DecoderRNN AttentionLayer LSTMAtt PositionalEmbedding SinusoidalPositionalEmbedding DecoderTransformer TransformerDecoderLayer Embedding LayerNorm Linear convert_state_dict_type clip_grad_norm_ load_model_state print_embed_overlap save_state set_incremental_state _upgrade_state_dict load_ensemble_for_inference parse_embedding move_to_cuda post_process_prediction load_embedding strip_pad buffered_arange load_align_dict _override_model_args item make_positions convert_padding_direction get_incremental_state torch_persistent_save _get_full_incremental_state_key fill_with_neg_inf replace_unk checkpoint_paths DC softIoU DCLoss softIoULoss update_error_counts compute_metrics targetDistLoss Visualizer VOCDetection download_extract load items format print setattr open setattr print setattr set_default_model_flags split use_json_config join get_args_from_json add_argument dataset set_flags_for_model model_name ArgumentParser image_model parse_args set_defaults stack zip VOC join current_epoch NUSWIDE RandomSamplerWithState Recipe1M COCO DataLoader set_state ADE20K SequentialSampler increase_epoch CenterCrop ToTensor models_path DataFrame open save_results_path eval_split len Resize crop_size compute_metrics load_state_dict append to get_vocab format glob Compose mean eval Normalize image_size enumerate load join print to_csv tqdm get_loader zeros get_model makedirs scatter_ unsqueeze array to max size range byte topk byte size clone log sum max range tf_layers maxnumlabels EncoderCNN n_att DCLoss CrossEntropyLoss format DecoderTransformer ff_layers image_model embed_size BCELoss SetPred print dropout_encoder pred_cardinality label_loss DecoderRNN dropout_decoder FFDecoder print rename save isfile state_dict param_groups makedirs perminv metric_to_checkpoint setLevel values seed str addHandler Adam iter StreamHandler info item keys FileHandler lr_decay_every time learning_rate parameters model_name isfile step getLogger zero_grad DataParallel save_checkpoint save_dir decay_lr list tensorboard current_epoch resume image_model INFO Visualizer isinstance train scalar_summary lr_decay_rate model RandomCrop dataset experiment_name sum range LongTensor ERROR increase_loader_epoch num_epochs remove write Formatter reset RandomAffine set_lr RandomHorizontalFlip setFormatter synchronize make_dir close items StreamToLogger backward Tensor normal_ weight constant_ normal_ LearnedPositionalEmbedding weight constant_ bias normal_ weight constant_ bias LinearizedConvolution sqrt normal_ weight constant_ ne arange size new clone unsqueeze expand_as masked_scatter_ xavier_uniform_ SinusoidalPositionalEmbedding range items isinstance OrderedDict is_tensor torch_persistent_save load _upgrade_state_dict load_state_dict upgrade_state_dict max_positions load build_model _override_model_args upgrade_state_dict load_state_dict append _upgrade_state_dict items setattr __name__ _get_full_incremental_state_key _get_full_incremental_state_key isinstance format print set symbols keys len range len get tokenize_line enumerate unk_string replace_unk tokenize string type_as arange LongTensor remainder size eq expand_as sum hasattr norm mul_ item fullmatch append listdir compile enumerate from_numpy unsqueeze float sum array sigmoid sum average download_url | ## Image-to-Set Prediction Companion code for [L. Pineda, A. Salvador, et al.: Elucidating image-to-set prediction: An analysis of models, losses and datasets](https://arxiv.org/abs/1904.05709). This repository contains a unified code-base to train and test strong image-to-set prediction (multi-label classification) baselines. The code comes with pre-defined train/valid/test splits for 5 datasets of increasing complexity ([Pascal VOC 2007](http://host.robots.ox.ac.uk/pascal/VOC/voc2007/), [MS COCO 2014](http://cocodataset.org/#home), [ADE20k](http://groups.csail.mit.edu/vision/datasets/ADE20K/), [NUS-WIDE](https://lms.comp.nus.edu.sg/wp-content/uploads/2019/research/nuswide/NUS-WIDE.html) and [Recipe1M](http://im2recipe.csail.mit.edu/dataset)) as well as a common evaluation protocol to compare all models. The top ranked baselines across datasets are released together with the code. If you find this code useful in your research, please consider citing with the following BibTeX entry: ``` @article{PinedaSalvador2019im2set, author = {Pineda, Luis and Salvador, Amaia and Drozdzal, Michal and Romero, Adriana}, title = {Elucidating image-to-set prediction: An analysis of models, losses and datasets}, journal = {CoRR}, | 2,058 |
facebookresearch/libri-light | ['speech recognition'] | ['Libri-Light: A Benchmark for ASR with Limited or No Supervision'] | data_preparation/split_librilight/split.py data_preparation/build_all_stats.py data_preparation/make_vad_inputs.py data_preparation/rebuild_limited_train/sample_10h.py data_preparation/rebuild_limited_train/clean_texts.py eval/eval_PER.py data_preparation/split_librilight/materialize_split.py data_preparation/metadata_completion/__init__.py data_preparation/split_librilight/extract_test_speakers.py eval/PER_src/simplePhonemLearner.py eval/PER_src/seq_alignment.py data_preparation/metadata_completion/DuplicateSearch.py eval/ABX_src/setup.py data_preparation/text_retrieval/hathitrust.py eval/eval_WER.py data_preparation/rebuild_limited_train/utils.py data_preparation/text_retrieval/main_lesson.py eval/ABX_src/abx_iterators.py data_preparation/metadata_completion/ReaderScapper.py eval/ABX_src/__init__.py data_preparation/split_librilight/prepare_vads_tests.py data_preparation/metadata_completion/utilities.py data_preparation/download_librivox.py data_preparation/complete_metadata.py data_preparation/metadata_completion/GenreScrapper.py data_preparation/calculate_snr.py eval/eval_ABX.py eval/PER_src/setup.py eval/WER_src/simple_dataset.py data_preparation/unzip_and_convert.py eval/PER_src/__init__.py data_preparation/split_librilight/puts_json.py eval/CPC_loader.py data_preparation/text_retrieval/bartleby.py data_preparation/text_retrieval/guttenberg.py data_preparation/unit_tests.py data_preparation/text_retrieval/__init__.py data_preparation/cut_by_vad.py data_preparation/plot.py data_preparation/rebuild_limited_train/split_1h_in10min.py eval/ABX_src/abx_group_computation.py data_preparation/metadata_completion/text_cleaner.py eval/ABX_src/unit_tests.py eval/WER_src/wl_decoder.py data_preparation/metadata_completion/genre_folding.py data_preparation/text_retrieval/archive_org.py data_preparation/split_librilight/prepare_vads.py eval/WER_src/letter_ctc.py data_preparation/rebuild_limited_train/get_stats.py data_preparation/rebuild_limited_train/select_1h.py calculate_snr cal_snr_librivox calculate_file_snr convert_wav_buf_f32 mp_file_snr cal_signal_power main parse_args cut_sequence cut_book save parse_args cut get_reader_data import_page RequestPBar load_tmp get_size_page MyHTMLParser reorder_vad findAllSeqs save_lst get_lst get_file_duration_ms plot_seq plot_pie plot_scatter plot_hist TestCutDB unzip convert _convert_dir getBaseTitle getPossibleMatches getBaseStringData prepareMatches getTitleSimilarityScore get_books_duplicates getSameAuthorGroups GenreScapper getGenreFromMetadata gather_all_genres ReaderScrapper updateDataWithNames update_all_speaker_data get_librivox_reader_from_id loadData find404Error clean_all_text_data get_tag_list get_all_metadata get_hour_tag_repartition save_cache remove_tag strToHours get_speaker_data build_reverse_folding get_metdata_from_id load_cache getTotalTime get_zip_name apply_folding get_speakers combine_reverse_foldings get_speaker_hours_data get_txt_name getJSON get_wav_name get_updated_metadata get_all_speakers remove_multiple_tags get_base_name_from_metadata get_speaker_data_name get_args get_args do_split_10h get_args get_args do_split get_args do_split get_filelength get_speakers get_speaker_fname get_histogram traverse_tree full_records materialize print_stats extract_holdout_speakers get_args _print _apply get_args split_vad process get_args parse_vad TestSplit process_dir get_duplicates get_args read_snr parse_downloaded_jsons get_voice_activities BookError normalize get_stats take_n subselect get_args get_genre2files get_fname2json get_genre2time is_archive_org_url get_archive_id download_text_data get_archive_org_text_data BarthelebyTitleParser BarthelebyParser is_bartheleby_url get_bartheleby_data is_guttenberg_url get_guttenberg_data CatalogParser is_hathitrust_url load_hathitrust_book HathitrustParser load_whole_book get_tag_value_in_url get_all_text_from_main_lesson ToCParser get_full_url ChapterParser is_main_lesson_url get_text_data download_state_dict CPCModel build_feature_from_file FeatureModule ChannelNorm get_features_state_dict CPCAR CPCEncoder load_cpc_features load_pt load_npy find_all_files ABX main reduce_sparse_data parse_args run_training find_all_files get_n_phones per train parse_phone_labels filter_seq val_step set_seed train_step get_eval_args prepare_data eval_wer main Worker run get_kl_distance_batch get_cosine_distance_batch check_dtw_group_validity get_euclidian_distance_batch loc_dtw get_abx_scores_dtw_on_group get_distance_group_dtw get_kl_distance_symmetric_batch get_distance_function_from_name get_theta_group_dtw ABXWithinGroupIterator ABXIterator normalize_with_singularity get_features_group ABXFeatureLoader ABXAcrossGroupIterator load_item_file testGroupMaker testItemLoader testSingularityNormalization testABXFeatureLoader TestDistancesDTW cut_data prepare_data beam_search get_local_per get_seq_PER per_step load valStep CTCPhoneCriterion cutData SingleSequenceDataset prepareData trainStep CTCLetterCriterion LetterClassifier cut_data SingleSequenceDataset parse_ctc_labels_from_root find_seqs WlDecoder astype float32 int sum power float round int zip concatenate print convert_wav_buf_f32 log10 split append power sum range len read fromstring time format print write set imap_unordered Pool len add_argument add_mutually_exclusive_group ArgumentParser clean_all_text_data get_all_metadata get_books_duplicates Path gather_all_genres build_reverse_folding load_cache str combine_reverse_foldings exit parse_args debug update_all_speaker_data mkdir get_updated_metadata path_metadata print in_place print_help out_dir parent stem hstack write mkdir append int read save parent name glob stem cut_sequence Path enumerate glob print get str feed _content MyHTMLParser get_reader_data join urlretrieve RequestPBar json print get_text_data urlopen print int urlopen json load join items print tqdm save append walk len update str ProgressBar start Path finish append get_file_duration_ms enumerate replace unlink Path with_suffix yscale use isinstance xlabel ylabel tight_layout title clf hist histogram savefig sum array xlabel ylabel tight_layout title scatter clf savefig range yscale plot xlabel xscale ylabel tight_layout title clf savefig legend range list use pie tight_layout title clf savefig legend keys join print path_out mkdir path_in run join format path_out mkdir splitext path_in run print map path_out mkdir Pool path_in join int print add set append append join lower split getBaseStringData set intersection union len int isdigit replace lower split len update join getBaseTitle print ProgressBar start finish append getSameAuthorGroups range enumerate len append sort range len print prepareMatches getPossibleMatches get str feed GenreScapper _content update join getGenreFromMetadata print ProgressBar start finish append enumerate get str feed ReaderScrapper _content append deepcopy get_librivox_reader_from_id update join print ProgressBar start updateDataWithNames mkdir abspath finish get_speaker_data_name enumerate len enumerate find update join get_txt_name find404Error print loadData ProgressBar start popen abspath finish append enumerate endswith join walk append update print ProgressBar start Path finish enumerate len print suffix Path fallback_function suffix print save_cache Path map split join strToHours set join get_speaker_data_name set update join get_zip_name replace getJSON length rate ProgressBar start finish info enumerate get_speaker_data_name len update length rate ProgressBar start finish enumerate len update join length rate ProgressBar start finish enumerate len split union set items get sort append split split join enumerate parse_args add_argument ArgumentParser sorted get_histogram shuffle set append float append list shuffle set str list sorted get_filelength resolve rglob append stem split parent FileRecord get_speaker_fname dict append defaultdict lambda_value lambda_key str sorted list move id set add copy Path mkdir book print get_histogram map values Path name stem dst_dir Path mkdir src_dir action print append min range len str parse_vad name p_threshold split_vad len_threshold_frames find str normalize name dict rglob normalize deepcopy parent glob stem add dict BookError zip round print add set defaultdict sum add set defaultdict append defaultdict items print get_genre2time append sum enumerate append get_item files len find join remove download_text_data get_archive_id replace Enum get str int feed BarthelebyTitleParser title loadText _content split get decode HathitrustParser feed get decode CatalogParser candidatesID feed find load_whole_book split split find get str chaptersList get_tag_value_in_url ToCParser get_full_url feed ChapterParser _content is_hathitrust_url is_bartheleby_url is_archive_org_url is_guttenberg_url is_main_lesson_url CPCModel FeatureModule CPCAR CPCEncoder load_state_dict config state_dict cpu size min append cuda cat endswith walk append load load tensor LongTensor print ones get_iterator size _indices get_abx_scores_dtw_on_group ABXFeatureLoader to_dense item reduce_sparse_data sum cuda get_distance_function_from_name symmetric load distance_mode path_data find_all_files path_checkpoint ABX path_item_file file_extension feature_size cuda load_cpc_features split append sort stem str valStep print save trainStep range CTCPhoneCriterion path_phone_converter get_output_dim nEpochs pathCheckpoint DataParallel DataLoader cuda filter_seq Popen list Adam get_n_phones freeze load_cpc_features debug run_training dup2 pathPhone shuffle SingleSequenceDataset file_extension eval mkdir pathDB parse_phone_labels pathVal load join int pathTrain print find_all_files output parameters fileno LSTM len load per_step parent CTCPhoneCriterion find_all_files path_checkpoint get_output_dim pathPhone get_n_phones SingleSequenceDataset eval file_extension DataLoader load_state_dict pathDB parse_phone_labels cuda load_cpc_features view cut_data criterion backward prepare_data zero_grad train step eval update Process join size ProgressBar put start eval finish append range enumerate len val_step train_step print save float range state_dict get join output is_available manual_seed_all manual_seed path_val downsampling_factor get_output_dim DataParallel DataLoader ArgumentParser run seed list set_seed parse_ctc_labels_from_root LetterClassifier path_train Adam lm_weight load_state_dict module eval_wer CTCLetterCriterion SingleSequenceDataset eval n_epochs path_wer join get_eval_args add_argument output parameters size log view size log view size pi acos view size view print size numpy sum size check_dtw_group_validity expand item get_distance_group_dtw range get_theta_group_dtw start ProgressBar finish view size sqrt zeros sum append join len split list sort append range enumerate len max str deepcopy sort add set shape append range enumerate cuda acquire get_seq_PER append array release update print ProgressBar start eval finish enumerate stem view max cutData cuda view optimize criterion model backward zero_grad prepareData train step eval rglob append rglob stem set | # Libri-Light: A Benchmark for ASR with Limited or No Supervision You can track papers that use Libri-Light and their relative performance on Papers With Code: [[test-clean]](https://paperswithcode.com/sota/speech-recognition-on-libri-light-test-clean) [[test-other]](https://paperswithcode.com/sota/speech-recognition-on-libri-light-test-clean) ## Description This repository contains code and models associated with the Libri-Light dataset, which can be [downloaded and prepared here](./data_preparation/README.md). More information about dataset creation and baselines can be found in this [arXiv Paper](https://arxiv.org/abs/1912.07875). Contained here is code for data preparation, pretrained models, and evaluation resources: data_preparation/ # code to download the data; VAD and SNR code; json generation; stats; audio segmentation eval/ # ABX, PER, WER (evaluation metrics on LibriSpeech dev-clean, dev-other, test-clean, test-other) baselines/ # code, pretrained wav2letter models, baselines, and examples To get started, first clone the repository: | 2,059 |
facebookresearch/madgrad | ['stochastic optimization'] | ['Adaptivity without Compromise: A Momentumized, Adaptive, Dual Averaged Gradient Method for Stochastic Optimization'] | madgrad/madgrad.py madgrad/fairseq_madgrad.py docs/source/conf.py setup.py madgrad/__init__.py tests/test_madgrad.py FairseqMADGRAD MADGRAD test_invalid_weight_decay step_test test_momentum_zero test_invalid_eps make_full_precision_params test_invalid_lr test_invalid_momentum set_torch_seed test_sparse test_step_full_precision_inferred manual_seed requires_grad_ requires_grad_ requires_grad_ requires_grad_ print __repr__ item step range requires_grad_ cuda MADGRAD step_test make_full_precision_params MADGRAD step_test make_full_precision_params requires_grad_ MADGRAD step rand_like to_sparse | # MADGRAD Optimization Method A Momentumized, Adaptive, Dual Averaged Gradient Method for Stochastic Optimization Documentation availiable at https://madgrad.readthedocs.io/en/latest/. ``` pip install madgrad ``` Try it out! A best-of-both-worlds optimizer with the generalization performance of SGD and at least as fast convergence as that of Adam, often faster. A drop-in torch.optim implementation `madgrad.MADGRAD` is provided, as well as a FairSeq wrapped instance. For FairSeq, just import madgrad anywhere in your project files and use the `--optimizer madgrad` command line option, together with `--weight-decay`, `--momentum`, and optionally `--madgrad_eps`. The madgrad.py file containing the optimizer can be directly dropped into any PyTorch project if you don't want to install via pip. If you are using fairseq, you need the acompanying fairseq_madgrad.py file as well. ## Things to note: - You may need to use a lower weight decay than you are accustomed to. Often 0. - You should do a full learning rate sweep as the optimal learning rate will be different from SGD or Adam. Best LR values we found were 2.5e-4 for 152 layer PreActResNet on CIFAR10, 0.001 for ResNet-50 on ImageNet, 0.025 for IWSLT14 using `transformer_iwslt_de_en` and 0.005 for RoBERTa training on BookWiki using `BERT_BASE`. On NLP models gradient clipping also helped. # Mirror MADGRAD | 2,060 |
facebookresearch/odin | ['out of distribution detection'] | ['Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networks'] | code/cal.py code/main.py code/densenet.py code/calData.py code/calMetric.py code/wideresnet.py testData testUni testGaussian DenseNet3 TransitionBlock BottleneckBlock DenseBlock BasicBlock main BasicBlock NetworkBlock WideResNet data argmax max cuda open exp net1 add ge sum format float enumerate time criterion backward print Variable write cpu numpy data randn argmax max cuda open exp net1 add ge sum format float enumerate time criterion backward print Variable clamp write cpu numpy data rand argmax max cuda open exp net1 add ge sum format float enumerate time criterion backward print Variable write cpu numpy nn test out_dataset temperature parse_args magnitude gpu | # ODIN: Out-of-Distribution Detector for Neural Networks This is a [PyTorch](http://pytorch.org) implementation for detecting out-of-distribution examples in neural networks. The method is described in the paper [Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networks](https://arxiv.org/abs/1706.02690) by S. Liang, [Yixuan Li](www.yixuanli.net) and [R. Srikant](https://sites.google.com/a/illinois.edu/srikant/). The method reduces the false positive rate from the baseline 34.7% to 4.3% on the DenseNet (applied to CIFAR-10) when the true positive rate is 95%. <p align="center"> <img src="./figures/original_optimal_shade.png" width="500"> </p> ## Experimental Results We used two neural network architectures, [DenseNet-BC](https://arxiv.org/abs/1608.06993) and [Wide ResNet](https://arxiv.org/abs/1605.07146). The PyTorch implementation of [DenseNet-BC](https://arxiv.org/abs/1608.06993) is provided by [Andreas Veit](https://github.com/andreasveit/densenet-pytorch) and [Brandon Amos](https://github.com/bamos/densenet.pytorch). The PyTorch implementation of [Wide ResNet](https://arxiv.org/abs/1605.07146) is provided by [Sergey Zagoruyko](https://github.com/szagoruyko/wide-residual-networks). The experimental results are shown as follows. The definition of each metric can be found in the [paper]().  | 2,061 |
facebookresearch/online_dialog_eval | ['dialogue evaluation'] | ['Learning an Unreferenced Metric for Online Dialogue Evaluation'] | backup/analysis.py backup/plot.py backup/corrupt.py logbook/filesystem_logger.py codes/models.py codes/net.py codes/trainer.py scripts/fix_data.py data.py args.py backup/train.py logbook/logbook.py backup/baselines/ruber/train.py utils.py scripts/compute_corrupt.py backup/ablation_models.py codes/inference.py codes/logger.py backup/baselines/ruber/model.py scripts/extract_task_based_samples.py codes/baseline_models.py backup/models.py scripts/interactive_file.py backup/baselines/ruber/train_word2vec.py get_args context_collate_flat_fn vector_collate_fn Data DialogDataLoader context_collate_nce_fn ParlAIExtractor id_collate_nce_fn id_collate_fn DialogDiskDataLoader id_collate_ruber_fn id_collate_flat_nce_fn id_collate_flat_fn context_collate_flat_nce_fn id_context_collate_fn batch_vectors batch_yhats batch_dialogs BaseTrainer getbuckets batchify cosine_similarity batch_words get_optimizer batch _import_module TransitionRotation TransitionPredictor TransitionPredictorCNN TransitionFn TransitionPredictorIS temporal_cluster semantic_cluster CorruptDialog BaseModel TransitionPredictorMaxPoolLearnedDownsample TransitionPredictorMaxPool scatterPlot plot_tsne DialogMetric UnreferencedMetric get_mlp ReferencedMetric BiLSTMEncoder RUBER CBOWModeler Word2VecTrainer RuberUnreferenced BERTNLI InferSent test_single get_dial_tokens test_context_corruption test_trainer WandbLogger TransitionPredictorMaxPool Net main optimize_on_cluster build_model write_metric_logs add_time_to_log format_custom_logs set_logger write_trajectory_logs write_message_logs pprint write_log write_config_log flatten_dict format_log get_logger read_log write_metadata_logs flatten_dict LogBook prepare_corruptions generateSample process_data get_long_sents buffered_read make_batches cli_main main get_num_lines HyperOptArgumentParser add_argument opt_list batchify zip batch_dialogs from_numpy batchify zip batch_yhats batch_dialogs from_numpy zip batchify from_numpy batch_words batchify zip batch_yhats from_numpy zip batch_words batchify batch_yhats batch_dialogs from_numpy zip batchify from_numpy batch_words batchify zip batch_yhats from_numpy zip batch_words batchify from_numpy batch_words batchify zip batch_dialogs from_numpy batchify zip int range len Adagrad Adadelta Adamax Adam RMSprop SGD Rprop float ASGD split max LongTensor zeros long enumerate long LongTensor enumerate size unsqueeze zeros max enumerate len long enumerate len rsplit import_module getattr int dial_vecs transform print LinearDiscriminantAnalysis scatterPlot getbuckets shape stack append float range len list permutations format euclidean append std print KMeans extend cityblock mean split cosine_similarity range array dial_vecs fit_predict enumerate format rainbow shape scatter savefig mkdir linspace sample max range TSNE KMeans max shape scatter savefig append fit_transform range format annotate sample fit_predict enumerate int print PCA extend figure transform array len BatchNorm1d Linear ReLU append range Dropout test Trainer model bert_input iterrows batch_dialogs convert_tokens_to_ids from_numpy append batch_words get_dataloader update format debug close mean build_inputs_with_special_tokens item sample tokenize batchify is_transition_fn print to_csv tqdm std read_csv split update dump next_corrupt_context_model dialogs convert_tokens_to_ids len get_dial_tokens CorruptDialog close tqdm open append forward tokenize enumerate id_context_collate_fn per_experiment_nb_gpus add_command per_experiment_nb_cpus SlurmCluster nb_gpu_nodes gpu_type job_time optimize_parallel_cluster_gpu add_slurm_cmd LogBook ParlAIExtractor InferSent print RuberUnreferenced vocab_size vars BERTNLI write_metadata_logs TransitionPredictorMaxPool int build_model test Trainer sleep use_cluster TestTubeLogger fit MutableMapping items isinstance extend append add_time_to_log info strftime loads format_custom_logs write_log format_custom_logs write_log format_custom_logs to_json loads write_log format_custom_logs write_log format_custom_logs items write_log print dumps setFormatter getLogger addHandler StreamHandler Formatter setLevel INFO FileHandler from_pretrained DataFrame list iterrows uniform append corrupt_pre format choice unique sample enumerate join remove print to_csv index split read_csv len join deepcopy replace writerow readlines sent_tokenize open split from_pretrained iterrows build_inputs_with_special_tokens append split format print len to_csv get_long_sents drop next_epoch_itr LongTensor setup_task import_user_module make_generation_fast_ cuda half build_bpe load_model_ensemble buffer_size input format close resolve_max_positions load_align_dict build_tokenizer fp16 print target_dictionary build_generator get_num_lines tqdm path source_dictionary replace_unk max_positions split main get_generation_parser parse_args_and_arch | # MaUde **M**etric for **a**utomatic **U**nreferenced **d**ialog **e**valuation. Contains code of the paper titled _"Learning an Unreferenced Metric for Online Dialogue Evaluation"_ to appear at **ACL 2020**, [Arxiv](https://arxiv.org/abs/2005.00583) ## Installation - `pip install -r requirements.txt` - Install [ParlAI](https://github.com/facebookresearch/ParlAI#installing-parlai) ## Getting the data - Get the `convai2` train and test data and pre-trained Distilbert [embeddings here](https://drive.google.com/file/d/1VVcsxmUrDSRIfunPWe9UO1aeCz-lITNy/view?usp=sharing). Download and unzip in the folder `convai2_data`. - Get the trained model checkpoints [from here](https://drive.google.com/file/d/1Ysso9hdzSenK13LjOFombyXYqA_kv-Vy/view?usp=sharing). Download and unzip into the folder `full_acl_runs`. - For individual licensing reasons we cannot release the train/test data of MultiWoz, Frames and DailyDialog. Please [send me a mail](mailto:[email protected]) if you need them! | 2,062 |
facebookresearch/open_lth | ['network pruning', 'style transfer'] | ['The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks', 'Training BatchNorm and Only BatchNorm: On the Expressive Power of Random Features in CNNs'] | testing/test_case.py utils/test/test_tensor_utils.py lottery/branch/randomly_reinitialize.py datasets/test/test_registry.py pruning/registry.py lottery/test/__init__.py models/cifar_resnet.py models/initializers.py lottery/branch/desc.py models/test/test_freezing.py lottery/branch/__init__.py pruning/test/test_sparse_global.py training/test/test_metric_logger.py models/test/__init__.py models/bn_initializers.py open_lth.py models/__init__.py pruning/sparse_global.py training/runner.py utils/__init__.py lottery/branch/runner.py models/test/test_cifar_resnet.py foundations/test/test_step.py cli/shared_args.py pruning/__init__.py testing/toy_model.py models/test/test_cifar_vgg.py foundations/test/test_hparams.py lottery/branch/registry.py models/test/test_mnist_lenet.py training/test/test_standard_callbacks.py lottery/__init__.py testing/__init__.py cli/__init__.py platforms/local.py utils/tensor_utils.py foundations/hparams.py models/registry.py datasets/test/test_imagedataset_and_dataloader.py foundations/desc.py foundations/step.py utils/test/__init__.py lottery/desc.py models/cifar_vgg.py cli/arg_utils.py training/desc.py lottery/branch/retrain.py testing/test/test_toy_model.py datasets/test/test_cifar10.py datasets/mnist.py lottery/runner.py foundations/test/__init__.py models/base.py platforms/platform.py datasets/__init__.py training/__init__.py pruning/pruned_model.py training/test/test_checkpointing.py training/checkpointing.py training/test/test_train.py platforms/registry.py __init__.py testing/test/__init__.py datasets/test/__init__.py datasets/base.py lottery/branch/base.py datasets/cifar10.py platforms/base.py datasets/imagenet.py training/optimizers.py datasets/registry.py cli/runner_registry.py lottery/test/test_runner.py pruning/test/test_mask.py training/test/test_optimizers.py foundations/paths.py training/test/__init__.py lottery/test/test_desc.py lottery/branch/randomly_prune.py training/metric_logger.py models/imagenet_resnet.py pruning/test/__init__.py models/mnist_lenet.py training/test/test_lr_scheduler.py models/test/test_save_load_exists.py pruning/base.py pruning/test/test_pruned_model.py datasets/test/test_mnist.py training/standard_callbacks.py pruning/mask.py foundations/runner.py foundations/__init__.py training/train.py main maybe_get_arg get JobArgs maybe_get_default_hparams ShuffleSampler ImageDataset DataLoader DistributedShuffleSampler Dataset Dataset CIFAR10 _get_samples Dataset Dataset get num_classes iterations_per_epoch TestDataset TestImageDatasetAndDataLoader TestDataset TestRegistry Desc PruningHparams DatasetHparams TrainingHparams Hparams ModelHparams hparams logger model mask sparsity_report checkpoint Runner Step TestHparams TestStep LotteryDesc LotteryRunner Branch make_BranchDesc Branch Branch get Branch BranchRunner TestDesc TestRunner Model DistributedDataParallel DataParallel fixed uniform oneone positivenegative Model Model ResNet Model binary kaiming_uniform orthogonal kaiming_normal Model get load exists get_default_hparams TestCifarResNet TestCifarVGG TestFreezing TestMnistLenet TestSaveLoadExists Platform Platform get_platform get Strategy Mask PrunedModel get get_pruning_hparams Strategy PruningHparams TestMask TestPrunedModel TestSparseGlobal Platform main TestCase InnerProductModel TestInnerProductModel restore_checkpoint save_checkpoint_callback TrainingDesc MetricLogger get_optimizer get_lr_schedule TrainingRunner run_at_step save_model run_every_epoch create_timekeeper_callback run_every_step save_logger standard_callbacks create_eval_callback standard_train train TestCheckpointing TestLrScheduler TestMetricLogger TestOptimizers TestStandardCallbacks TestTrain shuffle_state_dict unvectorize shuffle_tensor vectorize perm TestTensorUtils display_output_location items print add_argument exit create_from_args add_args ArgumentParser run parse_args run_job maybe_get_arg add_argument range ArgumentParser maybe_get_arg join blur subsample get_test_set get_train_set randomize_labels unsupervised_rotation num_train_examples data zeros_like isinstance rand shape BatchNorm2d data ones_like zeros_like isinstance BatchNorm2d data BatchNorm2d ones_like isinstance data zeros_like isinstance uniform BatchNorm2d data sign weight kaiming_normal_ std kaiming_normal_ weight kaiming_uniform_ weight weight orthogonal_ named_modules hasattr isinstance get_model_from_name batchnorm_init named_parameters others_frozen_exceptions getattr model_name BatchNorm2d model_init is_valid_model_name split get load_state_dict load_model model default_hparams is_valid_model_name is_primary_process save_model barrier checkpoint load_model create_from_string from_epoch is_parallel load_state_dict checkpoint append iteration warmup_steps milestone_steps save save training_steps run_every_epoch create_timekeeper_callback iterations_per_epoch append from_str create_eval_callback callback model zero_grad DataParallel DistributedDataParallel apex_fp16 restore_checkpoint torch_device get_optimizer initialize to range data_order_seed from_epoch is_distributed iterations_per_epoch shuffle get_lr_schedule enumerate backward barrier is_parallel ep loss_criterion step makedirs get training_steps iterations_per_epoch standard_callbacks from_str train reshape sorted keys normal ones Generator manual_seed zeros shape sorted keys enumerate shuffle_tensor | # OpenLTH: A Framework for Lottery Tickets and Beyond
### Welcome
This framework implements key experiments from recent work on the lottery ticket hypothesis and the science of deep learning:
* [_The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks._](https://openreview.net/forum?id=rJl-b3RcF7) Jonathan Frankle & Michael Carbin. ICLR 2019.
