Upload 6 files

config/HiLAT_5/config.json

@@ -0,0 +1,59 @@
+{
+  "model_name_or_path": "pretrained/ClinicalplusXLNet/",
+  "tokenizer_name": "xlnet-base-cased",
+  "output_dir": "../model/",
+  "overwrite_output_dir": true,
+
+  "task_name": "mimic3-50",
+  "max_seq_length": 512,
+  "train_file": "../data/mimic3/5/train_data_5_level_1.csv",
+  "validation_file": "../data/mimic3/5/dev_data_5_level_1.csv",
+  "test_file": "../data/mimic3/5/test_data_5_level_1.csv",
+  "label_dictionary_file": "../data/mimic3/5/labels_dictionary_5_level_1.csv",
+  "ignore_keys_for_eval": ["preds", "label_attention_weights", "chunk_attention_weights"],
+  "use_cached_datasets": true,
+  "data_segmented": true,
+  "lazy_loading": false,
+
+  "d_model": 768,
+  "dropout": 0.1,
+  "dropout_att": 0.1,
+  "num_chunks_per_document": 10,
+  "transformer_layer_update_strategy": "all",
+  "use_code_representation": false,
+  "multi_head_attention": false,
+  "chunk_attention": true,
+  "multi_head_chunk_attention": false,
+  "linear_init_mean": 0.0,
+  "linear_init_std": 0.03,
+  "document_pooling_strategy": "mean",
+
+  "do_train": true,
+  "do_eval": true,
+  "do_predict": true,
+  "evaluation_strategy": "steps",
+  "per_device_train_batch_size": 4,
+  "per_device_eval_batch_size": 4,
+  "gradient_accumulation_steps": 1,
+  "learning_rate": 5e-5,
+  "weight_decay": 0.1,
+  "num_train_epochs": 100,
+  "max_steps": 10500,
+  "warmup_steps": 500,
+  "log_level": "error",
+  "logging_strategy": "steps",
+  "logging_steps": 500,
+  "save_strategy": "steps",
+  "save_steps": 500,
+  "seed": 2022,
+  "dataloader_drop_last": false,
+  "disable_tqdm": false,
+  "label_names": ["targets"],
+  "load_best_model_at_end": false,
+  "metric_for_best_model": "micro_f1",
+  "greater_is_better": true,
+  "remove_unused_columns": false,
+  "num_hidden_layers": 0,
+  "report_to": "none",
+  "resume_from_checkpoint": "C:/Users/mkanneganti/Downloads/new/model/checkpoint-10000- Hilat main 05/"
+}

config/HiLAT_50/config.json

@@ -0,0 +1,57 @@
+{
+  "model_name_or_path": "pretrained/ClinicalplusXLNet/",
+  "tokenizer_name": "xlnet-base-cased",
+  "output_dir": "../model/",
+  "overwrite_output_dir": true,
+
+  "task_name": "mimic3-50",
+  "max_seq_length": 512,
+  "train_file": "../data/mimic3/50/train_data_50_level_1_new.csv",
+  "validation_file": "../data/mimic3/50/dev_data_50_level_1_new.csv",
+  "test_file": "../data/mimic3/50/test_data_50_level_1_new.csv",
+  "label_dictionary_file": "../data/mimic3/50/labels_dictionary_50_level_1.csv",
+  "ignore_keys_for_eval": ["preds", "label_attention_weights", "chunk_attention_weights"],
+  "use_cached_datasets": true,
+  "data_segmented": false,
+  "lazy_loading": false,
+
+  "d_model": 768,
+  "dropout": 0.1,
+  "dropout_att": 0.1,
+  "num_chunks_per_document": 10,
+  "transformer_layer_update_strategy": "all",
+  "use_code_representation": false,
+  "multi_head_attention": false,
+  "chunk_attention": true,
+  "multi_head_chunk_attention": false,
+  "linear_init_mean": 0.0,
+  "linear_init_std": 0.03,
+  "document_pooling_strategy": "mean",
+
+  "do_train": true,
+  "do_eval": true,
+  "do_predict": true,
+  "evaluation_strategy": "steps",
+  "per_device_train_batch_size": 4,
+  "per_device_eval_batch_size": 4,
+  "gradient_accumulation_steps": 1,
+  "learning_rate": 5e-5,
+  "weight_decay": 0.1,
+  "num_train_epochs": 100,
+  "max_steps": 10000,
+  "warmup_steps": 500,
+  "log_level": "error",
+  "logging_strategy": "steps",
+  "logging_steps": 500,
+  "save_strategy": "steps",
+  "save_steps": 500,
+  "seed": 2022,
+  "dataloader_drop_last": false,
+  "disable_tqdm": false,
+  "label_names": ["targets"],
+  "load_best_model_at_end": false,
+  "metric_for_best_model": "micro_f1",
+  "greater_is_better": true,
+  "remove_unused_columns": false,
+  "report_to": "none"
+}

