I have run train.py with this config and after 13 epoch I still get 0 fscore.
{
"task_info":{
"label_type": "multi_label",
"hierarchical": true,
"hierar_taxonomy": "data/rcv1.taxonomy",
"hierar_penalty": 0.000001
},
"device": "cuda",
"model_name": "TextCNN",
"checkpoint_dir": "checkpoint_dir_rcv1",
"model_dir": "trained_model_rcv1",
"data": {
"train_json_files": [
"data/rcv1_train.json"
],
"validate_json_files": [
"data/rcv1_dev.json"
],
"test_json_files": [
"data/rcv1_test.json"
],
"generate_dict_using_json_files": false,
"generate_dict_using_all_json_files": false,
"generate_dict_using_pretrained_embedding": false,
"dict_dir": "dict_rcv1",
"num_worker": 4
},
"feature": {
"feature_names": [
"token"
],
"min_token_count": 2,
"min_char_count": 2,
"token_ngram": 0,
"min_token_ngram_count": 0,
"min_keyword_count": 0,
"min_topic_count": 2,
"max_token_dict_size": 1000000,
"max_char_dict_size": 150000,
"max_token_ngram_dict_size": 10000000,
"max_keyword_dict_size": 100,
"max_topic_dict_size": 100,
"max_token_len": 256,
"max_char_len": 1024,
"max_char_len_per_token": 4,
"token_pretrained_file": "",
"keyword_pretrained_file": ""
},
"train": {
"batch_size": 64,
"start_epoch": 1,
"num_epochs": 20,
"num_epochs_static_embedding": 0,
"decay_steps": 1000,
"decay_rate": 1.0,
"clip_gradients": 100.0,
"l2_lambda": 0.0,
"loss_type": "BCEWithLogitsLoss",
"sampler": "fixed",
"num_sampled": 5,
"visible_device_list": "0",
"hidden_layer_dropout": 0.5
},
"embedding": {
"type": "embedding",
"dimension": 64,
"region_embedding_type": "context_word",
"region_size": 5,
"initializer": "uniform",
"fan_mode": "FAN_IN",
"uniform_bound": 0.25,
"random_stddev": 0.01,
"dropout": 0.0
},
"optimizer": {
"optimizer_type": "Adam",
"learning_rate": 0.008,
"adadelta_decay_rate": 0.95,
"adadelta_epsilon": 1e-08
},
"TextCNN": {
"kernel_sizes": [
2,
3,
4
],
"num_kernels": 100,
"top_k_max_pooling": 1
},
"TextRNN": {
"hidden_dimension": 64,
"rnn_type": "GRU",
"num_layers": 1,
"doc_embedding_type": "Attention",
"attention_dimension": 16,
"bidirectional": true
},
"DRNN": {
"hidden_dimension": 5,
"window_size": 3,
"rnn_type": "GRU",
"bidirectional": true,
"cell_hidden_dropout": 0.1
},
"eval": {
"text_file": "data/rcv1_test.json",
"threshold": 0.5,
"dir": "eval_dir",
"batch_size": 1024,
"is_flat": true,
"top_k": 30,
"model_dir": "checkpoint_dir_rcv1/TextCNN_best"
},
"TextVDCNN": {
"vdcnn_depth": 9,
"top_k_max_pooling": 8
},
"DPCNN": {
"kernel_size": 3,
"pooling_stride": 2,
"num_kernels": 16,
"blocks": 2
},
"TextRCNN": {
"kernel_sizes": [
2,
3,
4
],
"num_kernels": 100,
"top_k_max_pooling": 1,
"hidden_dimension":64,
"rnn_type": "GRU",
"num_layers": 1,
"bidirectional": true
},
"Transformer": {
"d_inner": 128,
"d_k": 32,
"d_v": 32,
"n_head": 4,
"n_layers": 1,
"dropout": 0.1,
"use_star": true
},
"AttentiveConvNet": {
"attention_type": "bilinear",
"margin_size": 3,
"type": "advanced",
"hidden_size": 64
},
"log": {
"logger_file": "log_test_rcv1_hierar",
"log_level": "warn"
}
}
Size of doc_token dict is 0
Size of doc_char dict is 0
Size of doc_token_ngram dict is 0
Size of doc_keyword dict is 0
Size of doc_topic dict is 0
Shrink dict over.
Size of doc_label dict is 100
Size of doc_token dict is 0
Size of doc_char dict is 0
Size of doc_token_ngram dict is 0
Size of doc_keyword dict is 0
Size of doc_topic dict is 0
Train performance at epoch 1 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.176090.
Validate performance at epoch 1 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.175878.
test performance at epoch 1 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.176301.
Epoch 1 cost time: 273 second
Train performance at epoch 2 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.185657.
Validate performance at epoch 2 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.185401.
test performance at epoch 2 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.185863.
Epoch 2 cost time: 275 second
Train performance at epoch 3 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.179540.
Validate performance at epoch 3 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.179308.
test performance at epoch 3 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.179765.
Epoch 3 cost time: 274 second
Train performance at epoch 4 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.181358.
Validate performance at epoch 4 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.181102.
test performance at epoch 4 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.181564.
Epoch 4 cost time: 274 second
Train performance at epoch 5 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.177109.
Validate performance at epoch 5 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.176923.
test performance at epoch 5 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.177328.
Epoch 5 cost time: 274 second
Train performance at epoch 6 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.180688.
Validate performance at epoch 6 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.180476.
test performance at epoch 6 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.180903.
Epoch 6 cost time: 275 second
Train performance at epoch 7 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.177683.
Validate performance at epoch 7 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.177484.
test performance at epoch 7 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.177902.
Epoch 7 cost time: 278 second
Train performance at epoch 8 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.180226.
Validate performance at epoch 8 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.179987.
test performance at epoch 8 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.180454.
Epoch 8 cost time: 277 second
Train performance at epoch 9 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.176957.
Validate performance at epoch 9 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.176734.
test performance at epoch 9 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.177183.
Epoch 9 cost time: 277 second
Train performance at epoch 10 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.192425.
Validate performance at epoch 10 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.192209.
test performance at epoch 10 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.192622.
Epoch 10 cost time: 275 second
Train performance at epoch 11 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.181247.
Validate performance at epoch 11 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.181020.
test performance at epoch 11 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.181468.
Epoch 11 cost time: 277 second
Train performance at epoch 12 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.179442.
Validate performance at epoch 12 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.179230.
test performance at epoch 12 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.179660.
Epoch 12 cost time: 276 second
Train performance at epoch 13 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 259686.
Loss is: 0.182252.
Validate performance at epoch 13 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 28687.
Loss is: 0.182036.
test performance at epoch 13 is precision: 0.000000, recall: 0.000000, fscore: 0.000000, macro-fscore: 0.000000, right: 0, predict: 0, standard: 72472.
Loss is: 0.182464.```