* [_Stabilizing the LTH/The LTH at Scale._](https://arxiv.org/abs/1903.01611) Jonathan Frankle, Gintare Karolina Dziugaite, Daniel M. Roy, & Michael Carbin. Arxiv.
* [_Linear Mode Connectivity and the LTH._](https://arxiv.org/abs/1912.05671) Jonathan Frankle, Gintare Karolina Dziugaite, Daniel M. Roy, & Michael Carbin. Arxiv.
* [_The Early Phase of Neural Network Training._](https://openreview.net/forum?id=Hkl1iRNFwS) Jonathan Frankle, David J. Schwab, and Ari S. Morcos. ICLR 2020.
| 2,063 |
facebookresearch/qhoptim | ['stochastic optimization'] | ['Quasi-hyperbolic momentum and Adam for deep learning'] | qhoptim/common/__init__.py qhoptim/pyt/qhm.py test_qhoptim/tf/util.py test_qhoptim/pyt/__init__.py qhoptim/tf/util.py qhoptim/pyt/__init__.py test_qhoptim/pyt/test_qhadam.py qhoptim/tf/__init__.py test_qhoptim/tf/test_qhm.py qhoptim/tf/qhm.py qhoptim/__init__.py test_qhoptim/pyt/util.py test_qhoptim/tf/__init__.py test_qhoptim/tf/test_qhadam.py test_qhoptim/__init__.py qhoptim/tf/qhadam.py docs_src/conf.py test_qhoptim/pyt/test_qhm.py qhoptim/version.py qhoptim/pyt/qhadam.py setup.py qhoptim/common/param_conv.py setup from_accsgd from_synthesized_nesterov from_nadam from_robust_momentum from_pid from_two_state_optimizer QHAdamW QHAdam QHM QHAdamOptimizer QHMOptimizer call_if_callable test_adam_equiv test_nesterov_equiv test_momentum_equiv test_plain_sgd_equiv assert_optimizers_equal test_adam_equiv test_momentum_equiv test_plain_sgd_equiv assert_optimizers_equal allclose build_net add_js_file sqrt sqrt assert_optimizers_equal assert_optimizers_equal assert_optimizers_equal assert_optimizers_equal deepcopy reference_optim_ctor test_m view randn backward mm zero_grad step test_optim_ctor parameters mean double range reference_m normal list Random Variable shuffle gather range normal minimize allclose float64 placeholder matmul reduce_mean build_net ConfigProto | facebookresearch/qhoptim | 2,064 |
facebookresearch/simmc | ['response generation'] | ['Situated and Interactive Multimodal Conversations'] | mm_dst/gpt2_dst/scripts/run_generation.py mm_action_prediction/loaders/loader_vocabulary.py mm_action_prediction/tools/extract_actions_fashion.py mm_dst/gpt2_dst/utils/convert.py mm_action_prediction/models/encoders/history_agnostic.py mm_action_prediction/eval_simmc_agent.py mm_action_prediction/tools/response_evaluation.py mm_action_prediction/models/encoders/__init__.py mm_dst/gpt2_dst/scripts/preprocess_input.py mm_action_prediction/models/__init__.py mm_action_prediction/models/encoders/hierarchical_recurrent.py mm_action_prediction/models/action_executor.py mm_dst/utils/evaluate_dst.py mm_action_prediction/tools/embed_furniture_assets.py mm_dst/gpt2_dst/scripts/evaluate.py mm_action_prediction/tools/build_multimodal_inputs.py mm_action_prediction/tools/weight_init.py mm_action_prediction/models/encoders/tf_idf_encoder.py mm_action_prediction/models/self_attention.py mm_action_prediction/models/user_memory_embedder.py mm_action_prediction/tools/embed_fashion_assets.py mm_action_prediction/models/encoders/memory_network.py mm_action_prediction/loaders/loader_simmc.py mm_action_prediction/loaders/__init__.py mm_action_prediction/tools/retrieval_evaluation.py mm_action_prediction/train_simmc_agent.py mm_action_prediction/models/assistant.py mm_action_prediction/tools/extract_attribute_vocabulary.py mm_action_prediction/loaders/loader_base.py mm_action_prediction/tools/action_evaluation.py mm_action_prediction/tools/extract_actions.py mm_action_prediction/models/carousel_embedder.py mm_action_prediction/tools/rnn_support.py mm_action_prediction/tools/extract_vocabulary.py mm_dst/gpt2_dst/scripts/run_language_modeling.py mm_action_prediction/models/positional_encoding.py mm_action_prediction/tools/support.py mm_action_prediction/tools/data_support.py mm_action_prediction/models/decoder.py mm_action_prediction/tools/torch_support.py mm_action_prediction/options.py main evaluate_agent read_command_line LoaderParent DataloaderSIMMC Vocabulary ActionExecutor Assistant CarouselEmbedder GenerativeDecoder PositionalEncoding SelfAttention UserMemoryEmbedder HierarchicalRecurrentEncoder HistoryAgnosticEncoder MemoryNetworkEncoder TFIDFEncoder register_encoder main evaluate_action_prediction get_save_path convert_pool_matrices build_multimodal_inputs convert_pool_matrices_pretrained_tokenizer main get_intents FurnitureDatabase read_furniture_metadata pretty_print_dict setup_cuda_environment get_object_references ExponentialSmoothing sort_eval_metrics main main get_roundwise_dialog_actions update_carousel_state get_keystrokes_with_args get_carousel_prefabs get_turn_keystrokes examine_get_info_action gen_getinfo_from_annotation is_prefab_in_focus get_prefab_in_focus get_viewed_text_actions extract_actions get_args_for_furniture_click main collate_and_insert_actions get_relevant_actions gen_addtocart_from_annotation get_carousel_state matching_carousels extract_info_attributes main extract_actions extract_action_attributes print_fashion_attributes main main evaluate_response_generation normalize_sentence main evaluate_response_retrieval dynamic_rnn OutputForm rearrange get_sorted_order print_distribution pretty_print_dict setup_cuda_environment ExponentialSmoothing extract_split_from_filename sort_eval_metrics flatten gather_states unflatten weight_init set_seed prepare_xlm_input prepare_transfoxl_input prepare_ctrl_input adjust_length_to_model main prepare_xlnet_input TextDataset set_seed evaluate LineByLineTextDataset train _sorted_checkpoints mask_tokens main _rotate_checkpoints load_and_cache_examples convert_json_to_flattened parse_flattened_result represent_visual_objects parse_flattened_results_from_file add_dicts evaluate_from_flat_list evaluate_frame d_f1 evaluate_turn b_stderr b_arr update load format replace evaluate_agent print pretty_print_dict load_state_dict Assistant DataloaderSIMMC get_data_related_arguments get int update format exp replace print evaluate_response_retrieval evaluate_action_prediction num_instances train sum evaluate_response_generation print add_argument pretty_print_dict setup_cuda_environment ArgumentParser vars cuda parse_args sorted defaultdict format isinstance print union mean unique zip append zeros keys enumerate len evaluate_action_prediction from_pretrained get_candidate_ids strip retrieval_candidate_file max defaultdict array convert_pool_matrices_pretrained_tokenizer append range update format print_distribution concatenate action_json_path fill enumerate pad_token_id vocab_file print convert_pool_matrices int32 zeros full len sorted isinstance progressbar astype array fill max enumerate len sorted pad_token_id isinstance progressbar astype array fill max enumerate len join pretrained_tokenizer replace get_save_path save_path json_path build_multimodal_inputs save array print str format sorted format print keys max items sorted print int format enumerate literal_eval get int items vector concatenate size mean stack append zeros strip read_furniture_metadata lower loads append isclose startswith insert list reversed enumerate is_prefab_in_focus get_carousel_prefabs list get_carousel_prefabs reversed lower startswith append search_furniture enumerate word_tokenize defaultdict set append len deepcopy join format get_roundwise_dialog_actions replace get_keystrokes_with_args print get_turn_keystrokes extend get_viewed_text_actions gen_getinfo_from_annotation collate_and_insert_actions get_relevant_actions gen_addtocart_from_annotation get_min_max_price_per_class append get_intents deepcopy extend set add any ASSISTANT items get_object_references match any append compile get_intents get_object_references items ASSISTANT deepcopy index append update_carousel_state get_carousel_state enumerate deepcopy get_carousel_state all len metadata_path FurnitureDatabase save_root subtask extract_actions shutdown get int save_root literal_eval extract_info_attributes len extend literal_eval any append enumerate len get items defaultdict format isinstance print tuple shape range items sorted format print keys sorted len format SmoothingFunction print sqrt sentence_bleu append std normalize_sentence len evaluate_response_generation format print size len append sum array enumerate evaluate_response_retrieval sort append index_select get_sorted_order pack_padded_sequence pad_packed_sequence index_select rnn_model items sorted format print float sum max values shape reshape shape reshape data children constant_ ConvTranspose3d Embedding BatchNorm3d Conv3d normal_ xavier_normal_ BatchNorm1d GRUCell GRU BatchNorm2d ConvTranspose1d LSTMCell Conv1d ConvTranspose2d Linear isinstance orthogonal_ Conv2d parameters LSTM seed manual_seed_all manual_seed encode info str list input xlm_language keys from_pretrained decode prepare_input ArgumentParser device squeeze_ set_seed length tolist model_type dirname generate encode parse_args to path_output prompts_from_file adjust_length_to_model info enumerate join model_name_or_path add_argument makedirs TextDataset examples line_by_line LineByLineTextDataset print per_gpu_train_batch_size len join sorted format glob match output_dir append format save_total_limit rmtree _sorted_checkpoints info max len pad_token_id bool mask_token convert_tokens_to_ids clone randint shape masked_fill_ eq tensor mlm_probability full len gradient_accumulation_steps resize_token_embeddings get_linear_schedule_with_warmup clip_grad_norm_ zero_grad DataLoader DataParallel DistributedDataParallel max_grad_norm output_dir device save max initialize set_seed logging_steps load_state_dict master_params to state_dict SummaryWriter format _rotate_checkpoints close mean save_pretrained num_train_epochs info fp16 trange per_gpu_train_batch_size max_steps enumerate load join int n_gpu items evaluate model_name_or_path AdamW backward add_scalar makedirs tqdm parameters step train_batch_size len DataLoader DataParallel output_dir device tensor max eval_batch_size exp per_gpu_eval_batch_size to SequentialSampler format eval info join n_gpu makedirs tqdm load_and_cache_examples len enable_attach config_name should_continue resize_token_embeddings block_size warning do_train output_dir eval_all_checkpoints setLevel basicConfig add_special_tokens set_device _sorted_checkpoints WARN init_process_group tokenizer_name save_pretrained fp16 from_config wait_for_attach train n_gpu evaluate min barrier max_len bool load_and_cache_examples local_rank join format strip makedirs set add dirname append enumerate get items join append strip group append finditer compile split add_dicts d_f1 evaluate_turn range len add_dicts range evaluate_frame len set b_stderr zeros int | # Situated Interactive MultiModal Conversations (SIMMC) Challenge 2020 Welcome to the Situated Interactive Multimodal Conversations (SIMMC) Track for [DSTC9][dstc9] 2020. The SIMMC challenge aims to lay the foundations for the real-world assistant agents that can handle multimodal inputs, and perform multimodal actions. We thus focus on **task-oriented** dialogs that encompass a **situated** multimodal user context in the form of a co-observed image or virtual reality (VR) environment. The context is **dynamically** updated on each turn based on the user input and the assistant action. Our challenge focuses on our SIMMC datasets, both of which are shopping domains: (a) furniture (grounded in a shared virtual environment) and, (b) fashion (grounded in an evolving set of images). **Organizers**: Ahmad Beirami, Eunjoon Cho, Paul A. Crook, Ankita De, Alborz Geramifard, Satwik Kottur, Seungwhan Moon, Shivani Poddar, Rajen Subba <figure> | 2,065 |
facebookresearch/unlikelihood_training | ['text generation'] | ['Neural Text Generation with Unlikelihood Training'] | custom/metrics.py custom/sequence_penalty_loss.py custom/sequence_generator.py custom/__init__.py custom/evaluation.py custom/evaluate_utils.py custom/baseline_cross_entropy.py custom/transformer_arch.py custom/candidate_penalty_ce_loss.py custom/report_metrics.py custom/gpt2/run_gpt2.py custom/language_modeling_with_generation.py CrossEntropyCriterionWCustomMetrics CandidatePenaltyCrossEntropyCriterion load eval_single_token_prediction generate_completions batch_input_sequence_by_prefix_length main LanguageModelingWithGenerationTask Metrics TrainingMetrics ngram_metrics merge_dicts get_dataframe_for_model print_metrics find_tuple process_file main process_files get_metric SequenceGenerator top_k_logits SequencePenaltyCriterion ngram_repeat_mask transformer_lm_baevski_wiki103 sample_sequence ngram_repeat_mask mle_loss get_datasets ul_seq eval_singletoken top_k_top_p_filtering save_completion_metrics main save_singletoken_metrics get_text_continuation tokenize seed setup_task load_model_ensemble format print build_generator import_user_module half load_dataset path make_generation_fast_ next_epoch_itr manual_seed fp16 cuda gen_subset split update batch_input_sequence_by_prefix_length generate_completion tolist extend tqdm report generate Metrics cuda enumerate size contiguous model pad_index log get_normalized_probs view tolist move_to_cuda ranking_metrics numel append sum strip_pad size nll_loss item float aggregate_and_normalize enumerate tqdm pow ckpt parse_args_and_arch ckpt_epoch ArgumentParser batch_size_completions ckpt_step open topk basename topp data_dir len dictionary batch_size_single_prediction eval_single_token_prediction normpath beam_ngram_block singletoken_topk parse_args SequenceGenerator format dump data_prefix_length glob save_path singletoken_topp base_dir model_path enumerate load join completion_length beam_size items print target_dictionary data_split add_argument generate_completions get_generation_parser split extend defaultdict range Counter len update merge_dicts enumerate process_file tqdm tuple get_metric items get join int update merge_dicts load readlines len extend set report open Metrics enumerate append split append items format isinstance append items find_tuple DataFrame model_names set_option extend tqdm eval_dirs process_files expand_as zeros_like tuple tolist add set range enumerate len transformer_lm_big getattr cumsum sort size squeeze min clone softmax load items view size TensorDataset model multinomial stack top_k_top_p_filtering append argmax range cat model log_softmax nll_loss ranking_metrics numel item cuda batch_input_sequence_by_prefix_length top_k gather cuda defaultdict view tolist numel continuation_length sum log_softmax size prefix_length top_p ngram_metrics items sample_sequence contiguous type_as decode join tokenize replace join dump format print output_dir open join dump format print open aggregate_logging_outputs get_datasets set save_singletoken_metrics eval DataLoader SequentialSampler to_dict len from_pretrained gradient_accumulation_steps zero_grad DataLoader device output_dir token_loss save seed save_vocabulary list get_datasets device_count to_json_file append to SequentialSampler save_singletoken_metrics state_dict manual_seed_all aggregate_logging_outputs eval num_train_epochs info manual_seed trange max_steps backward AdamW RandomSampler named_parameters ul_seq eval_singletoken model_name WarmupLinearSchedule model_load_dir step to_dict makedirs | # Neural Text ~~de~~Generation with Unlikelihood Training PyTorch implementation of the paper: [Neural Text Generation with Unlikelihood Training](https://arxiv.org/pdf/1908.04319.pdf)\ Sean Welleck\*, Ilia Kulikov\*, Stephen Roller, Emily Dinan, Kyunghyun Cho, Jason Weston\ \*Equal contribution. The order was decided by a coin flip. We present code for training models described in the paper, as well as pre-trained models. The code includes: - An **implementation of unlikelihood training, fine-tuning, and evaluation** for [fairseq](https://github.com/pytorch/fairseq). - A script for **fine-tuning a GPT-2 model** from [pytorch-transformers](https://github.com/huggingface/pytorch-transformers) with the unlikelihood sequence loss. | Table of Contents | |-| | 2,066 |
faisal-iut/linkPrediction | ['information retrieval', 'link prediction'] | ['Mining Temporal Evolution of Knowledge Graph and Genealogical Features for Literature-based Discovery Prediction'] | versions.py feature_selection.py documentCentrality.py PathSim.py graph.py utils.py classification.py non_edge_feature_dataframe print_attributes classification_train_data_static classification_test_data_static reshape_feature_data_for_classification model_evaluate classification_model classification_train_data_dynamic document_centrality feature_y_weight dynamic_graph_feature_set node_and_article_feature train_data_frame_dynamic feature_node_type parent_key_from_parent_graph citation_feature_build build_feature_set drop_feature_columns feature_citation graph_norm build_graph static_graph_build graph_load reverseTuple load_graph save_graph nodes_intersection node_key_find node_label_find dynamic_train_test_graph_build commute_matrix scaling lighten_color save_data min_max_norm shuffling keyword_split arrayToList load_data scale load_dataset list sorted tuple non_edges append DataFrame sorted list tuple non_edges append DataFrame DataFrame update print_attributes print classification_train_data_static classification_test_data_static shuffling set classification_train_data_dynamic range len print len asarray print shape zeros range enumerate len Sequential fit Adam add Dense LSTM BatchNormalization compile Dropout evaluate roc_curve precision_recall_curve predict auc print transpose log dot shape nanmean nanstd nan zeros sum array range document_centrality list nodes set closeness_centrality append zeros DataFrame len node_and_article_feature apply len update set difference intersection all_neighbors len reset_index min_max_norm train_data_frame_dynamic nodes feature_node_type apply parent_key_from_parent_graph set build_feature_set sort_values intersection power sum range preferential_attachment list adamic_adar_index apply dict resource_allocation_index zip reset_index min_max_norm apply set build_feature_set sort_values range drop norm enumerate has_node append iterrows node_key_find update combinations iterrows list add_edge add set node_key_find has_node add_node write_gpickle read_gpickle str join build_graph static_graph_build print Graph save_graph nodes_intersection nodes edges range len str join build_graph Graph print save_graph nodes edges len str join print load_graph nodes edges range len open str list nodes apply edges intersection range update add_edges_from reverseTuple Graph close copy set remove_edges_from remove_nodes_from load print difference median len pop insert transpose tolist index dot reindex drop list Series diagonal append DataFrame split range read_csv keyword_split len min max any sample concat reset_index len rgb_to_hls print str join print str join min max | # Link Prediction using metadata information Performs scientific literature keyword-keyword co-occurrence prediction based on associated metadata based features. ## Getting Started ### Prerequisites * [NetworkX 2.2](https://networkx.github.io/) * [Keras](https://keras.io/) * theano * [statsmodel](https://www.statsmodels.org/stable/index.html) * scipy * numpy | 2,067 |
fajri91/NeuralRST | ['discourse parsing', 'word embeddings'] | ['Transition-based Neural RST Parsing with Implicit Syntax Features'] | in_out/tree.py in_out/instance.py models/alphabet.py modules/function_variational_rnn.py in_out/reader.py models/vocab.py modules/variational_rnn.py models/config.py models/metric.py models/explorer.py modules/layer.py in_out/preprocess.py transition/atom_feature.py transition/state.py models/architecture.py modules/embedding.py train_rst_parser.py transition/action.py in_out/util.py main set_label_action EDU SubTree Instance SynFeat CResult get_max_parameter validate_gold_actions batch_data_variable construct_embedding_table create_alphabet modTupByIndex Reader Tree load_embedding_dict get_logger lower_with_digit_transform Alphabet MainArchitecture Config Explorer Metric Vocab Embedding assign_tensor StackedStep VarRNNTanhCell AutogradVarMaskedRNN VarGRUCell VarFastGRUCell VarMaskedRecurrent AutogradVarMaskedStep VarMaskedStep StackedRNN VarRNNReLUCell VarLSTMCell VarFastLSTMCell get_tensor_np Biaffine NonLinear AvgPooling MyLSTM orthonormal_initializer LSTM VarMaskedRNNBase VarGRUCell VarFastGRUCell VarRNNCell default_initializer VarMaskedLSTM VarRNNCellBase VarMaskedRNN VarMaskedFastLSTM VarLSTMCell VarMaskedGRU VarFastLSTMCell VarMaskedFastGRU CAction CNode AtomFeat CState set_label_id range gold_actions len ArgumentParser test_path save clip str Reader train_syn_feat_path create_alphabet alphabet_path parse_args manual_seed_all print_metric alpha2id get_subtrees dev_syn_feat_path info manual_seed clip_grad_norm word_embedding join time train_path evaluate makedirs now test_syn_feat_path parameters model_name step zero_grad set_label_action word_embedding_file get_logger load_embedding_dict max_state_size eval long use_dynamic_oracle dev_path divmod add_argument train array Metric max_edu_size cuda opt validate_gold_actions batch_data_variable exit construct_embedding_table word_dim Vocab range get_f_measure state_dict use_gpu max_sent_size seconds set_num_threads reset tag_dim Config batch_size generate_optimizer MainArchitecture max_iter read_data freeze format start_dynamic_oracle lr model_path backward etype_dim loss len items size astype float32 sqrt empty load get str get_str info print size strip etype lower_with_digit_transform edus gold_actions Alphabet label total_words save range len str print is_shift Metric is_reduce gold_actions words edus gold_actions len cuda get_str FloatTensor ones tolist from_numpy edus range use_gpu size words etype max_state_size word2id zero_ gold_actions type Variable zeros len list load vector_size print dict shape load_word2vec_format empty enumerate open setFormatter str getLogger addHandler StreamHandler Formatter isoformat setLevel INFO FileHandler data isinstance linear relu linear tanh sigmoid tanh baddbmm unsqueeze tanh linear chunk apply sigmoid is_cuda sigmoid tanh baddbmm unsqueeze tanh linear chunk apply sigmoid is_cuda len StackedRNN StackedStep VarMaskedStep randn print dot sqrt eye sum range transpose sum sqrt | # NeuralRST This is the reimplementation of Transition-based Neural RST Parsing with Implicit Syntax Features, COLING 2018 (https://www.aclweb.org/anthology/C18-1047/) in Python using Pytorch. The original code can be accessed here: https://github.com/yunan4nlp/NNDisParser and it was implemented in C++ using N3LDG framework. ## Dependencies 1. Python 2.7 2. Run `pip install -r requirements.txt` ## Data ## RST Tree Bank.</br> *https://catalog.ldc.upenn.edu/LDC2002T07* ## External Resource ## Pretrained word embeddings.</br> | 2,068 |