config/Long-HiLAT_5/config.json

@@ -0,0 +1,60 @@
+{
+  "model_name_or_path": "pretrained/ClinicalplusXLNet/",
+  "tokenizer_name": "xlnet-base-cased",
+  "output_dir": "../model/",
+  "overwrite_output_dir": true,
+
+  "task_name": "mimic3-50",
+  "max_seq_length": 512,
+  "train_file": "../data/mimic3/5/train_data_5_level_1_new.csv",
+  "validation_file": "../data/mimic3/5/dev_data_5_level_1_new.csv",
+  "test_file": "../data/mimic3/5/test_data_5_level_1_new.csv",
+  "label_dictionary_file": "../data/mimic3/5/labels_dictionary_5_level_1.csv",
+  "ignore_keys_for_eval": ["preds", "label_attention_weights", "chunk_attention_weights"],
+  "use_cached_datasets": true,
+  "data_segmented": true,
+  "lazy_loading": false,
+
+  "d_model": 768,
+  "dropout": 0.1,
+  "dropout_att": 0.1,
+  "num_chunks_per_document": 8,
+  "transformer_layer_update_strategy": "all",
+  "use_code_representation": false,
+  "multi_head_attention": false,
+  "chunk_attention": true,
+  "multi_head_chunk_attention": false,
+  "linear_init_mean": 0.0,
+  "linear_init_std": 0.03,
+  "document_pooling_strategy": "mean",
+
+  "do_train": true,
+  "do_eval": true,
+  "do_predict": true,
+  "evaluation_strategy": "steps",
+  "per_device_train_batch_size": 2,
+  "per_device_eval_batch_size": 2,
+  "gradient_accumulation_steps": 1,
+  "learning_rate": 5e-5,
+  "weight_decay": 0.1,
+  "num_train_epochs": 100,
+  "max_steps": 10500,
+  "warmup_steps": 500,
+  "log_level": "error",
+  "logging_strategy": "steps",
+  "logging_steps": 500,
+  "save_strategy": "steps",
+  "save_steps": 500,
+  "seed": 2022,
+  "dataloader_drop_last": false,
+  "disable_tqdm": false,
+  "label_names": ["targets"],
+  "load_best_model_at_end": true,
+  "metric_for_best_model": "micro_f1",
+  "greater_is_better": true,
+  "remove_unused_columns": false,
+  "report_to": "none",
+  "num_hidden_layers": 2,
+  "resume_from_checkpoint": "C:/Users/mkanneganti/Downloads/new/model/checkpoint-10500- Long HiLAT 05"
+
+}

config/Long-HiLAT_50/config.json

@@ -0,0 +1,60 @@
+{
+  "model_name_or_path": "pretrained/ClinicalplusXLNet/",
+  "tokenizer_name": "xlnet-base-cased",
+  "output_dir": "../model/",
+  "overwrite_output_dir": true,
+
+  "task_name": "mimic3-50",
+  "max_seq_length": 512,
+  "train_file": "../data/mimic3/50/train_data_50_level_1_new_8.csv",
+  "validation_file": "../data/mimic3/50/dev_data_50_level_1_new_8.csv",
+  "test_file": "../data/mimic3/50/test_data_50_level_1_new_8.csv",
+  "label_dictionary_file": "../data/mimic3/50/labels_dictionary_50_level_1.csv",
+  "ignore_keys_for_eval": ["preds", "label_attention_weights", "chunk_attention_weights"],
+  "use_cached_datasets": true,
+  "data_segmented": true,
+  "lazy_loading": false,
+
+  "d_model": 768,
+  "dropout": 0.1,
+  "dropout_att": 0.1,
+  "num_chunks_per_document": 8,
+  "transformer_layer_update_strategy": "all",
+  "use_code_representation": false,
+  "multi_head_attention": false,
+  "chunk_attention": true,
+  "multi_head_chunk_attention": false,
+  "linear_init_mean": 0.0,
+  "linear_init_std": 0.03,
+  "document_pooling_strategy": "mean",
+
+  "do_train": true,
+  "do_eval": true,
+  "do_predict": true,
+  "evaluation_strategy": "steps",
+  "per_device_train_batch_size": 2,
+  "per_device_eval_batch_size": 2,
+  "gradient_accumulation_steps": 1,
+  "learning_rate": 5e-5,
+  "weight_decay": 0.1,
+  "num_train_epochs": 100,
+  "max_steps": 20000,
+  "warmup_steps": 500,
+  "log_level": "error",
+  "logging_strategy": "steps",
+  "logging_steps": 500,
+  "save_strategy": "steps",
+  "save_steps": 500,
+  "seed": 2022,
+  "dataloader_drop_last": false,
+  "disable_tqdm": false,
+  "label_names": ["targets"],
+  "load_best_model_at_end": true,
+  "metric_for_best_model": "micro_f1",
+  "greater_is_better": true,
+  "remove_unused_columns": false,
+  "report_to": "none",
+  "num_hidden_layers": 2,
+  "resume_from_checkpoint": "C:/Users/mkanneganti/Downloads/new/model/checkpoint-10500-Long-HiLAT 50"
+
+}