falreis/hed-rcf-cuda | ['boundary detection', 'edge detection'] | ['Holistically-Nested Edge Detection'] | python/caffe/io.py python/caffe/test/test_python_layer.py scripts/download_model_binary.py python/caffe/net_spec.py examples/rcf/rcf_solve.py examples/hed/hed_solve.py python/caffe/test/test_net.py tools/extra/resize_and_crop_images.py python/draw_net.py python/caffe/test/test_net_spec.py src/caffe/test/test_data/generate_sample_data.py python/caffe/draw.py python/caffe/pycaffe.py tools/extra/extract_seconds.py scripts/cpp_lint.py python/classify.py python/caffe/proto/caffe_pb2.py python/caffe/test/test_solver.py python/caffe/classifier.py python/caffe/test/test_python_layer_with_param_str.py tools/extra/parse_log.py python/caffe/__init__.py python/caffe/test/test_layer_type_list.py scripts/copy_notebook.py python/caffe/detector.py python/detect.py interp_surgery upsample_filt main main main parse_args Classifier Detector get_edge_label draw_net get_layer_label get_pydot_graph choose_color_by_layertype get_pooling_types_dict draw_net_to_file Transformer blobproto_to_array datum_to_array array_to_blobproto arraylist_to_blobprotovecor_str array_to_datum resize_image blobprotovector_str_to_arraylist load_image oversample Layers Function Parameters Top NetSpec assign_proto param_name_dict to_proto _Net_blobs _Net_forward_all _Net_set_input_arrays _Net_backward _Net_params _Net_forward _Net_outputs _Net_forward_backward_all _Net_blob_loss_weights _Net_batch _Net_inputs ReductionParameter HingeLossParameter BlobProto BlobProtoVector NetStateRule LayerParameter PowerParameter FillerParameter ArgMaxParameter V0LayerParameter ClipParameter InnerProductParameter ConvolutionParameter SolverState EltwiseParameter LossParameter SliceParameter BatchNormParameter WindowDataParameter DummyDataParameter HDF5OutputParameter TanHParameter TransformationParameter SoftmaxParameter ConcatParameter DataParameter SPPParameter SwishParameter ParamSpec EmbedParameter SolverParameter InputParameter MVNParameter ContrastiveLossParameter NetState NetParameter RecurrentParameter BiasParameter CropParameter DropoutParameter PoolingParameter Datum ParameterParameter SigmoidParameter BlobShape ExpParameter AccuracyParameter LogParameter ThresholdParameter TileParameter MemoryDataParameter LRNParameter ReLUParameter ImageDataParameter ELUParameter ReshapeParameter InfogainLossParameter ScaleParameter V1LayerParameter HDF5DataParameter PReLUParameter FlattenParameter PythonParameter TestLayerTypeList simple_net_file TestNet lenet TestNetSpec silent_net anon_lenet exception_net_file parameter_net_file SimpleLayer TestPythonLayer ParameterLayer python_net_file ExceptionLayer SimpleParamLayer TestLayerWithParam python_param_net_file TestSolver ParseNolintSuppressions CheckVlogArguments CheckSectionSpacing FindNextMultiLineCommentEnd ReplaceAll CheckForFunctionLengths _SetOutputFormat _IsTestFilename _VerboseLevel CheckBraces RemoveMultiLineComments ResetNolintSuppressions CheckForNonStandardConstructs _SetVerboseLevel PrintUsage _NestingState CheckIncludeLine CheckAccess _CppLintState Search CheckInvalidIncrement RemoveMultiLineCommentsFromRange CleansedLines CheckForBadCharacters UpdateIncludeState FindPreviousMatchingAngleBracket CheckEmptyBlockBody FindNextMultiLineCommentStart Match _NamespaceInfo CheckMakePairUsesDeduction CheckCheck IsBlankLine _SetFilters ProcessLine _FunctionState CheckPosixThreading GetLineWidth GetHeaderGuardCPPVariable IsCppString _IncludeState CheckSpacing _ClassInfo CheckForCopyright IsErrorSuppressedByNolint ProcessFileData CheckForMultilineCommentsAndStrings CloseExpression _PreprocessorInfo _OutputFormat CheckForIncludeWhatYouUse CheckSpacingForFunctionCall FindEndOfExpressionInLine FindNextMatchingAngleBracket _SetCountingStyle ProcessFile _IncludeError CleanseRawStrings CheckAltTokens CheckForNewlineAtEOF ParseArguments CheckForNonConstReference PrintCategories _Filters main FilesBelongToSameModule CheckCStyleCast FileInfo _BlockInfo CheckForHeaderGuard CheckCaffeDataLayerSetUp ReverseCloseExpression CleanseComments _DropCommonSuffixes _ClassifyInclude CheckStyle CheckCaffeAlternatives FindStartOfExpressionInLine _ShouldPrintError CheckComment Error _GetTextInside CheckLanguage CheckCaffeRandom GetPreviousNonBlankLine reporthook parse_readme_frontmatter model_checks_out valid_dirname get_start_time extract_seconds extract_datetime_from_line get_log_created_year print shape upsample_filt model_def endswith ArgumentParser save mean_file channel_swap output_file dirname expanduser parse_args input_file predict Classifier set_mode_cpu load time isdir print add_argument set_mode_gpu pretrained_model gpu len DataFrame Detector format to_hdf detect_selective_search mean set_index to_csv detect_windows read_csv add_argument ArgumentParser read NetParameter output_image_file rankdir Merge draw_net_to_file items DESCRIPTOR batch_size str num_output get_pooling_types_dict add_edge get_edge_label Dot get_layer_label values name choose_color_by_layertype Edge Node bottom append type layer add_node top shape BlobProto extend flat extend BlobProtoVector ParseFromString BlobProtoVector extend tostring shape Datum flat data len astype float32 tile zoom tuple resize fill empty array concatenate shape tile empty array LayerParameter NetParameter _to_proto extend Counter OrderedDict values iteritems isinstance extend add getattr setattr iteritems layers index set outputs _forward len iteritems _backward layers inputs index set len iteritems asarray extend copy next _batch forward len iteritems izip_longest asarray backward extend next _batch zip forward len ascontiguousarray concatenate num zeros next range len GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType GeneratedProtocolMessageType NamedTemporaryFile str close write data Pooling pool1 conv2 pool2 ip1 relu1 SoftmaxWithLoss Convolution NetSpec DummyData ip2 ReLU InnerProduct label conv1 Pooling SoftmaxWithLoss Convolution DummyData ReLU InnerProduct data NetSpec DummyData Silence data2 error search add group clear compile compile compile SetOutputFormat SetCountingStyle SetFilters _Filters startswith IsErrorSuppressedByNolint _ShouldPrintError write IncrementErrorCount replace append Match group find startswith endswith range error FindNextMultiLineCommentEnd RemoveMultiLineCommentsFromRange FindNextMultiLineCommentStart rstrip find xrange len FindEndOfExpressionInLine xrange len FindStartOfExpressionInLine error min search I xrange len FileInfo RepositoryName sep sub ParseNolintSuppressions error startswith split GetHeaderGuardCPPVariable enumerate error enumerate error len error replace count error find error find error find error find error Search error match InnermostClass replace error escape Match Search error group Search Check error lines Count End group Begin xrange NumLines Match raw_lines Search error match group error Match group pop group append Search pop group append Search elided replace CheckSpacingForFunctionCall rfind error len group min CloseExpression NumLines sub xrange find CheckComment Match Search lines_without_raw_strings error group starting_linenum Match range Search error rfind len group ReverseCloseExpression Search Match CloseExpression find error Match CloseExpression find elided error strip group FindEndOfExpressionInLine xrange find Match CloseExpression len error Match finditer normalize isinstance GetLineWidth int InnermostClass CheckCheck error CheckAltTokens CheckBraces CheckSpacing CheckSectionSpacing CheckEmptyBlockBody CheckAccess GetHeaderGuardCPPVariable lines_without_raw_strings _DropCommonSuffixes RepositoryName match split CheckNextIncludeOrder CanonicalizeAlphabeticalOrder FileInfo error search group SetLastHeader match _ClassifyInclude Match pop end search set itervalues append M rstrip replace CheckCStyleCast error _GetTextInside CheckIncludeLine search group lstrip startswith Match ResetSection Search split rfind error group ReverseCloseExpression lstrip xrange findall Match Search ReplaceAll error Match Search endswith replace setdefault group search CleanseComments open FilesBelongToSameModule error search copy sub xrange NumLines FullName keys error search CheckPosixThreading ParseNolintSuppressions CheckVlogArguments CheckMakePairUsesDeduction CheckCaffeDataLayerSetUp CheckLanguage CheckInvalidIncrement CheckCaffeRandom CheckForNonConstReference check_fn Update CheckForNonStandardConstructs CheckStyle raw_lines CheckForMultilineCommentsAndStrings CheckCaffeAlternatives CheckForFunctionLengths CleansedLines _NestingState CheckForBadCharacters CheckForNewlineAtEOF _IncludeState NumLines RemoveMultiLineComments CheckForCopyright ResetNolintSuppressions CheckForHeaderGuard xrange CheckCompletedBlocks CheckForIncludeWhatYouUse ProcessLine _FunctionState Error rstrip endswith len write ProcessFileData _SetVerboseLevel range split write exit join write exit _VerboseLevel int getopt _SetOutputFormat set _SetVerboseLevel PrintCategories _SetFilters _OutputFormat PrintUsage _SetCountingStyle split getreader ParseArguments ResetErrorCounts stderr exit verbose_level PrintErrorCounts StreamReaderWriter ProcessFile getwriter int time write flush load join index int rfind datetime split getctime year strip extract_datetime_from_line get_start_time total_seconds strip write get_log_created_year close extract_datetime_from_line open | ## Holistically-Nested Edge Detection Created by Saining Xie at UC San Diego ### Introduction: <img src="http://pages.ucsd.edu/~ztu/hed.jpg" width="400"> We develop a new edge detection algorithm, holistically-nested edge detection (HED), which performs image-to-image prediction by means of a deep learning model that leverages fully convolutional neural networks and deeply-supervised nets. HED automatically learns rich hierarchical representations (guided by deep supervision on side responses) that are important in order to resolve the challenging ambiguity in edge and object boundary detection. We significantly advance the state-of-the-art on the BSD500 dataset (ODS F-score of .790) and the NYU Depth dataset (ODS F-score of .746), and do so with an improved speed (0.4s per image). Detailed description of the system can be found in our [paper](http://arxiv.org/abs/1504.06375). ### Citations If you are using the code/model/data provided here in a publication, please cite our paper: @InProceedings{xie15hed, author = {"Xie, Saining and Tu, Zhuowen"}, Title = {Holistically-Nested Edge Detection}, | 2,069 |
fandulu/MPLT | ['multi object tracking'] | ['Using Panoramic Videos for Multi-person Localization and Tracking in a 3D Panoramic Coordinate'] | reid/__init__.py reid/backbones/senet.py sort/nn_matching.py sort/kalman_filter.py reid/baseline.py sort/linear_assignment.py sort/track.py reid/backbones/resnet.py sort/iou_matching.py utils.py sort/detection.py sort/preprocessing.py reid/backbones/__init__.py sort/tracker.py get_image plot_pano nms extract_image_patches plot_pano_detection get_world_locations rotate distance plot_pano_tracking get_pano_locations soft_nms weights_init_classifier weights_init_kaiming Baseline build_encoder ResNet conv3x3 BasicBlock Bottleneck SENet SEResNetBottleneck SEBottleneck SEResNeXtBottleneck Bottleneck SEModule Detection iou iou_cost KalmanFilter _cosine_distance NearestNeighborDistanceMetric _nn_cosine_distance _nn_euclidean_distance _pdist non_max_suppression Track TrackState VideoCapture read int COLOR_BGR2RGB CAP_PROP_POS_MSEC set shape resize cvtColor rot90 release show str Circle subplots LINE_AA plot putText axis FONT_HERSHEY_SIMPLEX add_patch add_axes shape imshow scatter annotate tick_params range len axis FONT_HERSHEY_SIMPLEX subplot2grid tick_params show str shape imshow scatter savefig range LINE_AA plot annotate int deepcopy Circle putText subplots_adjust add_patch figure len append shape astype radians list rotate keys stack append array range len show str Circle subplots LINE_AA plot putText axis FONT_HERSHEY_SIMPLEX add_patch add_axes shape imshow scatter annotate tick_params range len cos sin append minimum concatenate astype maximum delete argsort append len minimum ones_like exp concatenate maximum append argmax affine bias kaiming_normal_ weight __name__ constant_ bias normal_ weight __name__ constant_ Baseline prod maximum to_tlwh INFTY_COST asarray arange iou zeros enumerate len T inf dot float clip norm asarray _pdist _cosine_distance minimum concatenate astype maximum delete argsort append float len | # Using panoramic videos for multi-person localization and tracking in a 3D panoramic coordinate <!-- TABLE OF CONTENTS --> ## Table of Contents * [About the Project](#about-the-project) * [Dataset Download Link](#dataset-download-link) * [Getting Started](#getting-started) * [Installation](#installation) * [Run code](#run-code) * [Demos](#demos) * [License](#license) | 2,070 |
fanny-yang/MABFDR | ['unity'] | ['Online control of the false discovery rate with decaying memory'] | exp_new.py plot_results.py parse_mu.py onlineFDR_proc/AlphaInvest.py plot_main.py exp_new_punif.py onlineFDR_proc/wrongFDR.py main_real.py plotting.py rowexp_new.py onlineFDR_proc/Lord.py LUCB.py onlineFDR_proc/GAI_MW.py importme.py generate_mu.py main_onlocal.py main_prep.py confidence_bounds.py onlineFDR_proc/Bonferroni.py parse_new_yorker.py main_onlocal_pll.py onlineFDR_proc/GAIPlus.py UniformSampling.py plot_results_punif.py _binary_search test_cb_validity plot_cb ConfidenceBound run_single saveres run_single saveres saveres create_hyp create_mu generate_mu generate_hyp LUCB main main wrapper generate_ny_mu saveres saveplot plot_errors_mat plotsingle_shaded_mat main plot_results_punif bin_search rowexp Uniform ALPHA_proc BONF_proc GAI_proc GAI_MW_proc LORD_proc wrongFDR_proc _binary_search test_cb_validity plot_cb ConfidenceBound run_single saveres run_single saveres create_hyp create_mu generate_mu generate_hyp LUCB main main wrapper generate_ny_mu saveres saveplot plot_errors_mat plotsingle_shaded_mat main plot_results_punif bin_search rowexp Uniform ALPHA_proc BONF_proc GAI_proc GAI_MW_proc LORD_proc wrongFDR_proc f show plot xlabel title legend ConfidenceBound range randn cumsum print ConfidenceBound float range print join savetxt makedirs GAI_proc saveres next_alpha LORD_proc true_divide BONF_proc max open rowexp multi_ab strftime sum range GAI_MW_proc close ALPHA_proc wrongFDR_proc get_mu enumerate time makedirs write rightarm zeros len rand randn ones sort concatenate rand saveres zeros range int concatenate ones saveres choice true_divide ceil zeros sum range arange generate_mu arange range generate_hyp run_single list time mu_style dist_style ny_data arange product print load_balanced_view map Client ids dict execute range push len run_single load list append saveres min set open zeros float range enumerate len join print close savefig makedirs saveplot errorbar set_xlabel set_xlim add_subplot grid set_ylabel figure legend array range len saveplot plot xlabel min add_subplot ylabel grid tight_layout mean ylim array figure fill_between max range len plot_results plot_results_punif sorted loadtxt min set argsort average sqrt array plot_errors_mat true_divide plotsingle_shaded_mat zeros std range enumerate len f range sum exp ones multiply maximum sqrt true_divide float range log len sum exp ones multiply maximum sqrt true_divide float range log len sum exp ones multiply maximum sqrt true_divide float range log len sum exp ones multiply maximum sqrt true_divide float range log len | fanny-yang/MABFDR | 2,071 |
faresbs/san | ['sign language recognition'] | ['Context Matters: Self-Attention for Sign Language Recognition'] | source/evaluate_slr.py source/bleu.py source/dataloader_slt.py source/beam.py source/rouge.py source/transformer_slr.py source/evaluate_slt.py source/transformer_slt.py source/train_slr.py source/train_slt.py source/viz.py source/preprocessing.py source/dataset_stats.py source/utils.py source/download.py source/beam_search.py source/resize.py source/models/mb2.py source/dataloader_slr.py Beam translate_batch beam_decode _get_ngrams compute_bleu show_batch collate_fn loader PhoenixDataset SubtractMeans get_statistics download decoding lookup resize _len_lcs _get_ngrams rouge rouge_l_summary_level rouge_n _recon_lcs _split_into_words _lcs rouge_l_sentence_level _union_lcs _f_p_r_lcs _get_word_ngrams run_epoch run_epoch src_2Dembeddings FullTransformer HandExtractor make_model clones Output_layer PositionWise LayerNorm Encoder EncoderStack ResidualSkipWithLayerNorm ScaledDotProductAttention MultiHeadedAttention PositionalEncoding src_2Dembeddings FullTransformer HandExtractor make_model clones Output_layer PositionWise Decoder LayerNorm Encoder EncoderStack ResidualSkipWithLayerNorm ScaledDotProductAttention tgt_embeddings DecoderStack MultiHeadedAttention PositionalEncoding path_data Batch get_std_opt NoamOpt subsequent_mask LabelSmoothing greedy_decode relative_window learning_curve_slt learning_curve_slr InvertedResidual _make_divisible ConvBNReLU mobilenet_v2 MobileNetV2 translate_batch src_emb unsqueeze position encode range append collect_hypothesis_and_scores tuple range Counter len _get_ngrams exp Counter zip float sum range len append pad size numpy make_grid len ioff Compose axis show_batch DataLoader imshow savefig PhoenixDataset enumerate len update str view print size ProgressBar start ImageFolder DataLoader enumerate len beam_decode rel_mask Variable print src_mask greedy_decode quit update join len read_csv listdir range split update join ANTIALIAS ProgressBar start save isfile listdir open add set _lcs dict max range _recon tuple _lcs map intersection _get_word_ngrams len _len_lcs _split_into_words len _recon_lcs set _split_into_words union len _split_into_words len mean map zip ctc_beam_search_decoder clip_grad_norm_ zero_grad arch position IntTensor forward hand_extractor rel_mask Batch transpose ctc_loss step output_layer encoder to range update ProgressBar src_mask start eval distributed item train enumerate join time backward print feature_extractor parameters cpu hand_query numpy len max view trg append trg_mask greedy_decode loss_fn Variable clone hybrid dropout size transpose matmul sqrt masked_fill softmax FullTransformer deepcopy src_2Dembeddings load PositionWise Output_layer print Encoder named_parameters EncoderStack xavier_uniform_ c load_state_dict MultiHeadedAttention PositionalEncoding Decoder Sequential tgt_embeddings DecoderStack join astype zeros range hand_emb decode append Variable size src_emb output_layer subsequent_mask unsqueeze position encode to max range cat join plot xlabel title savefig figure legend range len join plot xlabel title savefig figure legend range len int max print load_state_dict_from_url load_state_dict MobileNetV2 | # Sign Attention Network This repository provides a pytorch-based implementation of Context Matters: Self-Attention for Sign Language Recognition. Please Note that in the paper we considered using only the Sign Language Recognition part. ## Getting Started These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. ## Updates * I am well aware of the code errors. Please use this repository as a reference code as there are things that you may need to change to make it work. Thank you for your understanding. * Paper published in ICPR 2020. * Paper Arxiv link: https://arxiv.org/abs/2101.04632 ### Prerequisites | 2,072 |
faresbs/slrt | ['sign language recognition'] | ['Context Matters: Self-Attention for Sign Language Recognition'] | source/evaluate_slr.py source/bleu.py source/dataloader_slt.py source/beam.py source/rouge.py source/transformer_slr.py source/evaluate_slt.py source/transformer_slt.py source/train_slr.py source/train_slt.py source/viz.py source/preprocessing.py source/dataset_stats.py source/utils.py source/download.py source/beam_search.py source/resize.py source/models/mb2.py source/dataloader_slr.py Beam translate_batch beam_decode _get_ngrams compute_bleu show_batch collate_fn loader PhoenixDataset SubtractMeans get_statistics download decoding lookup resize _len_lcs _get_ngrams rouge rouge_l_summary_level rouge_n _recon_lcs _split_into_words _lcs rouge_l_sentence_level _union_lcs _f_p_r_lcs _get_word_ngrams run_epoch run_epoch src_2Dembeddings FullTransformer HandExtractor make_model clones Output_layer PositionWise LayerNorm Encoder EncoderStack ResidualSkipWithLayerNorm ScaledDotProductAttention MultiHeadedAttention PositionalEncoding src_2Dembeddings FullTransformer HandExtractor make_model clones Output_layer PositionWise Decoder LayerNorm Encoder EncoderStack ResidualSkipWithLayerNorm ScaledDotProductAttention tgt_embeddings DecoderStack MultiHeadedAttention PositionalEncoding path_data Batch get_std_opt NoamOpt subsequent_mask LabelSmoothing greedy_decode relative_window learning_curve_slt learning_curve_slr InvertedResidual _make_divisible ConvBNReLU mobilenet_v2 MobileNetV2 translate_batch src_emb unsqueeze position encode range append collect_hypothesis_and_scores tuple range Counter len _get_ngrams exp Counter zip float sum range len append pad size numpy make_grid len ioff Compose axis show_batch DataLoader imshow savefig PhoenixDataset enumerate len update str view print size ProgressBar start ImageFolder DataLoader enumerate len beam_decode rel_mask Variable print src_mask greedy_decode quit update join len read_csv listdir range split update join ANTIALIAS ProgressBar start save isfile listdir open add set _lcs dict max range _recon tuple _lcs map intersection _get_word_ngrams len _len_lcs _split_into_words len _recon_lcs set _split_into_words union len _split_into_words len mean map zip ctc_beam_search_decoder clip_grad_norm_ zero_grad arch position IntTensor forward hand_extractor rel_mask Batch transpose ctc_loss step output_layer encoder to range update ProgressBar src_mask start eval distributed item train enumerate join time backward print feature_extractor parameters cpu hand_query numpy len max view trg append trg_mask greedy_decode loss_fn Variable clone hybrid dropout size transpose matmul sqrt masked_fill softmax FullTransformer deepcopy src_2Dembeddings load PositionWise Output_layer print Encoder named_parameters EncoderStack xavier_uniform_ c load_state_dict MultiHeadedAttention PositionalEncoding Decoder Sequential tgt_embeddings DecoderStack join astype zeros range hand_emb decode append Variable size src_emb output_layer subsequent_mask unsqueeze position encode to max range cat join plot xlabel title savefig figure legend range len join plot xlabel title savefig figure legend range len int max print load_state_dict_from_url load_state_dict MobileNetV2 | # Sign Attention Network This repository provides a pytorch-based implementation of Context Matters: Self-Attention for Sign Language Recognition. Please Note that in the paper we considered using only the Sign Language Recognition part. ## Getting Started These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. ## Updates * I am well aware of the code errors. Please use this repository as a reference code as there are things that you may need to change to make it work. Thank you for your understanding. * Paper published in ICPR 2020. * Paper Arxiv link: https://arxiv.org/abs/2101.04632 ### Prerequisites | 2,073 |
faroit/CountNet | ['speaker diarization'] | ['CountNet: Estimating the Number of Concurrent Speakers Using Supervised Learning Speaker Count Estimation'] | eval.py predict.py predict_audio.py mae_by_count mae count class_mae int mae where append round max range transform T predict norm | # Speaker Count Estimation using Deep Neural Networks <img width="400" align="right" alt="screen shot 2017-11-21 at 12 35 28" src="https://user-images.githubusercontent.com/72940/33071669-be6c35b2-cebc-11e7-8822-9b998ad1ea09.png"> _CountNet_ is a deep learning model to estimate the number of concurrent speakers from single channel mixtures is a very challenging task that is a mandatory first step to address any realistic “cocktail-party” scenario. It has various audio-based applications such as blind source separation, speaker diarisation, and audio surveillance. This repo provides pre-trained models. ## Publications ### 2019: IEEE/ACM Transactions on Audio, Speech, and Language Processing * __Title__: CountNet: Estimating the Number of Concurrent Speakers Using Supervised Learning Speaker Count Estimation * __Authors__: [Fabian-Robert Stöter](https://faroit.com), Soumitro Chakrabarty, Bernd Edler, Emanuël A. P. Habets | 2,074 |
fastread/src | ['active learning'] | ['Finding Better Active Learners for Faster Literature Reviews'] | src/simulate.py src/util/mar.py src/index.py load hello restart plot est train search susp labeling export load_old auto MAR create get_numbers create_old join remove listdir int get_numbers save code code get_numbers int values remove create name filename get_numbers format enable_est est_num random save get_numbers latest_labeled format BM25_get format BM25 split | What is FASTREAD (or FAST2)?
-----
FASTREAD (FAST2) is a tool to support primary study selection in systematic literature review.
Latest Versions:
- On Github repo: [https://github.com/fastread/src](https://github.com/fastread/src).
- In the Seacraft repository: [](https://doi.org/10.5281/zenodo.852663)
Cite as:
| 2,075 |
faustomilletari/CFCM-2D | ['semantic segmentation'] | ['CFCM: Segmentation via Coarse to Fine Context Memory'] | cfcm/layers/layers.py cfcm/utilities/timing.py cfcm/utilities/tensorflow/summary.py cfcm/zoo/fitrunners.py cfcm/zoo/tf_resnet.py cfcm/zoo/models.py cfcm/utilities/tensorflow/saver.py cfcm/cli.py setup.py cfcm/utilities/fitting/fitrunners.py cfcm/data/transforms.py cfcm/data/loading.py cfcm/layers/losses.py cfcm/utilities/iterators/batchiterators.py cfcm/data/managers.py cfcm/utilities/visualization.py train cli load_EndoVis_transform load_ISBI_tif_transform readGTFrames readFrames load_montgomery_xray DataManager resize_img specularity normalize_between_zero_and_one convert_to_float make_onehot flip_up_down shuffle_data crop_actual_image threshold_labels_transform add_trailing_singleton_dimension_transform flip_left_right BasicConvLSTMCell ConvRNNCell _conv_linear dice binary_cross_entropy_2D dice_loss timeit PrecisionRecallCurve surfd ConfusionMatrix RoCCurve FitRunner ParallelBatchIterator BatchIterator Saver SummaryManager ResNetEndovis SegBaseclass MontgomeryXRay ENDOVIS2015 ResNet LSTMResNetEndovis LSTMResNetMontgomeryXray LSTMResNet ResNetMontgomeryXray upsample_tf residual_block resnet lstm_resnet conv2d_transpose seg_resnet_lstm_v2_generator block_layer_expanding building_block seg_resnet_v2_generator block_layer_compressing conv2d_fixed_padding seg_resnet_lstm_v2 batch_norm_relu seg_resnet_v2 bottleneck_block load ResNetEndovis format print LSTMResNetEndovis LSTMResNetMontgomeryXray run_epoch_train open run_epoch_valid range ResNetMontgomeryXray list array get_reader join list endswith array get_reader split function as_list weighted_cross_entropy_with_logits softmax_cross_entropy_with_logits reshape reduce_sum reduce_mean concatenate distance_transform_edt astype ndim atleast_1d generate_binary_structure bool binary_erosion conv2d batch_norm as_list int round resize_images seg_resnet_lstm_v2 model seg_resnet_v2 model batch_norm relu conv2d conv2d_fixed_padding batch_norm_relu projection_shortcut conv2d_fixed_padding batch_norm_relu projection_shortcut append block_fn range relu block_fn relu concat append range | # CFCM This repository contains a publicly available version of CFCM: segmentation via Coarse to Fine Context Memory, a paper accepted at MICCAI 2018 for presentation. An arXiv version of the paper is available [here](https://arxiv.org/abs/1806.01413), and the final version is available [through Springer](https://link.springer.com/chapter/10.1007%2F978-3-030-00937-3_76). In this repository we provide only training/validation code. In the future it will be possibile to use CFCM to obtain predictions on the Montgomery XRay lung dataset over the cloud using [TOMAAT](https://tomaatcloud.github.io) and a compatible client (for example through 3D Slicer and its TOMAAT extension). The code is provided without guarantees of correctness and functionality. This implementation is derived directly from the original implementation used for the CFCM paper, | 2,076 |
fcUalberta/tless_edge_based | ['boundary detection', 'edge detection'] | ['Holistically-Nested Edge Detection'] | feature_enhancement/hed/hed_features.py train_test_main.py utils/utils.py feature_enhancement_main.py feature_enhancement/hed/hed_overlay.py feature_enhancement/hed/hed.py feature_enhancement/feature_enhancement_main.py preprocessing/manual_augmentation.py preprocessing/preprocessing.py initial_augmentation_main.py preprocessing/automatic_augmentation.py train_batches create_feature_matrix1 create_labels test holistically_nested load_hed_model CropLayer holistically_nested load_hed_model CropLayer holistically_nested load_hed_model CropLayer create_annotations automatic_augmentor image_processing manual_augmentor bounding_box crop_at_bb copy_separate_files copy_files data_balancing data_distribution join listdir print open DataFrame append split join print flatten resize append imread listdir len compute dump asarray Incremental print score classification_report predict shuffle partial_fit open transform train_test_split StandardScaler range append len imwrite canny DataFrame open len append imread range asarray COLOR_BGR2GRAY insert GaussianBlur listdir load join create_feature_matrix1 print to_csv transform read_csv makedirs print readNetFromCaffe dnn_registerLayer astype setInput resize forward blobFromImage fromarray convert blend skew rotate90 rotate270 flip_top_bottom mkdir random_erasing sample flip_left_right Pipeline crop_random join int str print write close shape mkdir imread listdir open adjust_gamma join show str imwrite normal convertScaleAbs copyfile choice shape GaussianBlur int join image_processing mkdir imread listdir range show subplots threshold_otsu rgb2gray closing regionprops square add_patch tight_layout imshow clear_border Rectangle label bbox subplots imwrite round open show str shape imshow imread bounding_box close tight_layout listdir join int print write add_patch Rectangle makedirs str join print append listdir range int len append round range data_distribution join listdir mkdir copy join listdir mkdir copy | # Textureless Object Recognition - An Edge-based approach This project is a part of Ualberta Masters program Course project: Textureless Object Recognition using an edge-based Approach ## About Textureless object recognition has become a significant task in Computer Vision with the advent of Robotics and its applications in manufacturing sector. It has been challenging to obtain good accuracy in real time because of its lack of discriminative features and reflectance properties which makes the techniques for textured object recognition insufficient for textureless objects. In this project, by applying image processing techniques we created a robust augmented dataset from initial imbalanced smaller dataset. We extracted edge features, feature combinations and all its combination on the RGB dataset to create 15 datasets, each with a size of ~340,000. We then trained four classifiers on these 15 datasets to arrive at a conclusion as to which dataset performs the best overall and whether edge features are important for textureless objects. Based on our experiments and analysis, RGB images enhanced with combination of 3 edge features performed the best compared to all other. Model performance on dataset with HED edges performed comparatively better than other edge detectors like Canny or Prewitt. [Link to full Research Report](https://github.com/fcUalberta/tless_edge_based/blob/master/Industrial%20Textureless%20Object%20Recognition.pdf) ## Implementation structure Part 1: Creation of Initial Augmented Data - Image Data Acquisition - Obtaining ground truth data | 2,077 |
fderoos/probabilistic_hessian | ['stochastic optimization'] | ['Active Probabilistic Inference on Matrices for Pre-Conditioning in Stochastic Optimization'] | setup.py probabilistic_hessian/probabilistic_hessian.py probabilistic_hessian/__init__.py Valid_Cov Estimate_Hessian MemoizeJacHessp Estimate_Hessian_Scalar Estimate_Scalar_Parameters norm arange zeros_like print hessp flatten jac abs max arange eigh hessian_product abs derivative solve fhess_p MemoizeJacHessp copy LinearOperator callable float Estimate_Scalar_Parameters T Valid_Cov dot eye zeros bool diag svds arange eigh hessian_product derivative solve fhess_p MemoizeJacHessp sqrt LinearOperator callable float Estimate_Scalar_Parameters T dot eye zeros bool diag svds | # Probabilistic Hessian Estimation This is a scipy-inspired implementation of the Hessian-inference algorithm presented in the paper [Active Probabilistic Inference on Matrices for Pre-Conditioning in Stochastic Optimization][article] ## Installation Install via pip install git+https://github.com/fderoos/probabilistic_hessian.git ## Usage The ``Probabilistic_Hessian`` module contains the function ``Estimate_Hessian`` which requires handles to the gradient and a Hessian-vector product similar to `scipy.optimize.minimize`. See below for two examples. ```python # (1) With handles import numpy as np | 2,078 |
fdtomasi/regain | ['time series'] | ['Latent Variable Time-varying Network Inference'] | regain/clustering.py regain/utils.py regain/bayesian/wishart_distribution_.py regain/covariance/kernel_latent_time_graphical_lasso_.py regain/plotting/graph.py regain/linear_model/lasso_.py regain/generalized_linear_model/glm_time_poisson.py regain/tests/test_update_rules.py regain/tests/test_graphical_lasso.py regain/covariance/latent_time_graphical_lasso_.py regain/norm.py regain/data/__init__.py notebooks/performance_utils.py regain/datasets/base.py regain/bayesian/__init__.py regain/forward_backward/time_graphical_lasso_.py regain/covariance/missing_graphical_lasso_.py regain/generalized_linear_model/glm_ising.py regain/datasets/ising.py regain/linear_model/group_lasso_.py regain/scores.py regain/covariance/kernel_time_graphical_lasso_.py regain/model_selection/stability_optimization.py regain/prox.py regain/tests/test_prox.py regain/tests/test_time_graphical_lasso.py regain/wrapper/paspal/glopridu.py regain/plotting/plotly.py regain/discriminant_analysis.py regain/generalized_linear_model/glm_time_ising.py regain/bayesian/sampling.py regain/bayesian/wishart_process_.py regain/covariance/graphical_lasso_.py regain/covariance/__init__.py regain/bayesian/gwishart_inference.py regain/plotting/results.py travis_tests.py regain/data/base.py regain/tests/test_datasets.py regain/tests/__init__.py regain/tests/test_utils.py regain/linear_model/__init__.py regain/generalized_linear_model/glm_poisson.py regain/update_rules.py regain/datasets/kernels.py regain/tests/test_latent_graphical_lasso.py regain/covariance/latent_graphical_lasso_.py regain/forward_backward/__init__.py notebooks/_graph-clustering/local_utils.py regain/model_selection/__init__.py regain/__init__.py regain/datasets/multi_class.py regain/forward_backward/forward_backward.py regain/linear_model/_group_lasso_overlap_old_.py regain/linear_model/group_lasso_overlap_.py regain/validation.py regain/covariance/infimal_convolution_.py regain/bayesian/gaussian_process_.py regain/covariance/latent_time_matrix_decomposition.py regain/datasets/__init__.py regain/bayesian/_laplace_approximation.py regain/datasets/gaussian.py regain/bayesian/stats.py notebooks/_kernel/kltgl_performance.py regain/tests/test_latent_time_graphical_lasso.py regain/wrapper/__init__.py regain/covariance/time_graphical_lasso_.py setup.py regain/plotting/covariance.py regain/model_selection/_bayesian_optimization.py regain/plotting/__init__.py regain/generalized_linear_model/glm_gaussian.py regain/datasets/poisson.py regain/generalized_linear_model/base.py tgl_results gl_results likelihood_score glasso_results ltgl_results chandresekeran_results hallac_results friedman_results lgl_results convert_dict_to_df highlight_max set_results ensure_ticc_valid highlight_max_std tgl_results ktgl_results kltgl_results base_results wp_results run_results ltgl_results use_bscv get_representative graph_k_means compute_distances DiscriminantAnalysis PrecomputedDiscriminantAnalysis vector_p_norm l1_od_norm node_penalty l1_norm blockwise_soft_thresholding_symmetric prox_FL prox_node_penalty prox_linf_1d blockwise_soft_thresholding soft_thresholding soft_thresholding_vector prox_laplacian prox_logdet_ala_ma prox_logdet prox_trace_indicator _soft_thresholding_od_2d _blockwise_soft_thresholding_2d prox_linf soft_thresholding_od EBIC_m_t log_likelihood_t BIC BIC_t EBIC_t EBIC_m EBIC log_likelihood update_rho update_gamma load_pickle threshold top_n_indexes flatten logentropy_normalize display_topics compose structure_error save_pickle alpha_heuristic namedtuple_with_defaults positive_definite ensure_posdef mean_structure_error is_pos_def write_network init_logger error_norm_time is_pos_semidef upper_to_full retain_top_n error_rank suppress_stdout error_norm normalize_matrix read_network convert_data_to_2d _ensure_filename_ending check_norm_prox check_array_dimensions check_input check_input_3d sample markov_blankets precision_selection BayesianGraphicalLasso GWishartScore bayesian_graphical_lasso _get_graphs GWishartFit score_blankets covsel compute_score mk_all_ugs sample_ell elliptical_slice GWP_construct _sample_ell_comp sample_hyper_kernel log_lik_frob lognormal_pdf lognormal_logpdf t_mvn_logpdf lognstat WishartDistribution main InverseWishartDistribution NormalInverseWishartDistribution _periodic_kernel _rbf_kernel WishartProcess predict fit first_derivative_h second_derivative_h_version2 h first_derivative_h_version2 GraphicalLasso graphical_lasso logl objective init_precision infimal_convolution objective SimilarityLatentTimeGraphicalLasso kernel_latent_time_graphical_lasso objective KernelLatentTimeGraphicalLasso kernel_time_graphical_lasso objective_kernel objective_similarity SimilarityTimeGraphicalLasso KernelTimeGraphicalLasso precision_similarity objective latent_graphical_lasso objective LatentGraphicalLasso LatentTimeGraphicalLasso latent_time_graphical_lasso objective LatentTimeMatrixDecomposition objective latent_time_matrix_decomposition missing_graphical_lasso compute_mean compute_empirical_covariance compute_cs MissingGraphicalLasso objective TimeGraphicalLasso time_graphical_lasso loss load_finance load_food_search_trends make_dataset _ising_case _poisson_case _gaussian_case _update_theta_l1 make_fede make_sin_cos make_fixed_sparsity make_covariance _update_theta_l2 make_ma_xue_zou_rand_k _permute_locations make_sin data_Meinshausen_Yuan_sparse_latent _theta_my make_starting _compute_probabilities make_ell _update_ell_l2 make_ma_xue_zou data_Meinshausen_Yuan make_sparse_low_rank hamiltonian ising_sampler p_plus update_ising_l1 ising_metropolis_hastings ising_exact p_plus_minmax direct_sampling ising_theta_generator genRandInv make_cluster_representative make_exp_sine_squared _block_matrix make_RBF genInvCov make_ticc_dataset make_ticc_dataset_v3 make_ticc_dataset_new make_ticc sample generate_multiple_class_dataset make_multiclass_dataset poisson_sampler _adjacency_to_A poisson_theta_generator update_l1 fista_step upper_diag_3d _scalar_product choose_gamma choose_lamda grad_loss_laplacian loss_laplacian grad_loss penalty penalty_laplacian objective_laplacian _J tgl_forward_backward TimeGraphicalLassoForwardBackward objective loss GLM_GM build_adjacency_matrix Gaussian_GLM_GM fit_each_variable objective IsingGraphicalModel loss _fit _gradient_ising objective loss_single_variable fit_each_variable PoissonGraphicalModel objective loss objective_single_variable objective_kernel _fit_time_ising_model loss_ising SimilarityTemporalIsingModel precision_similarity objective TemporalIsingModel SimilarityTemporalPoissonModel _fit_time_poisson_model loss_poisson objective_kernel TemporalPoissonModel objective group_lasso objective P_star_x_bar_function _remove_unused_features GroupLassoOverlap GroupLassoOverlapClassifier group_lasso_overlap D_function _overlapping_group_lasso objective lasso lu_factor objective group_lasso_feat_split x_update GraphicalModelStabilitySelection upper_bound graphlet_instability global_instability _check_param_order _BayesianOptimization _fit_score plot_cov_2d plot_cov_ellipse plot_graph_with_latent_variables plot_meinshausen_yuan plot_precision_recall_comparison plot_curve plot_roc_curves plot_roc_comparison test_make_dataset_gaussian_sklearn test_make_dataset_gaussian_fede test_make_dataset_gaussian_sin test_quickstart test_make_dataset_ising test_make_dataset_gaussian test_make_dataset_gaussian_fixed_sparsity test_make_dataset_poisson test_make_dataset_gaussian_sincos test_make_dataset_gaussian_ma test_make_dataset_gaussian_gp test_gl test_lgl_zero test_ltgl_zero test_ltgl_prox_l1 test_soft_thresholding_vector test_soft_thresholding_od test_soft_thresholding test_ltgl_zero test_update_gamma test_update_rho test_error_norm test_error_norm_time test_structure_error test_flatten test_ensure_filename_ending test_upper_to_full test_error_rank test_suppress_stdout glopridu_algorithm glo_prox reshape sum array GraphicalLasso time n_iter_ precision_ transpose dict error_norm array append max fit LatentGraphicalLasso time n_iter_ precision_ transpose error_rank dict error_norm array latent_ append max fit TimeGraphicalLasso time n_iter_ precision_ dict error_norm fit time n_iter_ precision_ error_rank dict error_norm LatentTimeGraphicalLasso latent_ structure_error fit time GLsk n_iter_ precision_ transpose dict error_norm array append structure_error max fit empirical_covariance int time max transpose dict error_norm append array glasso time print transpose dict TVGL error_norm vstack append structure_error array T lvglasso Bunch transpose error_rank sqrt dict error_norm structure_error update items list reset_index upper rename DataFrame items iterrows alltrue iter range append v_measure_score structure_error error_norm_time apply applymap astype where apply DataFrame BayesSearchCV fit set_params time hasattr precision_ dict error_rank error_norm latent_ structure_error use_bscv fit astype vstack KernelTimeGraphicalLasso KernelLatentTimeGraphicalLasso vstack astype time precision_ astype dict error_norm vstack WishartProcess structure_error fit data tgl_results IndexSlice print transpose thetas ktgl_results kltgl_results thetas_observed ltgl_results tau ells alpha enumerate append astype zeros_like zeros size astype enumerate T arange all compute_distances argmin shuffle copy dot range len T fun fill_diagonal diag soft_thresholding _soft_thresholding_od_2d repeat range empty_like empty_like repeat _blockwise_soft_thresholding_2d ravel range norm ones empty_like dot repeat zip ravel diag enumerate T range zeros_like eigh sqrt square eigh sqrt square eigh maximum zeros_like transpose update_rho shape range concatenate hstack copy empty_like sqrt zip enumerate blockwise_soft_thresholding_symmetric squared_norm inv dot convergence eye full tvgen fill_diagonal zeros_like soft_thresholding diag transpose triu_indices_from empty_like vstack func zip power array range sum log log_likelihood sum log log_likelihood zip array log_likelihood_t sum log log_likelihood_t sum log append join print enumerate sum log1p abs top_n_indexes zeros_like T isinstance namedtuple tuple _fields Mapping len string_types isinstance setFormatter list basicConfig getLogger handlers addHandler close StreamHandler removeHandler Formatter removeFilter setLevel filters shutdown _ensure_filename_ending flush _ensure_filename_ending to_csv int sorted fillna fill_diagonal itertuples endswith warn unique read_csv values int triu_indices sqrt zeros vstack astype array reshape sqrt T svdvals copy dot sqrt triu sum amax count_nonzero matthews_corrcoef size astype copy average_precision_score flatten dict prod array squareform items list str dict mean append structure_error std eigvalsh eigvalsh cholesky _ensure_posdef_2d zeros array shape T pdf dot sqrt diagonal max norm partial isinstance warn check_array_dimensions array isinstance len issparse int comb astype argsort append zeros sum range array unique enumerate exp repeat append zeros sum array enumerate transpose array zeros sum pinvh partial minimize eye zeros x empirical_covariance precision_selection S0 inf zeros_like shape d0 Bunch lognormconst pi multi_dot log log_likelihood svd shape triu sum range nonzero empty S0 argsort dot cholesky zeros diag d0 Bunch lognormconst GWishartFit shape cov compute_score zeros diag GraphicalLasso markov_blankets _get_graphs shape score_blankets logspace eye T transpose matmul normal reshape cos L pi likelihood shape xx dot uniform real sin GWP_construct V range log log_likelihood logp_post min lognormal logpdf lognstat ones size sqrt zeros _sample_ell_comp range logp_post normal min copy logpdf sqrt log squared_norm size log pi sum make_sparse_spd_matrix norm multivariate_normal print NormalInverseWishartDistribution marginal_mean marginal_mode eye posterior zeros kern Bunch randn sample_gp set_description vstack sample_hyper_kernel log_likelihood GWP_construct real append format xx elliptical_slice trange tril sample_ell likelihood zeros dot T kern pinvh rbf_kernel zeros_like empty_like dot trace range sum range argsort nonzero triu empty pinvh ndarray zeros_like isinstance empty_like copy zip pinvh enumerate norm zeros_like print update_rho copy prox_logdet convergence append init_precision range soft_thresholding_od squared_norm norm zeros_like soft_thresholding print pinvh prox_laplacian update_rho copy prox_trace_indicator convergence append range obj_ktgl ndarray range isinstance zeros_like prox_psi soft_thresholding ones update_rho shape append check_norm_prox sum range concatenate copy sqrt print squared_norm convergence eye array prox_phi init_precision loss zeros_like prox_psi soft_thresholding ones update_rho shape check_norm_prox append sum range concatenate copy sqrt print squared_norm convergence eye array init_precision check_norm_prox range triu_indices range eye list fill_diagonal map zeros range obj_gl norm zeros_like soft_thresholding print pinvh update_rho copy prox_logdet prox_trace_indicator convergence append range init_precision obj_tgl prox_psi append zeros_like soft_thresholding ones concatenate squared_norm print update_rho copy sqrt convergence array check_norm_prox full range prox_phi init_precision prox_psi append zeros_like soft_thresholding concatenate print squared_norm update_rho copy sqrt convergence array check_norm_prox full range prox_phi copy nan_to_num meshgrid zeros sum range T zeros_like dot meshgrid range nan_to_num copy graphical_lasso print copy dict nan_to_num compute_mean eye append zeros compute_empirical_covariance range compute_cs logl ones prox_psi zeros_like soft_thresholding ones concatenate squared_norm print update_rho copy sqrt convergence array check_norm_prox full range append init_precision join list items dirname zeros DataFrame read_csv join dirname read_csv update list partial transpose inv astype map dict vstack func array transpose astype vstack array ising_theta_generator transpose astype vstack poisson_theta_generator array int norm pdf sqrt uniform zeros array range len append uniform tolist fill_diagonal shuffle append zeros range check_random_state multivariate_normal inv _permute_locations _theta_my dot _compute_probabilities uniform append zeros range fill_diagonal check_random_state multivariate_normal inv shuffle _permute_locations _theta_my dot _compute_probabilities append zeros range int rand randint dot zeros range count_nonzero list set choice make_ell eye zeros sum range fill_diagonal zeros_like randn randint range len randint copy any zip zeros rand copy _update_theta_l1 reshape _update_theta_l2 copy warn dot _update_ell_l2 make_starting append randint ravel range get sum T copy range nonzero make_starting append abs diag triu_indices list ensure_posdef rand square pi maximum map linspace tile sin array len get triu_indices int permutation clip randn pi make_ell linspace sin zeros ravel array range len rand randint maximum dot eigh append zeros multi_dot range append make_sparse_spd_matrix make_low_rank_matrix range int T multi_dot randn print reshape diag min ravel dot eigh eye zeros round max clip int triu_indices T print size rand sign dot vstack eye multi_dot randint copy any zip zeros todense fill_diagonal barabasi_albert_graph warn choice update_ising_l1 array append fast_gnp_random_graph watts_strogatz_graph range range exp fill_diagonal arange choice array len ones zeros exp range ones uniform any append p_plus_minmax range range uniform range p_plus fill_diagonal ones vstack ising_metropolis_hastings ising_exact zeros direct_sampling range get T pi make_ell append array squareform get T dot squareform append array enumerate int randint gnm_random_graph edges zeros zeros range randint seed genRandInv eig min genInvCov zeros abs range T list asarray arange multivariate_normal Bunch make_ticc _block_matrix dot flatten repeat unique append zeros empty pinvh enumerate len arange Bunch flatten pinvh seed list append asarray block _block_matrix eig unique empty enumerate multivariate_normal min dot repeat zeros len seed arange multivariate_normal Bunch block _block_matrix eig min len dot flatten repeat unique append zeros empty pinvh enumerate arange Bunch rand where vstack list len random_regular_graph append range fill_diagonal size astype shuffle copy fit_predict int repeat squareform generate_multiple_class_dataset int T fill_diagonal arange barabasi_albert_graph rand copy where choice sign zip erdos_renyi_graph range append list keys randint copy choice any zip zeros todense fill_diagonal barabasi_albert_graph gnm_random_graph warn update_l1 append fast_gnp_random_graph watts_strogatz_graph range zeros hstack range eye T exp reshape _adjacency_to_A dot zeros array range poisson sqrt prox_FL _scalar_product function_f range soft_thresholding_od _scalar_product objective_f sqrt gradient_f function_f range function_g loss_tgl array sum ndarray map isinstance loss grad_loss range empty_like sum ndarray map isinstance loss_laplacian prox_FL prox_FL warn choose_gamma max shape append range choose_lamda grad_loss_laplacian partial grad_loss copy gradient_f soft_thresholding_od norm print min convergence array init_precision T len astype eye enumerate dot soft_thresholding print gradient shape convergence append zeros range dot shape exp range shape shape range gradient zeros_like T norm print check_array copy dot shape trace _gradient_ising append zeros range loss soft_thresholding_od dot exp range dot exp range loss_single_variable norm dot loss_single_variable ones loss_ising prox_psi zeros_like concatenate print update_rho shape sqrt convergence eye check_norm_prox zeros sum array range append print squareform ones loss_poisson max prox_psi build_adjacency_matrix zeros_like concatenate print update_rho shape sqrt convergence fit_each_variable eye check_norm_prox zeros sum array range append soft_thresholding dot shape append zeros lu_factor range sum sorted list set flatten dict zip range len zeros range enumerate zeros range enumerate P_star_x_bar_function soft_thresholding print inv dot shape D_function append zeros range diag enumerate P_star_x_bar_function inv dot shape eye zeros range enumerate prox len soft_thresholding dot shape append zeros lu_factor range T dot shape eye cholesky objective T norm ones eig dot shape sqrt append zeros ravel max range x_update dot shape range zeros range array triu_indices_from list nodes from_numpy_array edges append range diag graphlet_degree_vectors len sum range array triu_indices_from append items list hasattr atleast_2d plot pi dot eigh sqrt linspace array arctan2 Ellipse degrees eigsorted add_artist sqrt ylim gca xlim circular_layout draw_networkx_nodes subplots set_title set_axis_off draw draw_networkx_labels from_numpy_matrix show plot savefig nonzero figure zip subplots linspace show set_xlabel legend append update plot set_xlim mean interp enumerate minimum auc roc_curve maximum set_ylabel fill_between ravel std set_ylim subplots grid curve_func linspace show set_xlabel savefig legend append update plot set_xlim mean interp auc minimum int items enumerate maximum dict set_ylabel fill_between ravel std set_ylim seed make_dataset make_dataset make_dataset make_dataset make_dataset make_dataset make_dataset make_dataset make_dataset make_dataset make_dataset seed multivariate_normal precision_ eye assert_array_almost_equal zeros precision_ latent_ get_precision assert_array_equal zeros get_observed_precision precision_ latent_ assert_array_equal zeros get_observed_precision precision_ latent_ assert_array_equal zeros arange soft_thresholding reshape assert_array_equal array arange reshape assert_array_equal array soft_thresholding_od blockwise_soft_thresholding_symmetric T arange blockwise_soft_thresholding soft_thresholding_vector reshape assert_array_almost_equal assert_array_equal array update_rho update_gamma _ensure_filename_ending assert_equal assert_array_equal flatten arange assert_array_equal reshape upper_to_full error_rank reshape assert_equal reshape error_norm assert_equal reshape error_norm_time assert_equal structure_error eye assert_equal norm print size min maximum copy dot zeros sum array range len int norm T ones print copy glo_prox dot shape sqrt zeros array range len | [](https://travis-ci.org/fdtomasi/regain) [](http://codecov.io/github/fdtomasi/regain?branch=master) [](http://opensource.org/licenses/BSD-3-Clause) [](https://pypi.python.org/pypi/regain) [](https://anaconda.org/fdtomasi/regain) [](https://requires.io/github/fdtomasi/regain/requirements/?branch=master) # regain Regularised graph inference across multiple time stamps, considering the influence of latent variables. It inherits functionalities from the [scikit-learn](https://github.com/scikit-learn/scikit-learn) package. ## Getting started ### Dependencies `REGAIN` requires: - Python (>= 3.6) - NumPy (>= 1.8.2) - scikit-learn (>= 0.17) | 2,079 |