config/Long-LAT_5/config.json

@@ -0,0 +1,59 @@
+{
+  "model_name_or_path": "../pretrained/ClinicalplusXLNet/",
+  "tokenizer_name": "yikuan8/Clinical-Longformer",
+  "output_dir": "../model/",
+  "overwrite_output_dir": true,
+
+  "task_name": "mimic3-50",
+  "max_seq_length": 4096,
+  "train_file": "../data/mimic3/5/train_data_5_level_1_new_combined.csv",
+  "validation_file": "../data/mimic3/5/dev_data_5_level_1_new_combined.csv",
+  "test_file": "../data/mimic3/5/test_data_5_level_1_new_combined.csv",
+  "label_dictionary_file": "../data/mimic3/5/labels_dictionary_5_level_1.csv",
+  "ignore_keys_for_eval": ["preds", "label_attention_weights", "chunk_attention_weights"],
+  "use_cached_datasets": true,
+  "data_segmented": false,
+  "lazy_loading": false,
+
+  "d_model": 768,
+  "dropout": 0.1,
+  "dropout_att": 0.1,
+  "num_chunks_per_document": 1,
+  "transformer_layer_update_strategy": "all",
+  "use_code_representation": false,
+  "multi_head_attention": false,
+  "chunk_attention": false,
+  "multi_head_chunk_attention": false,
+  "linear_init_mean": 0.0,
+  "linear_init_std": 0.03,
+  "document_pooling_strategy": "mean",
+
+  "do_train": true,
+  "do_eval": true,
+  "do_predict": true,
+  "evaluation_strategy": "steps",
+  "per_device_train_batch_size": 4,
+  "per_device_eval_batch_size": 4,
+  "gradient_accumulation_steps": 2,
+  "learning_rate": 5e-5,
+  "weight_decay": 0.1,
+  "num_train_epochs": 100,
+  "max_steps": 10000,
+  "warmup_steps": 500,
+  "log_level": "error",
+  "logging_strategy": "steps",
+  "logging_steps": 500,
+  "save_strategy": "steps",
+  "save_steps": 500,
+  "seed": 2022,
+  "dataloader_drop_last": false,
+  "disable_tqdm": false,
+  "label_names": ["targets"],
+  "load_best_model_at_end": false,
+  "metric_for_best_model": "micro_f1",
+  "greater_is_better": true,
+  "remove_unused_columns": false,
+  "report_to": "none",
+  "num_hidden_layers": 12,
+  "resume_from_checkpoint": "C:/Users/mkanneganti/Downloads/new/model/checkpoint-10000- Long LAT 05"
+}

config/Long-LAT_50/config.json

@@ -0,0 +1,59 @@
+{
+  "model_name_or_path": "pretrained/ClinicalplusXLNet/",
+  "tokenizer_name": "yikuan8/Clinical-Longformer",
+  "output_dir": "../../model/",
+  "overwrite_output_dir": true,
+
+  "task_name": "mimic3-50",
+  "max_seq_length": 4096,
+  "train_file": "../../data/mimic3/50/train_data_50_level_1_new_combined.csv",
+  "validation_file": "../../data/mimic3/50/dev_data_50_level_1_new_combined.csv",
+  "test_file": "../../data/mimic3/50/test_data_50_level_1_new_combined.csv",
+  "label_dictionary_file": "../../data/mimic3/50/labels_dictionary_50_level_1.csv",
+  "ignore_keys_for_eval": ["preds", "label_attention_weights", "chunk_attention_weights"],
+  "use_cached_datasets": true,
+  "data_segmented": false,
+  "lazy_loading": false,
+
+  "d_model": 768,
+  "dropout": 0.1,
+  "dropout_att": 0.1,
+  "num_chunks_per_document": 1,
+  "transformer_layer_update_strategy": "all",
+  "use_code_representation": false,
+  "multi_head_attention": false,
+  "chunk_attention": false,
+  "multi_head_chunk_attention": false,
+  "linear_init_mean": 0.0,
+  "linear_init_std": 0.03,
+  "document_pooling_strategy": "mean",
+
+  "do_train": true,
+  "do_eval": true,
+  "do_predict": true,
+  "evaluation_strategy": "steps",
+  "per_device_train_batch_size": 4,
+  "per_device_eval_batch_size": 4,
+  "gradient_accumulation_steps": 2,
+  "learning_rate": 5e-5,
+  "weight_decay": 0.1,
+  "num_train_epochs": 100,
+  "max_steps": 10500,
+  "warmup_steps": 500,
+  "log_level": "error",
+  "logging_strategy": "steps",
+  "logging_steps": 500,
+  "save_strategy": "steps",
+  "save_steps": 500,
+  "seed": 2022,
+  "dataloader_drop_last": false,
+  "disable_tqdm": false,
+  "label_names": ["targets"],
+  "load_best_model_at_end": false,
+  "metric_for_best_model": "micro_f1",
+  "greater_is_better": true,
+  "remove_unused_columns": false,
+  "report_to": "none",
+  "num_hidden_layers": 16,
+  "resume_from_checkpoint": "C:/Users/mkanneganti/Downloads/new/model/checkpoint-10000-Long LAT 50_5e-5"
+}