feiaxyt/Winner_ECCV20_TAO | ['multi object tracking'] | ['1st Place Solution to ECCV-TAO-2020: Detect and Represent Any Object for Tracking'] | tao_detection_release/configs/bags/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis.py tao_detection_release/tools/train.py tao_detection_release/mmdet/core/bbox/geometry.py tao_detection_release/mmdet/core/evaluation/eval_hooks.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110657.py tao_detection_release/mmdet/ops/dcn/deform_pool.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812113844.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0_trainall.py tao_detection_release/configs/bags/gs_htc_x101_64x4d_fpn_20e_16gpu_lvis.py tao_detection_release/mmdet/models/losses/focal_loss.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812113900.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812114458.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804114151.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_smalllr.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804112435.py tao_detection_release/mmdet/core/evaluation/mean_ap.py tao_detection_release/tools/lvis_tao_analyse.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804111844.py tao_tracking_release/application_util/visualization.py tao_detection_release/configs/transferred/faster_rcnn_x101_64x4d_fpn_1x_lvis_is.py tao_detection_release/mmdet/ops/sigmoid_focal_loss/__init__.py tao_detection_release/mmdet/core/anchor/__init__.py tao_detection_release/.history/tools/multi_process_inference/inference_20200804113705.py tao_detection_release/mmdet/datasets/pipelines/loading.py tao_detection_release/mmdet/datasets/tao.py tao_tracking_release/gen_detections.py tao_detection_release/mmdet/datasets/pipelines/transforms.py tao_detection_release/mmdet/ops/roi_pool/roi_pool.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0.py tao_detection_release/mmdet/core/bbox/samplers/sampling_result.py tao_detection_release/mmdet/models/bbox_heads/convfc_bbox_head.py tao_detection_release/mmdet/models/losses/ghm_loss.py tao_detection_release/mmdet/core/fp16/hooks.py tao_detection_release/.history/tools/multi_process_inference/inference_20200802135027.py tao_detection_release/mmdet/models/detectors/cascade_rcnn.py tao_detection_release/mmdet/core/mask/utils.py tao_detection_release/tools/lvis_analyse.py tao_detection_release/mmdet/datasets/wider_face.py tao_detection_release/.history/tools/multi_process_inference/inference_20200802135021.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_dcm.py tao_detection_release/lvis-api/setup.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110807.py tao_tracking_release/application_util/image_viewer.py tao_detection_release/mmdet/models/detectors/double_head_rcnn.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812114343.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110714.py tao_detection_release/tools/test_lvis_tnorm.py tao_detection_release/mmdet/models/detectors/fcos.py tao_detection_release/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats.py tao_detection_release/mmdet/ops/dcn/deform_conv.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812113906.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0_20e.py tao_detection_release/configs/baselines/htc_dconv_c3-c5_mstrain_400_1400_x101_32x4d_rfp.py tao_detection_release/.dev_scripts/gather_models.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804111841.py tao_detection_release/mmdet/core/fp16/__init__.py tao_tracking_release/deep_sort/__init__.py tao_detection_release/mmdet/datasets/__init__.py tao_detection_release/mmdet/core/evaluation/__init__.py tao_detection_release/mmdet/models/mask_heads/fused_semantic_head.py tao_detection_release/mmdet/models/anchor_heads/anchor_head.py tao_detection_release/configs/bags/gs_faster_rcnn_x101_64x4d_fpn_1x_lvis.py tao_detection_release/mmdet/models/necks/bfp.py tao_detection_release/mmdet/models/detectors/rpn.py tao_detection_release/mmdet/core/bbox/samplers/__init__.py tao_detection_release/.history/tools/multi_process_inference/inference_20200802135024.py tao_detection_release/tools/inference.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110808.py tao_detection_release/mmdet/core/anchor/guided_anchor_target.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110703.py tao_detection_release/mmdet/core/bbox/samplers/ohem_sampler.py tao_detection_release/mmdet/models/losses/__init__.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804110914.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_focalloss.py tao_detection_release/mmdet/ops/roi_pool/gradcheck.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_reweightall.py tao_detection_release/mmdet/models/detectors/grid_rcnn.py tao_detection_release/mmdet/models/utils/__init__.py tao_detection_release/mmdet/models/anchor_heads/fovea_head.py tao_detection_release/mmdet/core/utils/__init__.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804114153.py tao_detection_release/mmdet/core/fp16/utils.py tao_detection_release/mmdet/models/detectors/reppoints_detector.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_reweighthead_bf.py tao_detection_release/mmdet/models/backbones/resnext_ori.py tao_detection_release/mmdet/core/bbox/samplers/instance_balanced_pos_sampler.py tao_detection_release/mmdet/models/mask_heads/htc_mask_head.py tao_detection_release/mmdet/core/bbox/samplers/combined_sampler.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812113821.py tao_detection_release/tools/multi_process_inference/inference.py tao_detection_release/configs/tao/htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis.py tao_detection_release/mmdet/datasets/file_dataset.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812114457.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0_bgall.py tao_detection_release/mmdet/ops/masked_conv/__init__.py tao_detection_release/mmdet/models/anchor_heads/__init__.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804111845.py tao_detection_release/mmdet/models/anchor_heads/retina_head.py tao_detection_release/mmdet/models/shared_heads/res_layer.py tao_detection_release/tools/test_lvis.py tao_detection_release/tools/test_tao.py tao_detection_release/mmdet/core/evaluation/lvis_utils.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110804.py tao_detection_release/mmdet/core/utils/misc.py tao_detection_release/mmdet/core/mask/mask_target.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804111846.py tao_detection_release/mmdet/datasets/loader/sampler.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812113937.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110649.py tao_detection_release/mmdet/ops/roi_pool/__init__.py tao_detection_release/mmdet/models/anchor_heads/ga_rpn_head.py tao_detection_release/configs/bags/gs_mask_rcnn_r50_fpn_1x_lvis.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110710.py tao_detection_release/mmdet/core/utils/dist_utils.py tao_detection_release/mmdet/core/bbox/assign_sampling.py tao_detection_release/.history/tools/multi_process_inference/inference_20200804113701.py tao_detection_release/configs/baselines/htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_rfp.py tao_detection_release/setup.py tao_detection_release/mmdet/core/mask/__init__.py tao_detection_release/mmdet/datasets/dataset_wrappers.py tao_detection_release/mmdet/models/anchor_heads/fcos_head.py tao_detection_release/mmdet/models/plugins/non_local.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804110853.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812113859.py tao_detection_release/mmdet/models/detectors/base.py tao_detection_release/.history/tools/multi_process_inference/inference_20200804113709.py tao_detection_release/configs/ablations/faster_rcnn_r50_fpn_1x_lvis_is_with0-bg3.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0_20e_nosample.py tao_detection_release/mmdet/models/shared_heads/__init__.py tao_detection_release/mmdet/core/bbox/bbox_target.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812113846.py tao_detection_release/mmdet/core/bbox/__init__.py tao_detection_release/mmdet/models/mask_heads/maskiou_head.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_reweighthead_bfocal.py tao_detection_release/configs/baselines/detectors_dconv_c3-c5_mstrain_400_1400_x101_32x4d.py tao_detection_release/mmdet/models/utils/weight_init.py tao_detection_release/configs/tao/htc_x101_64x4d_fpn_20e_16gpu_tao_test.py tao_detection_release/mmdet/datasets/voc.py tao_detection_release/mmdet/models/losses/cross_entropy_loss.py tao_detection_release/mmdet/models/bbox_heads/double_bbox_head.py tao_detection_release/.history/tools/multi_process_inference/inference_20200804113707.py tao_detection_release/mmdet/models/detectors/mask_scoring_rcnn.py tao_detection_release/.history/tools/multi_process_inference/inference_20200802135019.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804114158.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804111837.py tao_detection_release/mmdet/models/losses/balanced_l1_loss.py tao_detection_release/mmdet/models/detectors/single_stage.py tao_detection_release/mmdet/models/anchor_heads/ga_retina_head.py tao_detection_release/mmdet/models/detectors/fast_rcnn.py tao_detection_release/mmdet/models/backbones/__init__.py tao_detection_release/mmdet/ops/nms/nms_wrapper.py tao_detection_release/mmdet/models/bbox_heads/reweight_bbox_head.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812114009.py tao_detection_release/lvis-api/lvis/lvis.py tao_detection_release/mmdet/core/evaluation/class_names.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812114449.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812113847.py tao_detection_release/mmdet/ops/sigmoid_focal_loss/sigmoid_focal_loss.py tao_detection_release/mmdet/models/backbones/ssd_vgg.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812113911.py tao_detection_release/mmdet/models/anchor_heads/ssd_head.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804114157.py tao_detection_release/mmdet/datasets/cityscapes.py tao_detection_release/lvis-api/lvis/tao_eval.py tao_detection_release/lvis-api/lvis/colormap.py tao_detection_release/mmdet/datasets/lvis.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0_bg1_tunehead.py tao_detection_release/tools/get_flops.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812114016.py tao_detection_release/mmdet/models/detectors/DCM.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_add12epoch.py tao_tracking_release/tao_post_processing/create_json_for_eval.py tao_detection_release/mmdet/models/bbox_heads/gs_bbox_head_with0.py tao_detection_release/.history/tools/lvis_tao_analyse_20200812113412.py tao_detection_release/mmdet/models/builder.py tao_detection_release/mmdet/utils/__init__.py tao_detection_release/mmdet/models/detectors/rfp.py tao_detection_release/mmdet/core/evaluation/coco_utils.py tao_tracking_release/deep_sort/track.py tao_detection_release/mmdet/core/anchor/anchor_target.py tao_detection_release/mmdet/datasets/pipelines/test_aug.py tao_detection_release/mmdet/core/bbox/assigners/__init__.py tao_detection_release/mmdet/datasets/pipelines/compose.py tao_detection_release/mmdet/models/detectors/group_softmax.py tao_detection_release/mmdet/utils/registry.py tao_detection_release/lvis-api/lvis/vis.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812113933.py tao_tracking_release/reid_pytorch/reid_extractor.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804111838.py tao_detection_release/lvis-api/lvis/__init__.py tao_detection_release/mmdet/core/bbox/assigners/point_assigner.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110826.py tao_detection_release/mmdet/core/__init__.py tao_detection_release/mmdet/models/necks/hrfpn.py tao_detection_release/mmdet/__init__.py tao_detection_release/mmdet/models/utils/scale.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0_reweight.py tao_detection_release/mmdet/ops/roi_align/roi_align.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0_bgn.py tao_detection_release/mmdet/models/backbones/hrnet.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200728161844.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110648.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0_bg8_8bin.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110715.py tao_detection_release/mmdet/ops/dcn/__init__.py tao_detection_release/tools/multi_process_inference/multi_inference.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_is.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_finetunefewshot.py tao_detection_release/mmdet/models/detectors/__init__.py tao_detection_release/mmdet/models/anchor_heads/rpn_head.py tao_detection_release/lvis-api/lvis/results.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812114342.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110825.py tao_detection_release/tools/test_robustness.py tao_detection_release/mmdet/models/detectors/faster_rcnn.py tao_detection_release/mmdet/models/utils/norm.py tao_detection_release/.history/tools/multi_process_inference/inference_20200801212626.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110806.py tao_tracking_release/track_tao.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110632.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0_bg1.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812113857.py tao_tracking_release/generate_detection_features.py tao_detection_release/mmdet/core/evaluation/recall.py tao_detection_release/mmdet/models/roi_extractors/single_level.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110644.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812114456.py tao_tracking_release/tao_post_processing/track_concat.py tao_detection_release/mmdet/ops/roi_align/gradcheck.py tao_tracking_release/tao_post_processing/onekey_postprocessing.py tao_detection_release/mmdet/apis/inference.py tao_tracking_release/deep_sort/nn_matching.py tao_detection_release/mmdet/core/anchor/point_generator.py tao_detection_release/mmdet/models/anchor_heads/guided_anchor_head.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_reweighthead_bours.py tao_detection_release/mmdet/core/bbox/samplers/random_sampler.py tao_detection_release/.history/tools/multi_process_inference/inference_20200802135033.py tao_detection_release/configs/tao/gs_cascade_rcnn_x101_64x4d_fpn_1x_lvis.py tao_detection_release/mmdet/models/necks/fpn.py tao_detection_release/.history/tools/multi_process_inference/inference_20200804113708.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_focalloss_all.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812113841.py tao_detection_release/mmdet/models/backbones/resnext.py tao_tracking_release/deep_sort/linear_assignment.py tao_detection_release/mmdet/models/bbox_heads/__init__.py tao_detection_release/mmdet/core/bbox/assigners/assign_result.py tao_detection_release/mmdet/datasets/xml_style.py tao_detection_release/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110717.py tao_detection_release/mmdet/datasets/registry.py tao_detection_release/lvis-api/test.py tao_detection_release/mmdet/models/detectors/test_mixins.py tao_detection_release/mmdet/models/losses/smooth_l1_loss.py tao_tracking_release/tao_post_processing/associate_track.py tao_detection_release/tools/robustness_eval.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_reweighthead.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0_bg1_tuneall.py tao_detection_release/configs/baselines/faster_rcnn_r50_fpn_1x_lvis.py tao_detection_release/mmdet/models/backbones/resnet.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0_bg1_trainhead.py tao_detection_release/mmdet/core/bbox/assigners/approx_max_iou_assigner.py tao_detection_release/mmdet/datasets/pipelines/formating.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110642.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804110859.py tao_detection_release/mmdet/models/necks/__init__.py tao_detection_release/mmdet/core/bbox/samplers/pseudo_sampler.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200801212617.py tao_detection_release/mmdet/core/post_processing/merge_augs.py tao_detection_release/mmdet/models/detectors/retinanet.py tao_detection_release/mmdet/models/losses/mse_loss.py tao_detection_release/mmdet/models/bbox_heads/bbox_head.py tao_detection_release/mmdet/utils/flops_counter.py tao_tracking_release/deep_sort/tracker.py tao_detection_release/mmdet/models/detectors/mask_rcnn.py tao_detection_release/mmdet/datasets/coco.py tao_tracking_release/deep_sort/detection.py tao_detection_release/configs/tao/htc_x101_64x4d_fpn_20e_16gpu_lvis.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804110913.py tao_detection_release/mmdet/ops/context_block.py tao_detection_release/mmdet/models/mask_heads/grid_head.py tao_detection_release/mmdet/models/losses/utils.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804110857.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_is_finetune.py tao_detection_release/mmdet/core/fp16/decorators.py tao_detection_release/mmdet/ops/nms/__init__.py tao_detection_release/mmdet/models/backbones/resnext_rs.py tao_detection_release/configs/baselines/mask_rcnn_r50_fpn_1x_lvis.py tao_detection_release/mmdet/core/bbox/samplers/iou_balanced_neg_sampler.py tao_detection_release/mmdet/core/post_processing/bbox_nms.py tao_detection_release/mmdet/datasets/loader/build_loader.py tao_detection_release/mmdet/models/losses/accuracy.py tao_detection_release/mmdet/models/mask_heads/__init__.py tao_detection_release/mmdet/models/plugins/__init__.py tao_detection_release/mmdet/models/detectors/htc.py tao_detection_release/configs/baselines/htc_x101_64x4d_fpn_20e_16gpu_lvis.py tao_detection_release/mmdet/datasets/pipelines/__init__.py tao_detection_release/.history/tools/multi_process_inference/inference_20200804110846.py tao_detection_release/mmdet/models/roi_extractors/__init__.py tao_detection_release/mmdet/apis/env.py tao_detection_release/mmdet/core/evaluation/bbox_overlaps.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110706.py tao_detection_release/configs/tao/gs_htc_x101_64x4d_fpn_20e_16gpu_lvis.py tao_detection_release/.history/tools/multi_process_inference/inference_20200802135034.py tao_tracking_release/application_util/__init__.py tao_detection_release/mmdet/core/bbox/transforms.py tao_detection_release/.history/configs/tao/gs_htc_dconv_x101_64x4d_fpn_1x_lvis_taocats_20200812114003.py tao_detection_release/mmdet/models/anchor_heads/reppoints_head.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804110910.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804114156.py tao_detection_release/configs/transferred/faster_rcnn_r50_fpn_1x_lvis_is_with0bg3.py tao_detection_release/configs/ablations/gs_faster_rcnn_r50_fpn_1x_lvis_with0_bg8_2bin.py tao_detection_release/.history/tools/multi_process_inference/inference_20200802135016.py tao_detection_release/mmdet/models/plugins/generalized_attention.py tao_detection_release/mmdet/datasets/loader/__init__.py tao_detection_release/mmdet/models/utils/saconv.py tao_detection_release/mmdet/ops/roi_align/__init__.py tao_detection_release/mmdet/apis/__init__.py tao_detection_release/mmdet/datasets/builder.py tao_detection_release/mmdet/core/bbox/assigners/base_assigner.py tao_detection_release/mmdet/models/losses/iou_loss.py tao_detection_release/configs/bags/gs_faster_rcnn_r50_fpn_1x_lvis_with0_bg8.py tao_detection_release/mmdet/ops/masked_conv/masked_conv.py tao_detection_release/mmdet/core/bbox/assigners/max_iou_assigner.py tao_detection_release/configs/tao/faster_rcnn_x101_64x4d_fpn_1x_tao.py tao_detection_release/configs/baselines/faster_rcnn_x101_64x4d_fpn_1x_lvis.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110711.py tao_detection_release/configs/tao/gs_faster_rcnn_x101_64x4d_fpn_1x_lvis.py tao_detection_release/mmdet/models/registry.py tao_detection_release/mmdet/models/bbox_heads/DCM_bbox_head.py tao_detection_release/mmdet/models/__init__.py tao_tracking_release/deep_sort_app.py tao_detection_release/mmdet/ops/__init__.py tao_detection_release/mmdet/models/mask_heads/fcn_mask_head.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110822.py tao_detection_release/configs/bags/gs_cascade_rcnn_x101_64x4d_fpn_1x_lvis.py tao_detection_release/configs/baselines/cascade_rcnn_x101_64x4d_fpn_1x_lvis.py tao_detection_release/mmdet/core/anchor/point_target.py tao_detection_release/.history/tools/lvis_tao_analyse_20200812113309.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110805.py tao_detection_release/mmdet/models/utils/conv_ws.py tao_detection_release/.dev_scripts/benchmark_filter.py tao_detection_release/mmdet/core/evaluation/tao_utils.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804114150.py tao_detection_release/.history/tools/multi_process_inference/multi_inference_20200804114155.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110829.py tao_detection_release/mmdet/models/bbox_heads/gs_bbox_head_with0_reweight.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110716.py tao_detection_release/mmdet/core/bbox/samplers/base_sampler.py tao_detection_release/mmdet/models/detectors/fovea.py tao_detection_release/.history/tools/multi_process_inference/inference_20200802135007.py tao_detection_release/lvis-api/lvis/eval.py tao_detection_release/mmdet/apis/train.py tao_detection_release/mmdet/core/anchor/anchor_generator.py tao_detection_release/mmdet/models/backbones/resnet_ori.py tao_detection_release/mmdet/core/post_processing/__init__.py tao_detection_release/configs/baselines/untitled.py tao_detection_release/mmdet/models/utils/conv_module.py tao_detection_release/tools/eval_lvis.py tao_detection_release/.history/configs/tao/gs_htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis_20200804110828.py tao_detection_release/configs/baselines/htc_dconv_c3-c5_mstrain_400_1400_x101_64x4d_fpn_20e_lvis.py tao_detection_release/mmdet/datasets/custom.py tao_detection_release/mmdet/models/detectors/two_stage.py make_cuda_ext write_version_py readme get_requirements get_version get_git_hash get_hash make_cython_ext main parse_args get_final_epoch process_checkpoint get_final_results main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args single_process single_process single_process single_process single_process single_process single_process single_process single_process single_process single_process single_process single_process single_process single_process single_process single_process single_process colormap Params LVISEval LVIS LVISResults TAOEval Params LVISVis _init_dist_pytorch _init_dist_slurm init_dist set_random_seed get_root_logger _init_dist_mpi inference_detector show_result_pyplot LoadImage init_detector show_result _dist_train build_optimizer batch_processor _non_dist_train train_detector parse_losses AnchorGenerator anchor_target unmap anchor_inside_flags images_to_levels anchor_target_single ga_loc_target ga_shape_target_single calc_region images_to_levels ga_shape_target PointGenerator images_to_levels point_target unmap point_target_single assign_and_sample build_assigner build_sampler bbox_target_single expand_target bbox_target bbox_overlaps delta2bbox roi2bbox bbox_flip distance2bbox bbox2delta bbox_mapping bbox2result bbox_mapping_back bbox2roi ApproxMaxIoUAssigner AssignResult BaseAssigner MaxIoUAssigner PointAssigner BaseSampler CombinedSampler InstanceBalancedPosSampler IoUBalancedNegSampler OHEMSampler PseudoSampler RandomSampler SamplingResult bbox_overlaps get_classes imagenet_vid_classes voc_classes imagenet_det_classes coco_classes cityscapes_classes wider_face_classes coco_eval segm2json proposal2json fast_eval_recall xyxy2xywh results2json det2json CocoDistEvalRecallHook DistEvalmAPHook DistEvalHook CocoDistEvalmAPHook lvis_fast_eval_recall lvis_eval segm2json proposal2json xyxy2xywh results2json det2json eval_map tpfp_imagenet print_map_summary average_precision get_cls_results tpfp_default plot_iou_recall set_recall_param print_recall_summary _recalls eval_recalls plot_num_recall tao_fast_eval_recall segm2json tao_eval proposal2json xyxy2xywh results2json det2json force_fp32 auto_fp16 Fp16OptimizerHook wrap_fp16_model patch_forward_method patch_norm_fp32 cast_tensor_type mask_target mask_target_single split_combined_polys multiclass_nms multiclass_nms_tao merge_aug_scores merge_aug_masks merge_aug_bboxes merge_aug_proposals DistOptimizerHook allreduce_grads _allreduce_coalesced unmap tensor2imgs multi_apply build_dataset _concat_dataset CityscapesDataset CocoDataset CustomDataset RepeatDataset ConcatDataset FilesDataset LvisDataset TaoDataset VOCDataset WIDERFaceDataset XMLDataset build_dataloader GroupSampler_addrepeat GroupSampler DistributedGroupSampler_addrepeat DistributedSampler DistributedGroupSampler_addrepeat_sampleout DistributedGroupSampler EpisodicSampler Compose DefaultFormatBundle Transpose ToTensor Collect to_tensor ImageToTensor ToDataContainer LoadImageFromFile LoadProposals LoadAnnotations MultiScaleFlipAug RandomFlip Pad Corrupt PhotoMetricDistortion MinIoURandomCrop Resize RandomCrop Albu SegResizeFlipPadRescale Normalize Expand build_shared_head build_detector build_loss build build_backbone build_roi_extractor build_head build_neck AnchorHead FCOSHead FeatureAlign FoveaHead GARetinaHead GARPNHead FeatureAdaption GuidedAnchorHead RepPointsHead RetinaHead RPNHead SSDHead HRModule HRNet ResNet BasicBlock make_res_layer Bottleneck ResNet BasicBlock make_res_layer Bottleneck ResNeXt make_res_layer Bottleneck ResNeXt make_res_layer Bottleneck ResNeXt make_res_layer Bottleneck SSDVGG L2Norm BBoxHead SharedFCBBoxHead ConvFCBBoxHead DCMBBoxHead DoubleConvFCBBoxHead BasicResBlock GSBBoxHeadWith0 GSBBoxHeadWith0Reweight ReweightBBoxHead BaseDetector CascadeRCNN DCM DoubleHeadRCNN FasterRCNN FastRCNN FCOS FOVEA GridRCNN GroupSoftmax HybridTaskCascade MaskRCNN MaskScoringRCNN RepPointsDetector RetinaNet RecursiveFeaturePyramid ASPP RPN SingleStageDetector MaskTestMixin BBoxTestMixin RPNTestMixin TwoStageDetector Accuracy accuracy BalancedL1Loss balanced_l1_loss binary_cross_entropy mask_cross_entropy _expand_binary_labels CrossEntropyLoss cross_entropy sigmoid_focal_loss py_sigmoid_focal_loss FocalLoss _expand_binary_labels GHMR GHMC bounded_iou_loss BoundedIoULoss iou_loss IoULoss MSELoss smooth_l1_loss SmoothL1Loss weight_reduce_loss weighted_loss reduce_loss FCNMaskHead FusedSemanticHead GridHead HTCMaskHead MaskIoUHead BFP FPN HRFPN GeneralizedAttention NonLocal2D SingleRoIExtractor ResLayer ConvModule build_conv_layer ConvAWS2d conv_ws_2d ConvWS2d build_norm_layer SAConv2d Scale xavier_init bias_init_with_prob uniform_init normal_init kaiming_init last_zero_init ContextBlock DeformConvFunction ModulatedDeformConv DeformConvPack ModulatedDeformConvPack DeformConv ModulatedDeformConvFunction DeformRoIPoolingPack DeformRoIPoolingFunction ModulatedDeformRoIPoolingPack DeformRoIPooling MaskedConv2dFunction MaskedConv2d nms soft_nms RoIAlign RoIAlignFunction RoIPool RoIPoolFunction SigmoidFocalLoss SigmoidFocalLossFunction add_flops_counting_methods add_flops_counter_hook_function bn_flops_counter_hook reset_flops_count deconv_flops_counter_hook relu_flops_counter_hook get_model_parameters_number add_flops_mask flops_to_string params_to_string remove_flops_mask remove_batch_counter_hook_function start_flops_count add_batch_counter_variables_or_reset pool_flops_counter_hook empty_flops_counter_hook add_flops_mask_variable_or_reset add_batch_counter_hook_function get_model_complexity_info conv_flops_counter_hook remove_flops_counter_hook_function batch_counter_hook add_flops_counter_variable_or_reset is_supported_instance stop_flops_count upsample_flops_counter_hook linear_flops_counter_hook compute_average_flops_cost print_model_with_flops build_from_cfg Registry accumulate_acc get_split_bin main parse_args multi_gpu_test reweight_cls single_gpu_test collect_results save_in_tao_format main parse_args get_cate_gs2 ana_coco_param get_hist get_split get_val get_draw_val_imgs get_bin_weight ana_param load_checkpoint_all del_tail get_cate_weight_bours get_cate_gs8 get_cate_gs update_cls del_nondense_cls get_dense_det construct_data test_node_map get_split2 trymapping load_checkpoint get_mask get_cate_weight_bf count_ins get_split8 get_cate_weight get_cate_gs2 ana_coco_param get_hist get_split get_val get_draw_val_imgs get_bin_weight ana_param load_checkpoint_all del_tail get_cate_weight_bours get_cate_gs8 get_cate_gs update_cls del_nondense_cls get_dense_det construct_data test_node_map get_split2 trymapping load_checkpoint get_mask get_cate_weight_bf count_ins get_split8 get_cate_weight get_distortions_from_results print_coco_results get_distortions_from_file get_coco_style_results get_voc_style_results get_results main multi_gpu_test reweight_cls single_gpu_test collect_results main parse_args reweight_cls single_gpu_test accumulate_acc collect_results get_split_bin reweight_cls_newhead main multi_gpu_test parse_args voc_eval_with_return single_gpu_test collect_results coco_eval_with_return main multi_gpu_test parse_args multi_gpu_test reweight_cls single_gpu_test collect_results main parse_args select_cascade_cls_params select_training_param select_mask_params main parse_args select_head multi_gpu_test reweight_cls parse_textfile single_gpu_test collect_results save_in_tao_format main parse_args single_process dump_testfile chunks parse_textfile create_detections gather_sequence_info run BoundingBoxDataset main infer init_model save_detections main save_detections_star main track_and_save_star track_and_save is_in_bounds ImageViewer view_roi create_unique_color_float Visualization create_unique_color_uchar NoVisualization Detection min_cost_matching matching_cascade _cosine_distance NearestNeighborDistanceMetric _nn_cosine_distance _nn_euclidean_distance _pdist Track TrackState Tracker change_input_dim ReID_Inference apply noverlap track_and_save_star associate track_and_save cosine_similarity main reid_similarity track_associate main remove_1len_track create_json main track_and_save_star track_and_save track_concat main track_and_save_star track_and_save decode _minimal_ext_cmd exists get_hash cythonize Extension format realpath dirname add_argument ArgumentParser join data_pipeline dump name print dict datasets scandir basic_arch append parse_args nn_module load decode format rstrip save Popen items dict mkdir_or_exist list get_final_epoch format rstrip copy process_checkpoint get_final_results root out exists update show_result size ProgressBar eval append dataset range enumerate len update get_dist_info size collect_results ProgressBar eval append dataset range enumerate len rstrip tensor broadcast list get_dist_info mkdtemp encode append range dump format bytearray zip load join barrier extend rmtree mkdir_or_exist full str local_rank format print copy_ pnorm state_dict append enumerate config batch_size load_filepaths model save_in_tao_format launcher parse_testfile cuda dirname build_detector fromfile get replace init_dist build_dataloader wrap_fp16_model gpuid eval test_pipeline img_dir checkpoint enumerate load_checkpoint makedirs img_list out_dir FilesDataset MMDataParallel print check_output dump_testfile reshape astype float32 _init_dist_mpi set_start_method _init_dist_slurm _init_dist_pytorch int set_device init_process_group device_count int str format init_process_group set_device device_count getoutput seed manual_seed_all manual_seed basicConfig setLevel get_dist_info getLogger get_classes isinstance model load_checkpoint warn eval build_detector fromfile to Compose cfg dict test_pipeline device bool concat_list isinstance concatenate imshow_det_bboxes astype copy vstack randint imread bgr2rgb imshow show_result figure items list isinstance OrderedDict mean item Tensor sum dict parse_losses model log_level _non_dist_train get_root_logger _dist_train pop get hasattr endswith search copy named_parameters getattr append optim module workflow log_level use_img_sampling MMDistributedDataParallel DistSamplerSeedHook setLevel cuda run total_epochs issubclass build_optimizer checkpoint_config work_dir module update get val format CocoDistEvalRecallHook load_from use_sample_out resume_from register_training_hooks resume type INFO optimizer DistOptimizerHook lr_config DistEvalmAPHook isinstance CocoDataset print load_checkpoint register_hook CocoDistEvalmAPHook Runner log_config Fp16OptimizerHook workflow log_level setLevel cuda run total_epochs build_optimizer checkpoint_config work_dir optimizer_config get load_from resume_from register_training_hooks resume INFO optimizer lr_config load_checkpoint Runner log_config Fp16OptimizerHook multi_apply images_to_levels any sum range cat len append stack squeeze assign_and_sample zeros_like PseudoSampler pos_gt_bboxes size pos_weight anchor_inside_flags unmap sample new_zeros build_assigner assign pos_inds bbox2delta allowed_border neg_inds pos_bboxes assigner uint8 type new_full clamp long new_full zeros_like calc_region size sqrt log2 floor full_like item append zeros float sum long range len multi_apply images_to_levels any append sum range cat len ga_assigner build_sampler ga_sampler PseudoSampler zeros_like reshape pos_gt_bboxes size unmap build_assigner assign pos_inds sample neg_inds pos_bboxes multi_apply images_to_levels any sum range cat len assign_and_sample zeros_like PseudoSampler pos_gt_bboxes size pos_weight unmap new_zeros build_assigner assign pos_inds sample neg_inds assigner BaseAssigner isinstance BaseSampler isinstance build_sampler sampler build_assigner assign sample assigner multi_apply cat bbox2delta size new_zeros squeeze new_zeros clamp size min max stack unsqueeze div_ float log exp clamp size repeat expand_as view_as abs log addcmul Tensor ndarray isinstance clone bbox_flip new_full new_zeros append cat enumerate cpu append unique numpy clamp minimum T astype maximum float32 zeros range items list eval is_str list format isinstance COCOeval print evaluate summarize is_str COCO accumulate getImgIds loadRes fast_eval_recall array enumerate load getAnnIds is_str mean getImgIds eval_recalls append zeros loadAnns array range len tolist dict append float xyxy2xywh range len dict append float xyxy2xywh range len decode isinstance dict append float xyxy2xywh range len dump format ndarray isinstance segm2json dict proposal2json det2json update dump format list print lvis_fast_eval_recall LVISEval is_str open run print_results LVIS array range enumerate load range is_str mean eval_recalls append get_img_ids get_ann_ids array zeros load_anns len print stack arange ones hstack maximum zeros sum range minimum zeros_like len argsort zeros bbox_overlaps range enumerate zeros_like len argsort bbox_overlaps zeros argmax max enumerate append zeros range len eps cumsum tuple maximum average_precision argsort enumerate mean any vstack print_map_summary item zip append zeros range get_cls_results len get_classes table print len is_str AsciiTable append zeros range enumerate sum sort hstack copy zeros float argmax fliplr range enumerate array isinstance min set_recall_param print_recall_summary _recalls array append zeros bbox_overlaps range len arange table insert print size tolist AsciiTable append array enumerate show ndarray plot isinstance xlabel tolist axis ylabel figure show ndarray plot isinstance xlabel tolist axis ylabel figure update dump format tao_fast_eval_recall TAOEval list print is_str open run print_results LVIS array range enumerate load range is_str mean eval_recalls append get_img_ids get_ann_ids array zeros load_anns len hasattr patch_norm_fp32 modules half children isinstance half patch_forward_method float forward ndarray isinstance Iterable Tensor Mapping list map cat mask_size imresize size astype maximum new_zeros int32 device append to numpy range _pair tolist append slice_list range len pop new_full sort copy nms_op new_zeros getattr append range cat pop new_full sort copy nms_op new_zeros getattr append range cat nms nms_thr sort min clone max_num zip append bbox_mapping_back cat append mean bbox_mapping_back zip Tensor isinstance average mean array list _take_tensors _flatten_dense_tensors zip _unflatten_dense_tensors OrderedDict all_reduce copy_ div_ append type values all_reduce _allreduce_coalesced get_world_size div_ uint8 transpose size astype ascontiguousarray append array range list map get deepcopy isinstance append build_dataset range len isinstance ConcatDataset _concat_dataset build_from_cfg RepeatDataset get pop get_dist_info DistributedGroupSampler_addrepeat print DistributedSampler DataLoader DistributedGroupSampler_addrepeat_sampleout DistributedGroupSampler Tensor ndarray isinstance isinstance block Sequential build_conv_layer append range expansion topk isinstance size t eq mul_ expand_as append sum max log abs e where float weight_reduce_loss new_full size squeeze expand size weight_reduce_loss binary_cross_entropy_with_logits _expand_binary_labels float squeeze arange type_as sigmoid pow weight_reduce_loss binary_cross_entropy_with_logits _sigmoid_focal_loss weight_reduce_loss view clamp view zeros_like size min where abs max abs where get_enum sum reduce_loss dict conv_layer pop copy size view pop str setdefault norm_layer copy parameters _specify_ddp_gpu_num hasattr bias xavier_uniform_ xavier_normal_ weight constant_ hasattr bias normal_ weight constant_ hasattr bias uniform_ weight constant_ kaiming_uniform_ hasattr bias weight kaiming_normal_ constant_ float Sequential isinstance constant_init ndarray isinstance new_zeros Tensor to numpy is_cuda ndarray soft_nms_cpu isinstance Tensor numpy flops_model get_model_parameters_number input_constructor stop_flops_count add_flops_counting_methods start_flops_count compute_average_flops_cost new_empty print_model_with_flops print compute_average_flops_cost apply sum __get__ reset_flops_count apply __batch_counter__ is_supported_instance modules add_batch_counter_hook_function apply remove_batch_counter_hook_function apply add_batch_counter_variables_or_reset apply apply apply isinstance numel shape affine prod groups kernel_size out_channels in_channels list kernel_size out_channels groups in_channels expand sum prod print len register_forward_hook hasattr remove hasattr is_supported_instance register_forward_hook is_supported_instance isinstance hasattr remove is_supported_instance hasattr is_supported_instance pop get list items setdefault copy is_str isclass print format exists items list format print float64 astype forward_dummy hasattr tuple get_model_complexity_info shape reweight_cls tmpdir MMDistributedDataParallel exit tau multi_gpu_test single_gpu_test zip type len items list cats from_numpy save LVIS zeros enumerate items list arange cats LVIS zeros array append load norm cats LVIS numpy array load items list norm hstack COCO zeros numpy load norm numpy mean load norm append zeros numpy range enumerate items list cats from_numpy save LVIS zeros load print from_numpy array range items list arange cats LVIS zeros array append items list ones_like cats where mean from_numpy save LVIS zeros items list ones_like cats power set_trace mean from_numpy save LVIS zeros sum items list ones_like ones cats power set_trace where mean from_numpy save LVIS zeros items list ones_like hstack cats set_trace where mean LVIS zeros numpy range append items list cats set_trace from_numpy save LVIS zeros enumerate items list arange cats LVIS zeros array append items list cats set_trace from_numpy save LVIS zeros enumerate items list arange cats LVIS zeros array append items list print cats add set img_ann_map LVIS append len items list dict img_ann_map LVIS len items list add set img_ann_map LVIS append len items list cats add set LVIS append items list int format join print add set choice img_ann_map mkdir LVIS append len join set_trace add set append items list cats set_trace LVIS join set_trace add set mkdir LVIS append join set_trace add set mkdir append len zeros _print load list format basename print_coco_results isinstance print mean zeros keys enumerate len load list format basename isinstance print mean zeros keys enumerate len print get_coco_style_results get_voc_style_results load append replace enumerate task add_argument get_results ArgumentParser filename ann_file show get_dist_info lvis_eval build_dataset CLASSES test json_out results2json lvis items list arange cats zeros array append get_ann_info build_assigner assign cuda cat_ids accumulate_acc hstack astype get_split_bin assigner int zeros numpy join list format items state_dict print copy_ startswith pnorm split reweight_cls_newhead list format evaluate COCOeval print summarize is_str COCO accumulate getImgIds loadRes stats fast_eval_recall array enumerate load eval_map CLASSES concatenate get_ann_info vstack append range len img_norm_cfg set_random_seed coco_eval_with_return final_prints seed corruptions voc_eval_with_return final_prints_aggregate obj_from_dict workers_per_gpu insert iou_thr deepcopy coco severities summaries tao_eval parameters parameters parameters children parameters log_level select_training_param autoscale_lr train_detector select_cascade_cls_params get_root_logger work_dir select_mask_params val resume_from info gpus select_head parse_textfile split join name close write open ceil int float len TemporaryDirectory int min max append int astype Detection NearestNeighborDistanceMetric NoVisualization gather_sequence_info Tracker str BoundingBoxDataset str concatenate print hstack apply DataLoader mkdir read_csv save ReID_Inference values init_model detection_file defaultdict imap_unordered mkdir detections_dir items infer tqdm FixedGpuPool get_output_path str sorted with_suffix print DataFrame extend to_csv mkdir zip append numpy array save_detections workers save_detections Pool load defaultdict concatenate ones reshape len astype hstack mkdir unique save run array append count track_and_save track_and_save output_dir create_json hsv_to_rgb create_unique_color_float linear_assignment arange append distance_metric enumerate len list set range min_cost_matching len T inf dot float clip norm asarray _pdist _cosine_distance load transform save norm mean len vstack unique append range reid_similarity noverlap concatenate astype copy associate unique append track_associate results_dir append unique concatenate load remove_1len_track sorted defaultdict update print extend tqdm set append keys track_result track_concat onnx_results1 onnx_results2 defaultdict astype copy vstack unique append count results_dir2 results_dir1 | # Code for the winner of TAO 2020 ## Note: This repo is **NOT** a fully reinplementation of our submitted result([[email protected]](https://motchallenge.net/results/TAO_Challenge/)). Instead, two models for appearance modeling are included, together with the open-source [BAGS](https://github.com/FishYuLi/BalancedGroupSoftmax) model and the full set of code for inference. With this code, you can achieve around mAP@23 with TAO test set (based on our estimation). # Start from here: Please find instructions for the detection model in [tao_detection_release/README.md](tao_detection_release/README.md) and the tracking model in [tao_tracking_release/README.md](tao_tracking_release/README.md) # Result file： The result for the TAO validation set can be downloaded in [TAO_val](https://drive.google.com/file/d/1NGLcQ40ci2MfbJK34hq37mv9iTOXV2sh/view?usp=sharing). # License | 2,080 |
felipecode/coiltraine | ['imitation learning', 'autonomous driving'] | ['Exploring the Limitations of Behavior Cloning for Autonomous Driving', 'On Offline Evaluation of Vision-based Driving Models'] | carla08/driving_benchmark/experiment_suites/corl_2017.py network/models/building_blocks/__init__.py carla08/image_converter.py configs/namer.py logger/tensorboard_logger.py drive/CoILBaseline.py drive/__init__.py tools/batch_rename.py coilutils/checkpoint_schedule.py tools/analize_results.py plotter/plotting_params/eccv_online_offline_plots.py carla08/client.py drive/coil_agent.py plotter/plot_on_map.py carla08/agent/modules/waypointer.py drive/suites/nocrash_training_suite.py coilutils/exporter.py tools/get_sample_datasets.py plotter/metrics.py model_view/carla08interface.py drive/suites/nocrash_new_weather_suite.py network/loss.py carla08/planner/__init__.py carla08/driving_benchmark/driving_benchmark.py carla08/agent/modules/controllers.py drive/suites/eccv_training_suite.py logger/monitorer.py drive/suites/nocrash_new_weather_town_suite.py carla08/util.py coil_core/train.py tools/download_sample_models.py coilutils/drive_utils.py carla08/planner/map.py logger/json_formatter.py carla08/driving_benchmark/experiment_suites/basic_experiment_suite.py input/coil_dataset.py carla08/driving_benchmark/experiment_suites/longcontrol_2018.py carla08/agent/modules/utils.py coilutils/attribute_dict.py tools/plot_predictions_on_images.py logger/printer.py coilutils/general.py tools/get_offline_online_data.py tools/get_baseline_dataset.py coiltraine.py manipulate_experiments.py coil_core/run_drive.py network/__init__.py carla08/planner/city_track.py tools/download_tools.py plotter/plotter.py carla08/planner/astar.py carla08/planner/converter.py carla08/transform.py carla08/driving_benchmark/results_printer.py plotter/__init__.py drive/suites/nocrash_new_town_suite.py run_plotting.py carla08/planner/graph.py drive/suites/corl_new_weather_town_suite.py logger/coil_logger.py coil_core/validate.py network/models/coil_icra.py network/models/building_blocks/conv.py plotter/scatter_plotter.py carla08/driving_benchmark/experiment_suites/__init__.py network/optimizer.py network/models/__init__.py configs/__init__.py carla08/settings.py carla08/agent/forward_agent.py drive/suites/test_t1_suite.py tools/screen_manager.py carla08/driving_benchmark/metrics.py drive/suites/corl_new_weather_suite.py coilutils/experiment_schedule.py coilutils/checking.py logger/carla_metrics_parser.py plotter/plotting_params/plotting_all_cameras.py coilutils/__init__.py carla08/agent/command_follower.py carla08/agent/lane_follower.py drive/suites/test_t2_suite.py carla08/agent/human_agent.py tools/plot_offline_evaluation.py plotter/plotting_params/sample_plot.py network/models/building_blocks/resnet.py input/data_parser.py carla08/agent/__init__.py input/augmenter.py coil_core/__init__.py carla08/driving_benchmark/__init__.py plotter/data_reading.py input/scheduler.py drive/suites/eccv_generalization_suite.py carla08/agent/modules/__init__.py input/__init__.py carla08/agent/modules/obstacle_avoidance.py carla08/sensor.py testing/full_tests/test_load_data.py carla08/driving_benchmark/experiment.py coil_core/executer.py view_model.py network/models/building_blocks/join.py carla08/planner/grid.py configs/coil_global.py drive/suites/corl_new_town_suite.py tools/download_nocrash_models.py carla08/tcp.py tools/get_town03_dataset.py carla08/planner/planner.py network/models/building_blocks/branching.py tools/view_npy.py carla08/agent/agent.py carla08/planner/bezier.py carla08/driving_benchmark/experiment_suites/experiment_suite.py logger/__init__.py network/coil_model.py network/loss_functional.py drive/suites/corl_training_suite.py carla08/carla_server_pb2.py carla08/driving_benchmark/recording.py input/coil_sampler.py input/splitter.py model_view/carla09interface.py network/models/building_blocks/fc.py make_carla_client _make_sensor_parsers CarlaClient labels_to_cityscapes_palette to_bgra_array depth_to_array depth_to_logarithmic_grayscale depth_to_local_point_cloud to_rgb_array labels_to_array Image Lidar SensorData Camera LidarMeasurement _append_extension Sensor PointCloud CarlaSettings TCPConnectionError TCPClient Transform make_connection print_over_same_line to_hex_str StopWatch Agent CommandFollower ForwardAgent HumanAgent LaneFollower Controller ObstacleAvoidance get_vec_dist get_angle sldist Waypointer angle_between sldist get_vec_dist DrivingBenchmark run_driving_benchmark Experiment Metrics Recording print_summary BasicExperimentSuite CoRL2017 ExperimentSuite LongitudinalControl2018 AStar Cell bezier_curve bernstein_poly CityTrack Converter string_to_floats sldist Graph sldist3 string_to_node Grid angle_between CarlaMap color_to_angle sldist angle_between Planner signal compare AttributeDict _is_tensor_image do_assert is_callable _is_numpy_image is_single_number is_hdf5_prepared check_loss_validation_stopped is_ready_to_save is_next_checkpoint_ready maximun_checkpoint_reach is_open get_next_checkpoint validation_stale_point get_latest_saved_checkpoint get_latest_evaluated_checkpoint checkpoint_parse_configuration_file get_gpu_resources get_remainig_exps execvec_to_names allocate_gpu_resources mount_experiment_heap dict_to_namevec export_csv_separate export_csv export_status erase_logs erase_wrong_plotting_summaries plot_test_image alphanum_key camelcase_to_snakecase write_data_point_control_summary compute_average_std command_number_to_index static_vars compute_average_std_separatetasks create_exp_path create_log_folder write_header_control_summary get_validation_datasets erase_validations get_latest_path snakecase_to_camelcase softmax unique tryint send_email get_driving_environments sort_nicely execute_validation folder_execute execute_drive execute_train execute driving_benchmark start_carla_simulator find_free_port execute execute write_waypoints_output write_regular_output get_names merge_with_yaml _merge_a_into_b _decode_cfg_value set_type_of_process _check_and_coerce_cfg_value_type generate_name parse_split_configuration get_dropout_sum CoILBaseline distance_vehicle CoILAgent CorlNewTown CorlNewWeather CorlNewWeatherTown CorlTraining EccvGeneralization EccvTraining NocrashNewTown NocrashNewWeather NocrashNewWeatherTown NocrashTraining TestT1 TestT2 Augmenter get_episode_weather parse_remove_configuration CoILDataset PreSplittedSampler LogitSplittedSampler RandomSampler get_rank SubsetSampler BatchSequenceSampler forward_speed get_speed orientation_vector check_available_measurements soft hard_harder medium soft_harder high medium_harder remove_traffic_lights get_boost_pedestrian_vehicle_traffic_lights partition_keys_by_percentiles label_split remove_angle order_sequence parse_split_configuration remove_angle_traffic_lights split_lateral_noise_longitudinal_noise split_brake split_pedestrian_vehicle_traffic_lights_move split_speed_module_throttle convert_measurements split_left_central_right select_balancing_strategy split_sequence float_split split_speed_module get_inverse_freq_weights select_data_sequence split_pedestrian_vehicle_traffic_lights get_averaged_metrics erase_csv write_on_error_csv add_message add_image recover_loss_window close check_finish write_on_csv create_log write_stop add_scalar logger closeFileLogger readJSONlog streamlogger JSONFormatter filelogger get_episode_number get_status get_latest_checkpoint get_number_episodes_completed get_latest_checkpoint_drive get_error_summary get_summary get_latest_checkpoint_validation get_latest_output print_validation_summary print_train_summary plot_folder_summaries print_drive_summary print_folder_process_names Logger make_carla_settings ColorText game_loop HUD CameraManager FadingText KeyboardControl HelpText World PlayerMeasurements Listener FadingText LaneInvasionSensor World game_loop find_weather_presets get_actor_display_name Camera HUD CollisionSensor CameraManager ColorText KeyboardControl HelpText SensorCollector Measurements CoILModel branched_loss l1 l2 Loss l1_attention compute_branches_masks weight_decay_l1 weight_decay_l2 l1_loss normalize l2_loss adjust_learning_rate_auto adjust_learning_rate CoILICRA Branching Conv FC Join ResNet resnet50 Bottleneck resnet152 conv3x3 resnet34 resnet18 BasicBlock resnet101 read_control_csv get_speed_ground_truth get_camera_labels _read_data read_summary_csv get_ground_truth augment_steering _read_step_data _read_control_data read_summary_tasks_csv compute_steering_error compute_displacement compute_steering_error_filter_gt compute_id compute_steering_avg_l1 compute_km_per_infraction compute_control_accuracy compute_steering_accuracy compute_count_errors_weighted compute_steering_avg_mse_filter_gt compute_step compute_and_aggregate compute_correlation compute_steering_classification_error compute_count_cumulative_displacement compute_experiment compute_control_average_completion aggregate_metric compute_control_success_rate compute_relative_error_smoothed compute_steering_accuracy_filter_gt compute_steering_avg_l1_speed compute_steering_avg_mse compute_cumulative_displacement compute_displacement_steer compute_count_errors_weighted_speed filter_data read_data plot_scatter process_data compute_metric sldist split_episodes plot_test_image get_causes_of_end plot_on_map plot_point plot_episodes_tracks plot plot_analysis compute_lims TestLoadData download_file_from_google_drive save_response_content get_confirm_token download_file_from_google_drive save_response_content get_confirm_token download_file_from_google_drive save_response_content get_confirm_token download_file_from_google_drive save_response_content get_confirm_token Control figure_plot join_classes join_classes_for Control figure_plot join_classes join_classes_for get_average_over_interval draw_pt ScreenManager draw_vbar_on calc_lookahead_offset perspective_tform get_average_over_interval_stride draw_path calc_curvature draw_path_on generate_ncolors error type SensorDefinition reshape frombuffer raw_data to_bgra_array zeros list labels_to_array items dot astype float32 to_bgra_array depth_to_array ones log shape clip tan height depth_to_array reshape inv identity pi where delete dot width fov array connect client_type columns write max flush len sqrt array atan2 items list print zip zeros range len dot array linspace len split split File close sleep exists stat st_size join EXPERIMENT_BATCH_NAME EXPERIMENT_NAME sort_nicely listdir join EXPERIMENT_BATCH_NAME EXPERIMENT_NAME get_latest_evaluated_checkpoint exists print get_latest_evaluated_checkpoint index join EXPERIMENT_NAME EXPERIMENT_BATCH_NAME exists join EXPERIMENT_NAME EXPERIMENT_BATCH_NAME exists append append items list append dict_to_namevec heappush execvec_to_names dict_to_namevec join isdir print read_summary_csv listdir update join argmax isdir print read_summary_csv append listdir join listdir exp max set sort sub upper join list split fromarray save join mkdir join mkdir join isdir add set listdir join isdir add set listdir join remove listdir isdir update join remove print loadtxt listdir len print join remove listdir glob join sort_nicely SMTP sendmail MIMEText as_string quit update items list print zip zeros sum range len join print close write open join print close write open update items list print zip zeros sum range len start Process create_exp_path start Process create_exp_path update Process start create_exp_path update join get_gpu_resources execute_drive execute_train print execute_validation mount_experiment_heap heappop allocate_gpu_resources plot_folder_summaries sleep append listdir add_message communicate print find_free_port Popen add_message weathers start_carla_simulator write_data_point_control_summary plot_episodes_tracks str compute_average_std_separatetasks call PROCESS_NAME range get_latest_path CoILAgent run_driving_benchmark load join kill print build_experiments len USE_ORACLE experiment_suite_module add_message camelcase_to_snakecase FINISH_ON_VALIDATION_STALE open str len __import__ is_next_checkpoint_ready getpid getattr sleep write_header_control_summary get_latest_evaluated_checkpoint PROCESS_NAME range get_next_checkpoint mkdir unique validation_stale_point TEST_SCHEDULE join driving_benchmark print merge_with_yaml build_experiments set_type_of_process split data model zero_grad Loss AUGMENTATION save PRELOAD_MODEL_CHECKPOINT abs cuda CoILDataset recover_loss_window squeeze tolist Adam check_finish MODEL_CONFIGURATION CoILModel load_state_dict extract_branch PRELOAD_MODEL_BATCH append Augmenter adjust_learning_rate_auto select_balancing_strategy PRELOAD_MODEL_ALIAS stack MODEL_TYPE TRAIN_DATASET_NAME load time is_ready_to_save add_image write_on_error_csv criterion backward LOSS_FUNCTION parameters randint get_latest_saved_checkpoint step add_scalar BATCH_SIZE min write_on_csv max range write_on_csv range len forward_branch DataLoader eval write_regular_output write_waypoints_output write_stop split generate_name _merge_a_into_b update join merge_with_yaml immutable listdir join add_message LOG_SCALAR_WRITING_FREQUENCY EXPERIMENT_BATCH_NAME immutable LOG_IMAGE_WRITING_FREQUENCY EXPERIMENT_NAME create_log mkdir PROCESS_NAME deepcopy list items _decode_cfg_value _check_and_coerce_cfg_value_type literal_eval str list ndarray isinstance tuple eval type array print list keys OrderedDict SPLIT USE_FULL_ORACLE USE_NOISE_DATA parse_split_configuration print list keys OrderedDict print len array deg2rad dot array orientation_vector update join list glob keys float Sequential float Sequential float Sequential float Sequential Sequential float min float Sequential append sum len append add_message range len append SEQUENCE_STRIDE range NUMBER_IMAGES_SEQUENCE int list isinstance set select_data_sequence append range len order_sequence partition_keys_by_percentiles convert_measurements print range append len items list fromkeys array append keys convert_measurements convert_measurements convert_measurements convert_measurements convert_measurements convert_measurements convert_measurements print append len PreSplittedSampler list BATCH_SIZE parse_split_configuration get_inverse_freq_weights SPLIT NUMBER_IMAGES_SEQUENCE RandomSampler splitter_function measurements set DataLoader getattr intersection array range len items list print zip append zeros range len join filelogger Logger isfile join closeFileLogger info join join str join str join join remove str join str scalar_summary print transpose image_summary shape append get_cmap numpy range setFormatter getLogger addHandler JSONFormatter setLevel FileHandler getLogger close copy removeHandler flush StreamHandler append filterfunction loads read_summary_csv join read_summary_csv join range len range len join list sorted EXPERIMENT_BATCH_NAME set difference EXPERIMENT_NAME sort_nicely listdir PROCESS_NAME join readline print EXPERIMENT_BATCH_NAME loadtxt close EXPERIMENT_NAME split expand_dims PROCESS_NAME open readJSONlog join open print print join time get_episode_number print exists USE_ORACLE print_train_summary get_names print_drive_summary print_validation_summary str sorted list append range get_latest_path get_status listdir join print system immutable merge_with_yaml len join print merge_with_yaml immutable sort_nicely listdir EXPERIMENT_GENERATED_NAME set_rotation set_position CarlaSettings randomize_seeds add_sensor set Camera set_image_size output_folder init makedirs compile join get_world port run_step KeyboardControl host HWSURFACE tick render get_agent_sensor width DOUBLEBUF set_mode World height show_image_mini Client Clock mkdir get_attentions flip print latest_image get_forward_speed apply_control parse_events HUD set_timeout get_command compute_branches_masks update loss_function range sum requires_grad min add parameters float abs requires_grad min add parameters float sum append cuda cat append range len append abs range len LEARNING_RATE_DECAY_INTERVAL print param_groups LEARNING_RATE LEARNING_RATE_DECAY_LEVEL max LEARNING_RATE_THRESHOLD print param_groups count_steps_without_decrease_robust LEARNING_RATE LEARNING_RATE_DECAY_LEVEL max count_steps_without_decrease update ResNet load_url load_state_dict state_dict update ResNet load_url load_state_dict state_dict update ResNet load_url load_state_dict state_dict load_url ResNet load_state_dict load_url ResNet load_state_dict radians fabs min atan max update readline loadtxt close open expand_dims split update readline print loadtxt len close set open expand_dims range append split update readline loadtxt close open expand_dims split update join loadtxt update join loadtxt update join loadtxt update loadtxt get_ground_truth join read_control_csv update int join str sorted items OrderedDict range len items list aggregate_metric metric_func append percentile ones mean isscalar nan float sum array len append items float list print sum array abs array digitize float array astype digitize float array astype digitize float array astype abs array fabs set_printoptions append sum range len sum multiply isnan float abs len abs multiply absolute isnan sqrt float sum len append sum range len print join _read_control_data _read_data items list get_camera_labels print where OrderedDict getattr metric_func compute_and_aggregate list items filter_data compute_metric update join list items read_data plot plot_analysis print debug get_names strftime rmtree gmtime process_data makedirs astype plot_square range pop readline print loadtxt len close open range append split readline concatenate print loadtxt len close where open zeros split load join split_episodes str asarray plot_test_image print astype get_causes_of_end convert_to_pixel plot_on_map rescale mkdir open CarlaMap append makedirs min max std set_cmap subplots set_yticklabels reduce list sorted set_xlabel title scatter savefig hash legend sum plot set_xticklabels set_xlim ScalarMappable Normalize get_cmap keys items print rc locator_params set_ylabel to_rgba array set_ylim len set_cmap subplots set_yticklabels reduce list set_xlabel title scatter savefig sum set_xticklabels set_xlim compute_lims items print rc locator_params argsort set_ylabel array set_ylim get get_confirm_token save_response_content Session items list startswith print items copy range copy str list subplots set_fontsize plot get_yticklabels set_xlabel set_xlim get_xticklabels close set_ylabel savefig legend range set_ylim len int range hsv_to_rgb append randint float range append sum range len append sum range len perspective_tform draw_pt zip pi arcsin calc_curvature tan clip calc_lookahead_offset arange draw_path | COiLTRAiNE: Conditional Imitation Learning Training Framework ------------------------------------------------------------- This repository can be used to easily train and manage the trainings of imitation learning networks jointly with evaluations on the CARLA simulator. Objectives: * To enable the user to perform several trainings with a single command. * To automatically test the trained systems using CARLA. * Allow the user to monitor several trainings and testings on CARLA with a single glance. * Allows [to perform the testing methodology proposed](docs/on_offline_evaluation.md) | 2,081 |
fenchri/edge-oriented-graph | ['relation extraction'] | ['Connecting the Dots: Document-level Neural Relation Extraction with Edge-oriented Graphs'] | src/nnet/walks.py data_processing/utils.py data_processing/readers.py src/converter.py data_processing/gda2pubtator.py data_processing/bin2txt.py src/loader.py src/reader.py src/nnet/modules.py src/nnet/attention.py data_processing/filter_hypernyms.py src/utils.py src/nnet/init_net.py data_processing/reduce_embeds.py data_processing/process.py src/nnet/trainer.py src/dataset.py evaluation/evaluate.py data_processing/statistics.py src/nnet/network.py src/eog.py data_processing/split_gda.py data_processing/tools.py chunks main readPubTator load_pretrained_embeddings main chunks sentence_split_genia fix_sent_break convert2sent find_cross generate_pairs find_mentions tokenize_genia adjust_offsets to_edges replace2symbol to_graph using_split2 replace2space main prf concat_examples _concat_arrays_with_padding _concat_arrays to_device DocRelationDataset main train set_seed test ConfigLoader str2bool DataLoader read read_subdocs overlap_chunk chunks write_preds load_mappings save_model load_model write_errors Tee print_options solve plot_learning_curve humanized_time print_results observe setup_log BaseNet EmbedLayer Classifier Encoder EOG Trainer WalkLayer range len readPubTator print add_argument exit output_file ArgumentParser parse_args makedirs OrderedDict join makedirs OrderedDict format print len list map set OrderedDict in_data load_pretrained_embeddings keys full_embeds combinations pmid OrderedDict PairStruct type find_cross zip abs len EntStruct pmid off2 off1 name bio sent_no len type append kb_id sum enumerate split join replace list connected_components to_graph set OrderedDict intersection append kb_id keys chdir system join replace strip tag append enumerate arange values list len exit using_split2 append format copy set enumerate join int off2 off1 print issubset write split Graph add_nodes_from add_edges_from to_edges next iter append _len index split replace replace isinstance _concat_arrays to_device append range len asarray _concat_arrays_with_padding concatenate insert tuple maximum shape any full array range len manual_seed_all seed manual_seed set_seed save_model print plot_learning_curve Trainer DataLoader test_loader run __call__ setup_log train_loader load_mappings load_model print Trainer test_loader DataLoader __call__ setup_log eval_epoch ConfigLoader train load_config test isinstance print items list format items list list abs sort map maxsize len print print subplot list arange plot yticks xlabel map ylabel savefig figure legend len format print humanized_time append tabulate indent divmod stdout format Tee makedirs open format print min named_parameters numpy max print join save state_dict print join load load_state_dict print format | # Edge-oriented Graph Source code for the EMNLP 2019 paper "[Connecting the Dots: Document-level Relation Extraction with Edge-oriented Graphs](https://www.aclweb.org/anthology/D19-1498.pdf)". <p align="center"> <img src="./network.svg" height="190"> </p> ### Environment `$ pip3 install -r requirements.txt` The model was trained on Tesla K80 GPU, Ubuntu 16.04. Results are reproducible with a fixed seed. ### Reproducibility & Bug Fixes In the original code, there was a bug related to the word embedding layer. | 2,082 |
fenchri/walk-based-re | ['relation extraction'] | ['A Walk-based Model on Entity Graphs for Relation Extraction'] | src/analysis/perf_by_ent.py src/nnet/layers.py data_processing/utils.py data_processing/split.py src/converter.py src/analysis/AR_sigtest.py src/walk_re.py src/walk_re_wrapper.py data_processing/bin2txt.py src/loader.py src/reader.py src/nnet/modules.py src/analysis/find_best_epoch.py src/analysis/eval.py src/utils.py src/nnet/init_net.py data_processing/reduce_embeds.py data_processing/process.py src/nnet/trainer.py src/dataset.py data_processing/statistics.py src/nnet/network.py src/hyperparam_tuning.py data_processing/tools.py src/analysis/perf_by_dist.py main write2file read_brat load_pretrained_embeddings main chunks doc2sent fix_sent_break ent2sent adjust_offsets check_entities generate_pairs_types generate_pairs offsets2tokids check_relations sentence_split_genia replace2symbol tokenize_stanford using_split2 replace2space tokenize_genia sentence_split_stanford concat_examples _concat_arrays_with_padding _concat_arrays to_device RelationDataset rand_search boham ent_search bayes_opt main objective_function ConfigLoader DataLoader read_relation_input print_hyperparams write_bingo2file save_model load_model ordered_load write_errors2file Tee print_options plot_learning_curve write_pred2file observe humanized_time print_results setup_log main train set_seed test set_seed ModelWrapper prf eval_ align sig_test load_data main f1 prf evaluation f1 prf evaluation f1 prf evaluation BaseNet VectorAttentionLayer WalkLayer EmbedLayer Classifier Encoder WalkBasedModel Trainer items list format type1 check_entities type2 write dir generate_pairs_types cross lower generate_pairs type entid check_relations format glob print len tqdm OrderedDict output_file sum input_folder makedirs read_brat list print add_argument tqdm ArgumentParser parse_args keys int format print len OrderedDict run format len map set OrderedDict in_data load_pretrained_embeddings dim full_embeds range len combinations format arg1 write docid OrderedDict PairStruct arg2 type len permutations format arg1 print write docid OrderedDict PairStruct arg2 zip type len join replace join int format list off2 off1 write exit copy repr append values len join int format off2 off1 name print write repr using_split2 docid append enumerate format arange off2 off1 issubset write set enumerate combinations remove write set word_id combinations remove write OrderedDict print relid enumerate append _len index split replace replace annotate join append join replace strip annotate append enumerate chdir system join replace strip tag append enumerate isinstance _concat_arrays to_device append range len _concat_arrays_with_padding concatenate insert tuple maximum shape any full array range len OrderedDict train update_hyperparams num_iterations acquisition_func format maximizer print model_type bayesian_optimization n_init num_iterations format maximizer print model_type n_init entropy_search print num_iterations format random_search num_iterations acquisition_func format maximizer print bohamiann basicConfig rand_search boham ent_search ModelWrapper bayes_opt array args stdout join format Tee makedirs open DEFAULT_MAPPING_TAG add_constructor divmod print namedparams subplot list arange plot print yticks xlabel map ylabel savefig figure legend xticks len dropo reg format pos_dim out_dim print gc lr beta dropi type_dim print format att walks_iter print join save state_dict print join load load_state_dict format print humanized_time append tabulate indent print makedirs print makedirs print makedirs manual_seed_all seed manual_seed set_seed print plot_learning_curve Trainer DataLoader test_loader run __call__ setup_log eval_epoch train_loader load_model print Trainer DataLoader __call__ setup_log eval_epoch ConfigLoader train load_config test zip print list keys split list not_equal where sum equal values len list format R print tqdm OrderedDict randint abs keys range align systemB sig_test copy load_data zip systemA values split prf sum frozenset tqdm | # Walk-based Relation Extraction Source code for the ACL 2018 paper "[A Walk-based model on Entity Graphs for Relation Extraction](https://www.aclweb.org/anthology/P18-2014/)". ### Requirements & Environment ``` pip3 install -r requirements.txt ``` The original model of the paper was implement in [Chainer](https://chainer.org/). This is the same version of the model in [PyTorch](https://pytorch.org/). Slight differences might occur due to the above change. #### Reproducability Results are reproducible with the a fixed seed. | 2,083 |
fenghansen/KinD-pytorch | ['low light image enhancement'] | ['Kindling the Darkness: A Practical Low-light Image Enhancer'] | losses.py pytorch_ssim/__init__.py restore_trainer.py illum_trainer_custom.py utils/img_generator.py base_layers.py dataloader.py base_trainer.py illum_trainer.py base_parser.py models.py test_your_pictures.py evaluate.py restore_MSIA_trainer.py utils.py decom_trainer.py SpatialAttention ChannelAttention DoubleConv Concat MSIA MaxPooling2D Conv_BN_Relu ResConv Conv2D AvgPooling2D ConvTranspose2D CBAM BaseParser BaseTrainer LOLDataset_Decom CustomDataset LOLDataset build_LOLDataset_Decom_list_txt build_LOLDataset_list_txt change_name divide_dataset Decom_Trainer KinD_noDecom_Trainer Illum_Trainer Illum_Trainer Illum_Loss Restore_Loss feature_map_hook gradient gradient_no_abs Decom_Loss Illum_Custom_Loss KinD_noDecom IllumNet IllumNet_Custom KinD RestoreNet_MSIA RestoreNet_Unet DecomNet KinD_plus Restore_Trainer Restore_Trainer TestParser KinD_Player standard_illum no_grad load_weights sample data_augmentation log data_prefetcher create_window gaussian _ssim SSIM ssim sample join log join append listdir log enumerate join move listdir log split join rename append listdir log split sample shape append array cat min expand conv2d pad div permute to abs max min expand conv2d pad div permute to abs max print rot90 load update load_state_dict state_dict print clamp log expand mean conv2d pow device to is_tensor std clip list format imwrite tuple makedirs waitKey log imshow zeros range len Tensor Variable contiguous unsqueeze pow conv2d create_window size type_as get_device cuda is_cuda print | # KinD-pytorch This is a PyTorch implementation of KinD. The official KinD project(TensorFlow) is [KinD](https://github.com/zhangyhuaee/KinD). The KinD net was proposed in the following [Paper](http://doi.acm.org/10.1145/3343031.3350926). Kindling the Darkness: a Practical Low-light Image Enhancer. In ACM MM 2019<br> Yonghua Zhang, Jiawan Zhang, Xiaojie Guo **** ## Enviroment ## 1. Python = 3.6 2. PyTorch = 1.2.0 | 2,084 |
fengxin-bupt/Application-of-Word2vec-in-Phoneme-Recognition | ['speech recognition'] | ['Application of Word2vec in Phoneme Recognition'] | data_preprocess.py model_module.py train.py embedding_dataset_make.py test.py error_statistics.py config.py embedding_model.py model.py Hparams convert_txt2phn label_file feature_label_save audio_file phone2index audioToInputVector make_index_dict read_pickle main load_txt_path main Embedded Levenshtein_distance error_nums split_target_pred LER search_min Attention_sr_model attention_cell dense_to_sparse compute_ler rnn_cell encoder_lstm_layer reshape_pyramidal label_extract compute_loss feature_extract index2label load_pickle paths_pickle_load preds2phone extract_labels main inference load_pickle print Path enumerate load len close open append range split print list keys insert append File close File close load std hstack logfbank mean delta max label_file audio_file print audioToInputVector shuffle abspath append range len load close open convert_txt2phn feature_label_save phone2index make_index_dict load_txt_path makedirs reset_default_graph train Embedded set_random_seed append split range len str shape append range len error_nums zeros range search_min len list Levenshtein_distance print split_target_pred mean append keys range len cudnnlstm_layer transpose CudnnLSTM floordiv reshape floormod shape pad LSTMCell BahdanauAttention dtype sequence_mask not_equal concat where shape gather_nd cast fill bool equal segment_min get File close abspath array get File close abspath array append append File close value append seed paths_pickle_load time load_pickle build_graph index2label Attention_sr_model save_path print phn2ind_path infer preds2phone set_random_seed extract_labels dataset_root test_path reset_default_graph ind2phn_path inference | # Application-of-Word2vec-in-Phoneme-Recognition Build an attention-based model for speech recogntion (Listen Attend and Spell).Use the Word2vec model to help to train the attention model. Some code is quoted from https://github.com/thomasschmied/Speech_Recognition_with_Tensorflow. But there are some mistakes in his file, which i have changed these mistakes. Our model is built based on phoneme recognition. The datasets used are librispeech(http://www.openslr.org/12) and TIMIT.The pronunciation dictionary used is the 39 phoneme pronunciation dictionary of CMU. We used a new method in the model. The word2vec model is used to initialize the embedding matrix in the attention model.This can make the distance between the embedding vectors larger, so it can improve the performance of the model.At the same time, in order to solve the overfitting problem of the attention dataset on the attention model.We use a new phoneme inverse mapping strategy to convert more 39 phoneme datasets to 61 phoneme datasets. Dataset: Librispeech, TIMIT Feature: 40 mel-filterbank + delta + delta delta Encoder: 512BLSTM Decoder: 512LSTM Network frame work: | 2,085 |
fepegar/SlicerTorchIO | ['data augmentation'] | ['TorchIO: A Python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learning'] | TorchIOTransforms/TorchIOTransformsLib/HistogramStandardization.py TorchIOTransforms/TorchIOTransformsLib/RandomAnisotropy.py TorchIOTransforms/TorchIOTransformsLib/RandomAffine.py TorchIOTransforms/TorchIOTransformsLib/Transform.py TorchIOTransforms/TorchIOTransformsLib/RandomMotion.py TorchIOModule/TorchIOModule.py TorchIOTransforms/TorchIOTransformsLib/RandomGamma.py TorchIOTransforms/TorchIOTransformsLib/__init__.py TorchIOTransforms/TorchIOTransformsLib/RandomGhosting.py TorchIOTransforms/TorchIOTransformsLib/RandomBlur.py TorchIOTransforms/TorchIOTransformsLib/RandomBiasField.py TorchIOTransforms/TorchIOTransformsLib/RandomElasticDeformation.py Screenshots/Icon/script.py TorchIOTransforms/TorchIOTransformsLib/RandomSpike.py TorchIOTransforms/TorchIOTransforms.py TorchIOTransforms/TorchIOTransformsLib/CoordinatesWidget.py TorchIOModuleLogic TorchIOModule TorchIOTransformsLogic TorchIOTransformsWidget TorchIOTransforms TorchIOTransformsTest CoordinatesWidget HistogramStandardization RandomAffine RandomAnisotropy RandomBiasField RandomBlur RandomElasticDeformation RandomGamma RandomGhosting RandomMotion RandomSpike Transform | <!-- markdownlint-disable --> <p align="center"> <a href="http://torchio.rtfd.io/"> <img src="https://raw.githubusercontent.com/fepegar/torchio/master/docs/source/favicon_io/for_readme_2000x462.png" alt="TorchIO logo"> </a> </p> <!-- markdownlint-restore --> > *Tools like TorchIO are a symptom of the maturation of medical AI research using deep learning techniques*. Jack Clark, Policy Director at [OpenAI](https://openai.com/) ([link](https://jack-clark.net/2020/03/17/)). | 2,086 |
fepegar/torchio | ['data augmentation'] | ['TorchIO: A Python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learning'] | torchio/data/image.py torchio/external/due.py torchio/transforms/fourier.py tests/data/sampler/test_patch_sampler.py torchio/transforms/augmentation/spatial/random_flip.py torchio/cli/apply_transform.py tests/transforms/preprocessing/test_z_normalization.py tutorials/example_heteromodal.py torchio/transforms/augmentation/intensity/random_labels_to_image.py torchio/datasets/bite.py docs/examples/plot_history.py torchio/data/sampler/label.py torchio/data/sampler/sampler.py torchio/transforms/preprocessing/spatial/resize.py tests/transforms/preprocessing/test_keep_largest.py torchio/data/__init__.py torchio/transforms/preprocessing/intensity/z_normalization.py torchio/data/sampler/__init__.py torchio/visualization.py print_system.py torchio/datasets/fpg.py torchio/transforms/augmentation/__init__.py tests/transforms/label/test_sequential_labels.py torchio/data/inference/aggregator.py torchio/transforms/preprocessing/label/keep_largest_component.py torchio/transforms/preprocessing/spatial/crop_or_pad.py tests/data/sampler/test_label_sampler.py torchio/transforms/preprocessing/label/label_transform.py torchio/data/sampler/uniform.py torchio/transforms/augmentation/intensity/random_bias_field.py torchio/transforms/augmentation/spatial/random_anisotropy.py tests/transforms/augmentation/test_random_motion.py torchio/datasets/mni/__init__.py tests/transforms/augmentation/test_random_flip.py torchio/transforms/preprocessing/spatial/copy_affine.py torchio/data/io.py setup.py torchio/data/queue.py torchio/transforms/preprocessing/spatial/ensure_shape_multiple.py tests/data/sampler/test_weighted_sampler.py torchio/transforms/preprocessing/intensity/rescale.py torchio/transforms/preprocessing/label/one_hot.py tests/transforms/preprocessing/test_contour.py torchio/datasets/rsna_miccai.py torchio/datasets/ixi.py torchio/transforms/__init__.py tests/transforms/augmentation/test_random_affine.py torchio/transforms/augmentation/intensity/random_blur.py tests/transforms/preprocessing/test_ensure_shape_multiple.py torchio/transforms/augmentation/composition.py torchio/download.py torchio/datasets/__init__.py torchio/transforms/preprocessing/intensity/__init__.py tests/datasets/test_medmnist.py torchio/transforms/lambda_transform.py tests/data/test_queue.py torchio/data/subject.py torchio/constants.py torchio/data/inference/__init__.py torchio/datasets/itk_snap/__init__.py torchio/transforms/data_parser.py tests/data/test_subject.py tests/data/sampler/test_uniform_sampler.py torchio/transforms/interpolation.py tests/transforms/augmentation/test_random_gamma.py torchio/transforms/spatial_transform.py docs/source/conf.py torchio/transforms/transform.py tests/transforms/augmentation/test_random_elastic_deformation.py torchio/transforms/preprocessing/label/contour.py tests/data/inference/test_inference.py tests/data/test_io.py tests/transforms/label/test_remove_labels.py tests/transforms/preprocessing/test_resize.py torchio/datasets/mni/sheep.py tests/transforms/augmentation/test_random_labels_to_image.py torchio/datasets/visible_human.py tests/data/inference/test_aggregator.py torchio/transforms/preprocessing/spatial/to_canonical.py torchio/cli/print_info.py tests/transforms/preprocessing/test_one_hot.py tests/transforms/preprocessing/test_copy_affine.py tests/transforms/preprocessing/test_clamp.py torchio/transforms/augmentation/random_transform.py torchio/transforms/preprocessing/spatial/crop.py tests/transforms/augmentation/test_oneof.py torchio/datasets/slicer.py torchio/transforms/augmentation/intensity/random_swap.py tests/transforms/test_reproducibility.py torchio/__init__.py torchio/transforms/preprocessing/intensity/histogram_standardization.py torchio/transforms/intensity_transform.py torchio/data/dataset.py torchio/transforms/preprocessing/intensity/normalization_transform.py tests/transforms/augmentation/test_random_noise.py torchio/typing.py torchio/transforms/preprocessing/spatial/resample.py tests/__init__.py torchio/transforms/preprocessing/spatial/pad.py torchio/transforms/augmentation/intensity/random_spike.py torchio/transforms/augmentation/intensity/random_noise.py torchio/datasets/episurg.py tests/data/inference/test_grid_sampler.py tests/transforms/augmentation/test_random_spike.py torchio/datasets/medmnist.py tests/transforms/augmentation/test_random_anisotropy.py tests/transforms/augmentation/test_random_swap.py docs/examples/plot_video.py torchio/transforms/augmentation/intensity/random_motion.py torchio/transforms/augmentation/spatial/__init__.py torchio/transforms/preprocessing/intensity/mask.py tests/transforms/preprocessing/test_mask.py docs/examples/plot_3d_to_2d.py torchio/datasets/itk_snap/itk_snap.py torchio/transforms/preprocessing/label/sequential_labels.py torchio/transforms/augmentation/spatial/random_affine.py tests/utils.py torchio/datasets/mni/pediatric.py torchio/data/sampler/weighted.py tests/transforms/preprocessing/test_resample.py torchio/datasets/mni/colin.py torchio/transforms/preprocessing/__init__.py docs/examples/plot_colin27.py tests/transforms/test_lambda_transform.py tests/transforms/preprocessing/test_crop.py torchio/reference.py torchio/transforms/augmentation/intensity/__init__.py tests/transforms/augmentation/test_random_ghosting.py tests/data/sampler/test_random_sampler.py torchio/transforms/preprocessing/intensity/clamp.py tests/transforms/preprocessing/test_histogram_standardization.py tests/transforms/preprocessing/test_crop_pad.py torchio/datasets/mni/icbm.py docs/examples/plot_include_exclude.py tests/transforms/preprocessing/test_pad.py tests/transforms/preprocessing/test_rescale.py torchio/utils.py torchio/datasets/mni/mni.py tests/test_utils.py torchio/transforms/preprocessing/label/remove_labels.py tests/data/test_image.py tests/datasets/test_ixi.py torchio/transforms/augmentation/spatial/random_elastic_deformation.py tests/transforms/test_invertibility.py tests/transforms/augmentation/test_random_bias_field.py tests/transforms/label/test_remap_labels.py torchio/transforms/augmentation/intensity/random_gamma.py tests/data/test_subjects_dataset.py tests/transforms/test_collate.py torchio/transforms/preprocessing/spatial/bounds_transform.py torchio/transforms/preprocessing/label/remap_labels.py torchio/data/sampler/grid.py torchio/transforms/augmentation/intensity/random_ghosting.py tests/test_cli.py tests/transforms/test_transforms.py tests/transforms/augmentation/test_random_blur.py tests/transforms/preprocessing/test_to_canonical.py plot_batch plot_gif read_clip TestCLI TestUtils TorchioTestCase get_all_random_transforms TestImage TestIO test_write_nd_with_a_read_it_with_b TestNibabelToSimpleITK TestQueue TestSubject TestSubjectsDataset TestAggregator TestGridSampler TestInference model TestLabelSampler TestPatchSampler TestRandomSampler TestUniformSampler TestWeightedSampler TestIXI test_load_all TestCollate TestInvertibility TestLambda TestReproducibility TestTransforms TestTransform TestOneOf TestRandomAffine TestRandomAnisotropy TestRandomBiasField TestRandomBlur TestRandomElasticDeformation TestRandomFlip TestRandomGamma TestRandomGhosting TestRandomLabelsToImage TestRandomMotion TestRandomNoise TestRandomSpike TestRandomSwap TestRemapLabels TestRemoveLabels TestSequentialLabels TestClamp TestContour TestCopyAffine TestCrop TestCropOrPad TestEnsureShapeMultiple TestHistogramStandardization TestKeepLargestComponent TestMask TestOneHot TestPad TestResample TestRescaleIntensity TestResize TestToCanonical TestZNormalization _is_tarxz _is_zip extract_archive _is_tgz _is_tar download_url download_and_extract_archive check_integrity check_md5 gen_bar_updater calculate_md5 _is_targz _is_gzip get_major_sitk_version add_images_from_batch parse_spatial_shape create_dummy_dataset compress history_collate get_subjects_from_batch guess_external_viewer to_tuple check_sequence get_first_item get_torchio_cache_dir get_subclasses guess_type get_batch_images_and_size apply_transform_to_file get_stem make_gif plot_subject import_mpl_plt rotate plot_histogram get_num_bins plot_volume color_labels main get_params_dict_from_kwargs main SubjectsDataset LabelMap ScalarImage Image _read_itk_matrix _write_itk_matrix _to_itk_convention get_ras_affine_from_sitk check_uint_to_int read_image ensure_4d write_image _from_itk_convention _read_dicom sitk_to_nib read_shape _write_niftyreg_matrix _matrix_to_itk_transform _read_nibabel nib_to_sitk _read_niftyreg_matrix _write_nibabel write_matrix get_reader read_matrix get_sitk_metadata_from_ras_affine _write_sitk _read_sitk read_affine get_rotation_and_spacing_from_affine Queue Subject GridAggregator GridSampler LabelSampler PatchSampler RandomSampler UniformSampler WeightedSampler BITE BITE3 EPISURG FPG get_subject_id sglob IXITiny IXI SynapseMNIST3D NoduleMNIST3D FractureMNIST3D MedMNIST OrganMNIST3D VesselMNIST3D AdrenalMNIST3D RSNAMICCAI Slicer VisibleFemale VisibleMale VisibleHuman SubjectITKSNAP T1T2 AorticValve BrainTumor Colin27 ICBM2009CNonlinearSymmetric SubjectMNI Pediatric format_age Sheep InactiveDueCreditCollector _donothing_func DataParser FourierTransform IntensityTransform get_sitk_interpolator Interpolation Lambda SpatialTransform Transform OneOf Compose RandomTransform _parse_order BiasField RandomBiasField RandomBlur blur Blur RandomGamma Gamma power RandomGhosting _parse_restore Ghosting RandomLabelsToImage _check_mean_and_std_length _parse_used_labels _parse_label_key LabelsToImage RandomMotion Motion Noise get_noise RandomNoise Spike RandomSpike Swap insert swap crop get_random_indices_from_shape RandomSwap get_borders_mean Affine _parse_scales_isotropic RandomAffine _parse_default_value RandomAnisotropy _parse_max_displacement _parse_num_control_points ElasticDeformation RandomElasticDeformation _parse_axes RandomFlip _flip_image _ensure_axes_indices Flip Clamp _get_percentiles HistogramStandardization _get_average_mapping normalize _standardize_cutoff Mask mask NormalizationTransform RescaleIntensity ZNormalization Contour KeepLargestComponent LabelTransform OneHot RemapLabels RemoveLabels SequentialLabels BoundsTransform CopyAffine Crop CropOrPad EnsureShapeMultiple Pad Resample Resize ToCanonical main subplots suptitle squeeze tight_layout get_first_item imshow DataLoader flatten Queue zip __name__ n_frames seek transpose convert append array range open imshow subplots get_frame assertTensorEqual randn squeeze gettempdir save_function getattr mkdir Path eye load_function DataLoader class_ md5 tqdm join extract_archive basename download_url expanduser _is_tarxz _is_zip join remove _is_tar dirname _is_gzip join basename urlretrieve print expanduser makedirs tuple iter isinstance rand randint is_dir Path str list append range ones_like astype mkdir trange uint8 print rmtree Nifti1Image eye to_filename Subject transform print Subject save result_type split replace literal_eval with_suffix iter getattr update hasattr isinstance __subclasses__ append items list len add_transform_to_subject_history range Path append get_batch_images_and_size klass Subject zip get_first_image class_ add_image list format glob Path platform which is_file to_tuple rot90 fliplr percentile show spacing subplots set_title isinstance set_xlabel import_mpl_plt tight_layout rotate shape imshow set_ylabel savefig color_labels array invert_xaxis show items subplots set_title import_mpl_plt len tight_layout plot_volume savefig zip check_consistent_spatial_shape enumerate percentile min round max len pop show xlabel import_mpl_plt ylabel hist items list isinstance to_rgb import_mpl_plt shape append zeros list warn roll round permute save numpy range len transform_class get_params_dict_from_kwargs getattr manual_seed apply_transform_to_file guess_type split load show plot print class_ _read_sitk load str as_tensor affine get_fdata transpose check_uint_to_int str as_tensor is_dir _read_dicom ReadImage sitk_to_nib check_uint_to_int str ImageSeriesReader GetGDCMSeriesFileNames Path SetFileNames Execute str ReadImageInformation GetSize SetFileName GetNumberOfComponents tuple ImageFileReader GetDimension get_ras_affine_from_sitk get_reader SetFileName ReadImageInformation ImageFileReader str _write_sitk str asarray suffix Nifti1Pair set_intent Nifti1Image permute save str uint8 astype warn nib_to_sitk WriteImage Path _read_itk_matrix _read_niftyreg_matrix suffix Path _write_niftyreg_matrix _write_itk_matrix suffix Path dot inv dot inv str reshape _from_itk_convention hstack GetParameters vstack ReadTransform str WriteTransform _matrix_to_itk_transform AffineTransform tolist _to_itk_convention loadtxt inv inv savetxt sqrt sum get_sitk_metadata_from_ras_affine asarray tuple float64 transpose astype SetSpacing shape GetImageFromArray SetOrigin SetDirection ensure_4d transpose GetDimension get_ras_affine_from_sitk GetNumberOfComponentsPerPixel check_uint_to_int GetSpacing reshape GetDirection dot eye append GetOrigin array len float64 astype dot flatten get_rotation_and_spacing_from_affine shape as_tensor ndim permute __doc__ __doc__ __doc__ __doc__ __doc__ __doc__ PARTS int value as_tensor array gaussian_filter abs sign float check_sequence len insert crop clone array array uint16 tolist astype any item append array GetThreshold Execute reshape hstack OtsuThresholdImageFilter GetImageFromArray mean any GetArrayViewFromImage to_tuple enumerate enumerate isinstance to_tuple sorted any get_first_image check_consistent_orientation copy set_data flip as_tensor numpy array asarray max min nan_to_num mean dot len sorted list tolist set percentile digitize ones_like as_tensor _get_percentiles inf reshape astype float32 shape zeros _standardize_cutoff numpy array diff numpy info model UniformSampler SubjectsDataset DataLoader shape Queue Identity range Subject | <!-- markdownlint-disable --> <p align="center"> <a href="http://torchio.rtfd.io/"> <img src="https://raw.githubusercontent.com/fepegar/torchio/main/docs/source/favicon_io/for_readme_2000x462.png" alt="TorchIO logo"> </a> </p> <!-- markdownlint-restore --> > *Tools like TorchIO are a symptom of the maturation of medical AI research using deep learning techniques*. Jack Clark, Policy Director at [OpenAI](https://openai.com/) ([link](https://jack-clark.net/2020/03/17/)). | 2,087 |
feziodoshi/Neural-Artistic-Style-Video- | ['style transfer'] | ['A Neural Algorithm of Artistic Style', 'Preserving Color in Neural Artistic Style Transfer'] | read_frames_from_video.py NetworkFunction.py videoCapture.py styleFramesToVid_py.py Stylize read_frames_from_video COLOR_RGB2YCrCb function gradients variable warn flatten Input load_mask pooling_func transpose placeholder Model uniform fmin_l_bfgs_b preprocess_image_wo_path append sum range style_loss concatenate imresize Evaluator copy lower load_weights convert_all_kernels_in_model preprocess_image float get_file int time original_color_transform print deprocess_image dict loss cvtColor len VideoCapture read isOpened imwrite COLOR_BGR2RGB destroyAllWindows write Stylize imshow release cvtColor | # Neural-Artistic-Style-Video- Implementation of Neural Style Transfer on Video.  Contributers : [Fenil Doshi](https://github.com/feziodoshi), [Rohan Pooniwala](https://github.com/rohanpooniwala), [Rohit Saha](https://github.com/RohitSaha), [Raghav Gupta](https://github.com/raghavgupta0296) # Neural Artistic Style Implementation We mainly used the following research papers to implement Neural Artistic Style <ul> <li> Implementation of Neural Style Transfer from the paper <a href="http://arxiv.org/abs/1508.06576">A Neural Algorithm of Artistic Style</a> in Keras 1.2.0 <li> Color Preservation is based on the paper [Preserving Color in Neural Artistic Style Transfer](https://arxiv.org/abs/1606.05897). <li> Masked Style Transfer is based on the paper [Show, Divide and Neural: Weighted Style Transfer](http://cs231n.stanford.edu/reports2016/208_Report.pdf) | 2,088 |
fgnt/nn-gev | ['speech recognition'] | ['Neural network based spectral mask estimation for acoustic beamforming'] | nn_models.py train.py fgnt/chainer_extensions/links/sequence_batch_norm.py fgnt/chainer_extensions/links/sequence_lstms.py fgnt/chainer_extensions/sequenze_batch_normalization.py beamform.py fgnt/utils.py fgnt/chainer_extensions/sequence_linear.py fgnt/chainer_extensions/sequence_lstm.py setup.py fgnt/chainer_extensions/binary_cross_entropy.py fgnt/chainer_extensions/weight_init.py fgnt/chainer_extensions/links/sequence_linear.py fgnt/mask_estimation.py chime_data.py fgnt/signal_processing.py fgnt/beamforming.py gen_flist_simu prepare_training_data get_audio_data_with_context get_audio_data gen_flist_real SimpleFWMaskEstimator BLSTMMaskEstimator MaskEstimator _create_batch apply_beamforming_vector gev_wrapper_on_masks get_pca_vector phase_correction get_power_spectral_density_matrix get_gev_vector get_mvdr_vector condition_covariance blind_analytic_normalization blind_analytic_normalization_legacy simple_ideal_soft_mask estimate_IBM _voiced_unvoiced_split_characteristic quantile_mask _samples_to_stft_frames _stft_frames_to_samples stft audioread _biorthogonal_window_loopy audiowrite istft mkdir_p segment_axis Timer BinaryCrossEntropy binary_cross_entropy sequence_linear_function SequenceLinearFunction _as_mat _sigmoid sequence_lstm_function _extract_gates _grad_sigmoid SequenceLSTMFunction _grad_tanh _make_initial_state sequence_batch_normalization_function SequenceBatchNormalizationFunction normal uniform eye orthogonal SequenceBatchNorm SequenceLinear SequenceLSTM SequenceBLSTM list concatenate astype float32 append int list concatenate astype float32 append max gen_flist_simu list join transpose tqdm mkdir_p get_audio_data append abs estimate_IBM Variable to_gpu gpu ones shape sum einsum conj trace eye reshape shape eigh argmax array solve swapaxes expand_dims einsum conj shape empty range eigh dot shape sqrt zeros abs range conj sqrt abs einsum conj shape range copy apply_beamforming_vector dtype T phase_correction complex128 astype get_power_spectral_density_matrix get_gev_vector condition_covariance blind_analytic_normalization int arange ones cos pi concatenate cumsum squeeze min len real conjugate sum max range split cumsum min sum conj logical_and logical_or _voiced_unvoiced_split_characteristic conjugate power ones zeros kron range len load _samples_to_stft_frames _stft_frames_to_samples window ndim segment_axis shape pad ascii_lowercase argmax window len pad _biorthogonal_window_loopy zeros range enumerate int16 format print min astype copy wav_write start float sum max clip list copy shape empty swapaxes ravel makedirs reshape data get_array_module Variable _make_initial_state minimum normal svd permutation zeros range sqrt sum | # Neural network based GEV beamformer ## Introduction This repository contains code to replicate the results for the 3rd CHiME challenge using a *NN-GEV* Beamformer. ## Install This code requires Python 3 to run (although most parts should be compatible with Python 2.7). Install the necessary modules: ``` pip install chainer pip install tqdm pip install SciPy pip install scikit-learn | 2,089 |
fhtanaka/directed_research_CS_2018 | ['data augmentation'] | ['Data Augmentation Using GANs'] | decision_tree.py discriminator.py evaluate_dbs.py generator.py train_generator_discriminator.py data_treatment.py compare_data.py escalonate_data.py utils.py compare_data create_comparing_table ToTensor DataAtts DataSet create_and_evaluate_DT train_discriminator DiscriminatorNet train_generator GeneratorNet noise Architecture save_model fake_data_target random_noise real_data_target show list print tail xlabel ylabel fname title DataAtts hist savefig clf read_csv values drop argmax round str class_name len DT DataAtts append graph_from_dot_file export_graphviz mean predict_proba int print tail reshape read_csv fit Image print export_graphviz DT mean predict_proba graph_from_dot_file argmax create_png fit real_data_target backward fake_data_target size zero_grad discriminator step loss real_data_target backward size zero_grad discriminator step loss save Variable is_available randn Variable ones is_available Variable zeros is_available | # Directed Research CS 2018 This repository is for the subject "direct research in computer science B" from the university of Tsukuba. The goal on this project is to produce syntethic data to generate new datasets and balance unbalanced databases. More information can be found on the pdf. The database credicard is not included on this github because of its size, you can download it on https://www.kaggle.com/mlg-ulb/creditcardfraud/home. If you do, include it in the directory directed_research_CS_2018/original_data/ ## Notebooks: -*classification_random_tree*: Evaluate a database using decision tree. The training and test sets can be chosen in the notebook -*compare_original_synthetic*: Compare the original and a synthetic dataset by showing visual graphs -*create_fake_data*: Uses an already trained model to create fake data -*data_analysis*: Show graphics of the original databases -*fake_data_analysis*: Compare the original and a synthetic database by graphs | 2,090 |
filipradenovic/cnnimageretrieval-pytorch | ['image retrieval'] | ['Fine-tuning CNN Image Retrieval with No Human Annotation', 'CNN Image Retrieval Learns from BoW: Unsupervised Fine-Tuning with Hard Examples'] | cirtorch/networks/imageretrievalnet.py cirtorch/datasets/genericdataset.py cirtorch/datasets/testdataset.py cirtorch/examples/test.py cirtorch/datasets/datahelpers.py cirtorch/layers/loss.py cirtorch/examples/train.py cirtorch/examples/test_e2e.py cirtorch/utils/whiten.py cirtorch/utils/download_win.py cirtorch/utils/download.py cirtorch/layers/functional.py cirtorch/layers/pooling.py cirtorch/utils/general.py cirtorch/datasets/traindataset.py cirtorch/__init__.py cirtorch/utils/evaluate.py cirtorch/layers/normalization.py cirtorch/examples/example_descriptor_extraction.py imresize collate_tuples default_loader cid2filename accimage_loader flip pil_loader ImagesFromList ImagesFromDataList config_qimname configdataset config_imname TuplesDataset main main main validate AverageMeter test save_checkpoint set_batchnorm_eval main train l2n triplet_loss mac rmac spoc contrastive_loss roipool gem powerlaw ContrastiveLoss TripletLoss PowerLaw L2N RMAC MAC SPoC Rpool GeMmp GeM extract_ss ImageRetrievalNet extract_vectors init_network extract_ssl extract_regional_vectors extract_ms extract_local_vectors extract_ssr download_train download_test download_train download_test compute_ap compute_map compute_map_and_print get_data_root get_root htime sha256_hash pcawhitenlearn cholesky whitenlearn whitenapply thumbnail ANTIALIAS size view join lower format len extract_ss print reshape Compose system init_network load_url eval default_loader load_state_dict extract_ms numpy cuda htime compute_map_and_print extract_vectors gpu_id whitening save whitenapply list parse_args download_test whitenlearn get_data_root get format network_path network_offtheshelf Normalize item image_size multiscale load download_train time join T meta_repr dot argsort configdataset split network pool TuplesDataset SGD pretrained DataLoader save_checkpoint arch seed test_datasets exp Adam epochs append regional range manual_seed_all val directory test lr resume manual_seed isfile local_whitening ExponentialLR training_dataset min train step makedirs update time format criterion backward print squeeze AverageMeter zero_grad apply create_epoch_tuples item step cuda range enumerate len update time format criterion print squeeze AverageMeter eval create_epoch_tuples item cuda range enumerate len htime compute_map_and_print extract_vectors whitenapply cuda whitenlearn get_data_root format Compose eval Normalize join time T print dot argsort configdataset test_whiten numpy split copyfile join save eval __name__ list max_pool2d size min tolist expand_as floor item Tensor abs max range int size min tolist unsqueeze floor item append Tensor abs max range eps clamp size pow sqrt permute sum size pow permute item sum get children list format basename join get_data_root print ImageRetrievalNet load_url load_state_dict startswith ReLU append Rpool Linear eval cuda ImagesFromList DataLoader pow zeros clone interpolate eval cuda ImagesFromList DataLoader eval cuda ImagesFromList DataLoader join format print len system mkdir range makedirs join format print len system mkdir range makedirs float arange len compute_ap max arange min zeros float sum array len format compute_map concatenate print around append range len round sha256 dot norm T eig inv mean dot sqrt diag T inv eig mean dot cholesky eye | ## CNN Image Retrieval in PyTorch: Training and evaluating CNNs for Image Retrieval in PyTorch This is a Python toolbox that implements the training and testing of the approach described in our papers: **Fine-tuning CNN Image Retrieval with No Human Annotation**, Radenović F., Tolias G., Chum O., TPAMI 2018 [[arXiv](https://arxiv.org/abs/1711.02512)] **CNN Image Retrieval Learns from BoW: Unsupervised Fine-Tuning with Hard Examples**, Radenović F., Tolias G., Chum O., ECCV 2016 [[arXiv](http://arxiv.org/abs/1604.02426)] <img src="http://cmp.felk.cvut.cz/cnnimageretrieval/img/cnnimageretrieval_network_medium.png" width=\textwidth/> --- | 2,091 |
fillantrophy02/FGAN_aa | ['anomaly detection'] | ['Fence GAN: Towards Better Anomaly Detection'] | untitled2.py utils/data.py utils/visualize.py 2D_experiment/custom_losses.py utils/cropping.py cropping.py utils/custom_losses.py utils/model.py 2D_experiment/2D_fgan.py main.py fgan_train.py untitled.py crop image_crop noise_data pretrain training_pipeline D_data set_trainability train crop noise_data animate real_data get_generative pretrain D_loss get_discriminative data_G data_D set_trainability train make_gan com crop image_crop com_conv get_cifar10 preprocess load_data get_mnist load_model D_loss get_mnist_model set_trainability get_cifar10_model compute_au show_images histogram deprocess len range extend split range crop extend layers noise_data list ones image_crop crop_size batch_load sample zeros range predict format batch_size print latent_dim range batch_size compute_au ano_class save dataset round open noise_data OrderedDict v_freq range predict format close latent_dim listdir show_images print makedirs write histogram evaluation epochs len seed load_model pretrain set_random_seed load_data train print str T int xlabel reshape close ylabel colorbar ylim scatter contourf figure linspace title meshgrid savefig xlim predict noise_data real_data ones zeros predict noise_data zeros binary_crossentropy Model Input Model Input compile G D Model set_trainability Input compile train_on_batch set_value data_D set_trainability append sum len train_on_batch set_value animate data_G data_D set_trainability append sum float64 clip concatenate print reshape delete preprocess load_data append enumerate int asarray concatenate choice preprocess load_data array G Adam D Model set_trainability Input compile G random_normal_initializer variable Adam D Model set_trainability Input compile uint8 clip subplot format squeeze len close tight_layout imshow savefig figure deprocess xticks range yticks squeeze roc_curve precision_recall_curve predict auc normal format xlabel squeeze close ylabel title hist savefig figure legend predict | # Fence GAN: Towards Better Anomaly Detection This is the official implementation of the paper: Fence GAN: Towards Better Anomaly Detection [(link)](https://arxiv.org/abs/1904.01209). ## Prerequisites 1. Linux OS 2. Python 3 3. CUDA ## Installation 1. Clone repository ``` git clone https://github.com/phuccuongngo99/Fence_GAN.git | 2,092 |
fillassuncao/denser-models | ['data augmentation'] | ['DENSER: Deep Evolutionary Network Structured Representation'] | load_data.py denser_models.py fashion_mnist.py augmentation load_data test load_data load_mnist_rotated load_rectangles_images load_mat_mnist_rotated_background load_mat_mnist_background load_mat load_mat_mnist_rotated load_mnist_background load_mnist_rotated_background load_rectangles load_svhn zeros randint load_mnist_rotated load_rectangles_images reshape to_categorical astype load_mnist_background load_mnist_rotated_background load_rectangles load_svhn load_model print average load_data append accuracy_score argmax array range predict urlretrieve int list map append float split int list map append float split int list map append float split print shape loadmat rollaxis load_mat_mnist_rotated load_mat_mnist_background load_mat_mnist_rotated_background load_mat load_mat_mnist_background load_mat_mnist_background | # DENSER: Deep Evolutionary Network Structured Representation Deep Evolutionary Network Structured Representation (DENSER) is a novel approach to automatically design Artificial Neural Networks (ANNs) using Evolutionary Computation. The algorithm not only searches for the best network topology, but also tunes hyper-parameters (e.g., learning or data augmentation parameters). The automatic design is achieved using a representation with two distinct levels, where the outer level encodes the general structure of the network, and the inner level encodes the parameters associated with each layer. The allowed layers and hyper-parameter value ranges are defined by means of a human-readable Context-Free Grammar. If you use this code, a reference to one of the following works would be greatly appreciated: ``` @inproceedings{assuncao2018evolving, title={Evolving the Topology of Large Scale Deep Neural Networks}, author={Assun{\c{c}}ao, Filipe and Louren{\c{c}}o, Nuno and Machado, Penousal and Ribeiro, Bernardete}, booktitle={European Conference on Genetic Programming (EuroGP)}, year={2018}, organization={Springer} } | 2,093 |
findoctor/YOPO_reproduce | ['adversarial defense'] | ['You Only Propagate Once: Accelerating Adversarial Training via Maximal Principle'] | MNIST/train.py CIFAR/attack.py yopo_trade/pre_res18.py yopo_trade/train.py CIFAR/plot.py yopo_trade/evaluate.py yopo_trade/attack.py MNIST/helper.py yopo_trade/config.py CIFAR/dataset.py CIFAR/train_one_epoch.py MNIST/train_pgd.py MNIST/attack.py CIFAR/network.py MNIST/network.py CIFAR/helper.py yopo_trade/main.py CIFAR/train.py CIFAR/train_pgd.py MNIST/train_one_epoch.py MNIST/plot.py yopo_trade/loss.py MNIST/dataset.py IPGD cw_l2_attack pgd train_step create_test_dataset create_train_dataset CrossEntropyLossWeighted Hamiltonian get_acc PreActBlock PreActResNet PreActResNet18 GlobalpoolFC create_network PreActBottleneck PreActResNet34 visualize eval_one_epoch visualize Yopo_train visualize2 visualize IPGD cw_l2_attack pgd train_step create_test_dataset create_train_dataset CrossEntropyLossWeighted Hamiltonian get_acc create_network SmallCNN visualize eval_one_epoch visualize Yopo_train visualize2 visualize clip_eta attacking training_setting eval CrossEntropyWithWeightPenlty cal_l2_norm Hamiltonian_Loss main PreActBlock PreActResNet PreActResNet18 GlobalpoolFC PreActBottleneck PreActResNet34 torch_accuracy inner_loop train outer_loop backward model clamp zero_ range str time criterion backward print train pgd zero_grad get_acc item append to step rand_like net enumerate backward print f zero_grad Adam step to sum range DataLoader Compose CIFAR10 DataLoader Compose CIFAR10 data sum max show plot xlabel add_subplot ylabel figure legend eval attack get_acc net show plot add_subplot figure legend rand_like MNIST MNIST criterion model clamp sign requires_grad_ eval clip_eta uniform_ to range str model print torch_accuracy attacking item to enumerate norm named_parameters PreActResNet18 Compose Hamiltonian_Loss layer_one SGD train MultiStepLR parameters DataLoader CIFAR10 to training_setting CrossEntropyLoss topk size t eq mul_ expand_as append sum max criterion backward clamp sign range inner_loop model item range detach KLDivLoss model zero_grad trades_criterion outer_loop str to range detach log_softmax requires_grad_ eval softmax item enumerate time criterion backward print clamp step | # YOPO_reproduce
Reproduce works of ["You Only Propagate Once: Accelerating Adversarial Training via Maximal Principle" ](https://arxiv.org/abs/1905.00877).
## Prerequisites
- PyTorch 1.0.1
- Python 3
- easydict
## How to run the code
###### PGD training:
| 2,094 |
fineline179/MEHTA_project | ['speech recognition'] | ['An exact mapping between the Variational Renormalization Group and Deep Learning'] | codeRBM/RBM_einsum.py codeRBM/RBM_torch.py codeRBM/RBM_testing.py codeIsing/setup.py codeRBM/RBM_testing_torch.py codeRBM/RBM.py RBM RBM RBM | # MEHTA_project_ Reconstruction of 1410.3831 - An exact mapping between the Variational Renormalization Group and Deep Learning. https://arxiv.org/abs/1410.3831 `codeIsing/` - Code to generate samples of Ising model via Metropolis monte carlo from: https://github.com/pablodecm/ising_model `codeRBM/` - Code for restricted Boltzmann machine. Both NumPy and PyTorch versions. | 2,095 |
fineline179/MEHTA_project_2 | ['speech recognition'] | ['An exact mapping between the Variational Renormalization Group and Deep Learning'] | codeRBM/RBM_einsum.py codeRBM/RBM_torch.py codeRBM/RBM_testing.py codeIsing/setup.py codeRBM/RBM_testing_torch.py codeRBM/RBM.py RBM RBM RBM | # MEHTA_project_ Reconstruction of 1410.3831 - An exact mapping between the Variational Renormalization Group and Deep Learning. https://arxiv.org/abs/1410.3831 `codeIsing/` - Code to generate samples of Ising model via Metropolis monte carlo from: https://github.com/pablodecm/ising_model `codeRBM/` - Code for restricted Boltzmann machine. Both NumPy and PyTorch versions. | 2,096 |
firmenich/MultiTask-GNN | ['data augmentation'] | ['Multitask Learning On Graph Neural Networks Applied To Molecular Property Predictions'] | torch_geometric/datasets/tosca.py torch_geometric/transforms/line_graph.py torch_geometric/datasets/dbp15k.py torch_geometric/datasets/dynamic_faust.py torch_geometric/nn/dense/dense_gcn_conv.py torch_geometric/nn/pool/max_pool.py torch_geometric/nn/glob/sort.py torch_geometric/read/ply.py torch_geometric/nn/glob/__init__.py torch_geometric/nn/dense/__init__.py torch_geometric/transforms/knn_graph.py torch_geometric/nn/unpool/knn_interpolate.py torch_geometric/data/makedirs.py torch_geometric/nn/inits.py torch_geometric/nn/conv/sage_conv.py torch_geometric/transforms/two_hop.py torch_geometric/read/sdf.py torch_geometric/utils/num_nodes.py torch_geometric/transforms/center.py torch_geometric/utils/undirected.py torch_geometric/nn/models/__init__.py torch_geometric/datasets/ppi.py torch_geometric/nn/conv/gat_conv.py torch_geometric/nn/models/re_net.py torch_geometric/data/dataset.py models/gain.py torch_geometric/nn/conv/gmm_conv.py torch_geometric/transforms/random_shear.py torch_geometric/transforms/local_cartesian.py torch_geometric/transforms/normalize_features.py torch_geometric/nn/reshape.py torch_geometric/transforms/radius_graph.py torch_geometric/utils/convert.py torch_geometric/utils/softmax.py torch_geometric/nn/pool/pool.py torch_geometric/datasets/cora_full.py torch_geometric/nn/conv/spline_conv.py torch_geometric/read/txt_array.py torch_geometric/datasets/amazon.py torch_geometric/datasets/mnist_superpixels.py torch_geometric/data/__init__.py torch_geometric/utils/dropout.py torch_geometric/transforms/random_translate.py torch_geometric/transforms/__init__.py torch_geometric/utils/degree.py torch_geometric/transforms/generate_mesh_normals.py torch_geometric/transforms/normalize_rotation.py torch_geometric/read/planetoid.py torch_geometric/transforms/to_dense.py torch_geometric/utils/subgraph.py torch_geometric/nn/conv/nn_conv.py torch_geometric/read/tu.py torch_geometric/transforms/random_scale.py torch_geometric/nn/pool/avg_pool.py torch_geometric/transforms/delaunay.py torch_geometric/utils/geodesic.py torch_geometric/nn/conv/feast_conv.py torch_geometric/nn/pool/sag_pool.py torch_geometric/data/data.py utils/features.py torch_geometric/nn/conv/cheb_conv.py torch_geometric/datasets/reddit.py torch_geometric/nn/conv/gcn_conv.py torch_geometric/transforms/target_indegree.py torch_geometric/datasets/qm7.py torch_geometric/transforms/random_rotate.py models/__init__.py torch_geometric/nn/pool/voxel_grid.py torch_geometric/utils/normalized_cut.py torch_geometric/datasets/ged_dataset.py torch_geometric/utils/metric.py torch_geometric/nn/pool/graclus.py torch_geometric/nn/__init__.py utils/__init__.py torch_geometric/datasets/pcpnet_dataset.py torch_geometric/transforms/compose.py torch_geometric/nn/glob/attention.py main.py torch_geometric/debug.py torch_geometric/utils/sort_edge_index.py torch_geometric/nn/models/deep_graph_infomax.py torch_geometric/transforms/laplacian_lambda_max.py torch_geometric/transforms/remove_isolated_nodes.py torch_geometric/data/batch.py torch_geometric/utils/isolated.py torch_geometric/nn/dense/dense_gin_conv.py torch_geometric/transforms/face_to_edge.py torch_geometric/nn/conv/edge_conv.py torch_geometric/datasets/karate.py torch_geometric/nn/conv/message_passing.py torch_geometric/visualization/influence.py torch_geometric/nn/conv/sg_conv.py torch_geometric/datasets/geometry.py torch_geometric/transforms/add_self_loops.py torch_geometric/data/dataloader.py torch_geometric/data/extract.py torch_geometric/utils/loop.py torch_geometric/datasets/planetoid.py torch_geometric/datasets/s3dis.py torch_geometric/transforms/sample_points.py torch_geometric/datasets/tu_dataset.py torch_geometric/utils/repeat.py torch_geometric/nn/glob/glob.py torch_geometric/utils/scatter.py torch_geometric/nn/conv/signed_conv.py torch_geometric/transforms/local_degree_profile.py torch_geometric/transforms/one_hot_degree.py torch_geometric/transforms/normalize_scale.py torch_geometric/nn/conv/ppf_conv.py torch_geometric/nn/conv/gated_graph_conv.py torch_geometric/nn/pool/topk_pool.py torch_geometric/utils/grid.py torch_geometric/nn/pool/__init__.py torch_geometric/nn/data_parallel.py torch_geometric/transforms/distance.py torch_geometric/datasets/coma.py torch_geometric/data/download.py torch_geometric/transforms/cartesian.py torch_geometric/datasets/shrec2016.py torch_geometric/read/off.py torch_geometric/datasets/bitcoin_otc.py torch_geometric/nn/conv/point_conv.py torch_geometric/nn/pool/edge_pool.py torch_geometric/nn/glob/set2set.py torch_geometric/transforms/linear_transformation.py models/ggrnet.py torch_geometric/datasets/willow_object_class.py torch_geometric/read/obj.py torch_geometric/datasets/icews.py torch_geometric/nn/conv/graph_conv.py torch_geometric/datasets/qm9.py torch_geometric/nn/pool/consecutive.py torch_geometric/nn/conv/x_conv.py torch_geometric/nn/dense/diff_pool.py torch_geometric/transforms/polar.py torch_geometric/nn/conv/hypergraph_conv.py torch_geometric/utils/to_dense_batch.py models/gin.py torch_geometric/transforms/point_pair_features.py torch_geometric/transforms/fixed_points.py torch_geometric/read/__init__.py torch_geometric/utils/random.py torch_geometric/nn/models/graph_unet.py torch_geometric/nn/unpool/__init__.py torch_geometric/transforms/spherical.py torch_geometric/datasets/modelnet.py torch_geometric/visualization/__init__.py torch_geometric/nn/meta.py torch_geometric/nn/models/autoencoder.py utils/validation.py torch_geometric/utils/to_dense_adj.py torch_geometric/nn/conv/appnp.py torch_geometric/utils/__init__.py torch_geometric/nn/conv/tag_conv.py torch_geometric/nn/conv/rgcn_conv.py torch_geometric/nn/conv/gin_conv.py torch_geometric/transforms/constant.py torch_geometric/utils/get_laplacian.py torch_geometric/nn/models/node2vec.py torch_geometric/__init__.py torch_geometric/data/sampler.py torch_geometric/datasets/entities.py torch_geometric/transforms/random_flip.py torch_geometric/datasets/coauthor.py torch_geometric/datasets/gdelt.py torch_geometric/data/in_memory_dataset.py torch_geometric/utils/sparse.py torch_geometric/nn/conv/__init__.py torch_geometric/datasets/shapenet.py torch_geometric/nn/models/signed_gcn.py torch_geometric/nn/conv/agnn_conv.py torch_geometric/nn/conv/arma_conv.py torch_geometric/nn/models/jumping_knowledge.py torch_geometric/read/npz.py torch_geometric/nn/dense/dense_sage_conv.py torch_geometric/datasets/faust.py torch_geometric/nn/conv/dna_conv.py torch_geometric/datasets/__init__.py prepare_dataset GINAttNet GGRNet GINet set_debug_enabled is_debug_enabled debug set_debug Batch Data size_repr DataListLoader DataLoader DenseDataLoader to_list Dataset files_exist download_url extract_zip extract_tar extract_bz2 extract_gz maybe_log InMemoryDataset makedirs NeighborSampler Block DataFlow Amazon BitcoinOTC Coauthor CoMA CoraFull DBP15K DynamicFAUST Entities FAUST GDELT GEDDataset GeometricShapes EventDataset ICEWS18 KarateClub MNISTSuperpixels ModelNet PCPNetDataset Planetoid PPI QM7b QM9 Reddit S3DIS ShapeNet SHREC2016 TOSCA TUDataset WILLOWObjectClass DataParallel normal ones kaiming_uniform uniform reset glorot zeros MetaLayer Reshape AGNNConv APPNP ARMAConv ChebConv DNAConv restricted_softmax MultiHead Attention Linear DynamicEdgeConv EdgeConv FeaStConv GatedGraphConv GATConv GCNConv GINConv GMMConv GraphConv HypergraphConv MessagePassing NNConv PointConv point_pair_features PPFConv RGCNConv SAGEConv SGConv SignedConv SplineConv TAGConv XConv DenseGCNConv DenseGINConv DenseSAGEConv dense_diff_pool GlobalAttention global_mean_pool global_add_pool global_max_pool Set2Set global_sort_pool ARGA negative_sampling GAE ARGVA VGAE InnerProductDecoder DeepGraphInfomax GraphUNet JumpingKnowledge Node2Vec RENet SignedGCN negative_triangle_sampling avg_pool _avg_pool_x avg_pool_x consecutive_cluster EdgePooling graclus max_pool_x _max_pool_x max_pool pool_batch pool_pos pool_edge SAGPooling topk filter_adj TopKPooling voxel_grid knn_interpolate read_npz parse_npz read_obj yield_file face_to_tri parse_off read_off read_planetoid_data sample_mask read_file edge_index_from_dict read_ply read_sdf parse_sdf cat read_file read_tu_data split parse_txt_array read_txt_array NormalizeFeatures AddSelfLoops Cartesian Center Compose Constant Delaunay Distance FaceToEdge FixedPoints GenerateMeshNormals KNNGraph LaplacianLambdaMax LinearTransformation LineGraph LocalCartesian LocalDegreeProfile NormalizeRotation NormalizeScale OneHotDegree PointPairFeatures Polar RadiusGraph RandomFlip RandomRotate RandomScale RandomShear RandomTranslate RemoveIsolatedNodes SamplePoints Spherical TargetIndegree ToDense TwoHop from_networkx from_scipy_sparse_matrix to_networkx to_scipy_sparse_matrix degree dropout_adj filter_adj geodesic_distance get_laplacian grid grid_pos grid_index remove_isolated_nodes contains_isolated_nodes add_self_loops segregate_self_loops contains_self_loops remove_self_loops add_remaining_self_loops true_positive false_negative accuracy false_positive precision mean_iou f1_score recall true_negative intersection_and_union normalized_cut maybe_num_nodes erdos_renyi_graph stochastic_blockmodel_graph barabasi_albert_graph repeat scatter_ softmax sort_edge_index dense_to_sparse subgraph to_dense_adj to_dense_batch is_undirected to_undirected influence one_of_k_encoding_unk bond_features atom_features maybe_num_nodes one_of_k_encoding safe_index get_bond_pair mol2vec val myloss EarlyStopping prepare_data test isnan clean_print CrossValidation train nonmissingvales overall_clean_print getcwd DataFrame read_csv merge is_tensor join print urlopen exists makedirs print maybe_log maybe_log maybe_log maybe_log expanduser normpath uniform_ sqrt uniform_ sqrt uniform_ sqrt size fill_ fill_ normal_ children _reset clamp exp sum dtype norm size transpose matmul numel mean softmax to new_full arange view sort size contiguous to_dense_batch item cat list uint8 view size astype from_numpy device sample tensor to range uint8 view astype from_numpy device randint to consecutive_cluster _avg_pool_x pool_batch _avg_pool_x Batch edge_index consecutive_cluster pool_edge transform edge_attr x size arange unique scatter_ namedtuple consecutive_cluster _max_pool_x pool_batch Batch edge_index consecutive_cluster pool_edge transform edge_attr _max_pool_x x size remove_self_loops view coalesce new_full arange view sort size clamp new_ones to long scatter_add cat new_full size maybe_num_nodes arange tensor size cat scatter_add todense tocoo to to_undirected size remove_self_loops tensor float long read close startswith open split Data contiguous yield_file append tensor Data face_to_tri parse_txt_array tensor to int64 Data arange size sample_mask edge_index_from_dict item zeros cat join lower format Tensor items list coalesce stack remove_self_loops zeros stack Data Data parse_txt_array one_hot coalesce t stack tensor cat join Data format unbind coalesce glob float t unsqueeze read_file remove_self_loops unique to long cat split arange cumsum edge_index tolist from_numpy bincount cat squeeze ones size maybe_num_nodes coo_matrix cpu data tocoo from_numpy stack to long update add_edge list DiGraph tolist nodes add_nodes_from num_nodes range enumerate items list from_dict number_of_nodes contiguous nodes edges append tensor zeros maybe_num_nodes new_full uint8 coalesce filter_adj maybe_num_nodes stack to cat dtype toarray arange view cross compute_gdist item append to numpy range cat len add_self_loops ones size pow maybe_num_nodes scatter_add cat grid_pos grid_index arange view coalesce size stack repeat tensor repeat arange view maybe_num_nodes remove_self_loops segregate_self_loops arange zeros_like size maybe_num_nodes full_like zeros sum full cat new_full arange maybe_num_nodes repeat cat arange maybe_num_nodes repeat full cat append sum range append sum range append sum range append sum range float to float to recall precision sum scatter_add mean float to intersection_and_union dtype degree combinations arange view to_undirected t stack uint8 arange view to_undirected size t stack tensor to cat to_undirected choice stack remove_self_loops numpy range cat Number isinstance format getattr op isinstance maybe_num_nodes exp argsort maybe_num_nodes contiguous arange size maybe_num_nodes item tensor zeros view size scatter_ new_zeros new_ones item zeros cat new_full arange view size scatter_ new_zeros new_ones item zeros cat maybe_num_nodes to_undirected coalesce maybe_num_nodes stack coalesce size requires_grad_ append sum range GetIdx one_of_k_encoding_unk GetRingInfo GetSubstructMatches GetSymbol sum str GetStereo one_of_k_encoding_unk GetBondType GetBonds Data bond_features GetAtoms append GetBonds tensor get_bond_pair MolFromSmiles dict append tensor range enumerate len criterion MSELoss dict sqrt isnan append tensor range len print range len print range len DataLoader ReduceLROnPlateau clean_print save device early_stopping SubsetRandomSampler Adam append to range KFold state_dict val test mean item print EarlyStopping prepare_data dict parameters split train step std early_stop len model backward zero_grad device to step model myloss float dict isnan eval device append to range len model dict eval device to range len | # Multitask Learning On Graph Neural Networks In this repository we provide the code that is necessary to reproduce the results from the paper [Multitask Learning On Graph Neural Networks Applied To Molecular Property Predictions](https://arxiv.org/pdf/1910.13124.pdf). # Dependencies <img width="40%" src="https://raw.githubusercontent.com/rusty1s/pytorch_geometric/master/docs/source/_static/img/pyg_logo_text.svg?sanitize=true" /> The list of libraries that you need to install to execute the code: - torchvision=0.3.0 - networkx=2.3 - rdflib=4.2.2 - torch_spline_conv=1.1.0 - plyfile=0.7 | 2,097 |
firstdayofjune/aire-20 | ['word embeddings'] | ["Topic Modeling on User Stories using Word Mover's Distance"] | python/mare/statistics.py python/mare/analysis.py python/mare/plotting.py python/mare/requirements.py SelfTrainedWord2VecAnalyzer RequirementsPreprocessor Word2VecAnalyzer RequirementsAnalyzer PreTrainedWord2VecAnalyzer LDAAnalyzer NLPAnalyzer RequirementsPlotter Plotter CrowdREReader Requirement embeddings_for_tokens items list defaultdict print set filter requirements lexical_words len | # Topic Modeling on the Crowd RE Dataset ## Build To build the pdf run `latexmk main` ## Contribute Content should be grouped into separate files and added to the *content* directory appropriately. The individual files can then be included in the main.tex using the `\input{contet/...}` directive. Any acronyms introduced to the document, may be added to the *misc/acronyms.tex* in alphabetical order. Bibliographic references go to the *references.bib* and can further be referenced using the `\cite{...}` or the `\cite[p123]{...}` command. ## LaTeX Tools When using the LaTeXTools Plugin for Sublime Text, make sure to configure your *.sublime-project* as follows: ```json | 2,098 |
fisherd3/fdm | ['stochastic optimization'] | ['Topic Modeling via Full Dependence Mixtures'] | assignments_calculation.py FDM.py synthetic_topic_test.py graph_communities_test.py fast_moments_build.py newsgroup_test.py _get_empirical_dist calculate_assignments calculate_assignments_sparse _csr_vappend FDM ent fit_transform CountVectorizer non_negative_factorization hstack time format csr_matrix print calculate_assignments range _csr_vappend len | # FDM This code is an implementation of the Full Dependence Mixture Topic Model (FDM) model described in: Topic Modeling via Full Dependence Mixtures, Fisher D., Kozdoba M., Mannor S. , ICML 2020 ([arxiv](https://arxiv.org/abs/1906.06181)) ## Examples: * `synthetic_topic_test.py`: Topic reconstruction for random Dirichlet distributed topics and documents. * `newsgroup_test.py`: Topic Model example with 20newsgroups dataset. * `graph_communities_test.py`: Communities on Political Blogs graph. ## Requirements: This version of FDM was tested with Python 3.6 and TensorFlow 2.0.0 | 2,099 |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.