Training in progress, step 32000, checkpoint

checkpoint-32000/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<mask>": 64000
+}

checkpoint-32000/config.json

@@ -0,0 +1,31 @@
+{
+  "_name_or_path": "demdecuong/vihealthbert-base-word",
+  "architectures": [
+    "RobertaForMaskedLM"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_id": 2,
+  "finetuning_task": "word-level",
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 258,
+  "model_type": "roberta",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "output_hidden_states": true,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "tokenizer_class": "PhobertTokenizer",
+  "torch_dtype": "float32",
+  "transformers_version": "4.40.2",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 64001
+}

checkpoint-32000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17276afb8d701a65aa0b1d0aeb38e4925a83a34933bb9811218efa3a9a695b3c
+size 540279452

checkpoint-32000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83b8489792615f1d36571c1731be62d29d291f775dd5e13c91aac2f8990b8903
+size 1080631290

checkpoint-32000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:889c14f94703ea0c8e7552487c51c2c1312c01d80180ba8bb11a8fe95e9f79b8
+size 14244

checkpoint-32000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54764741dce52dc1d3db53cacb054c817dd0b31badb8f0c8e2dde7574656377c
+size 1064

checkpoint-32000/special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": "<mask>",
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "unk_token": "<unk>"
+}

checkpoint-32000/tokenizer_config.json

@@ -0,0 +1,54 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64000": {
+      "content": "<mask>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": "<mask>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "tokenizer_class": "PhobertTokenizer",
+  "unk_token": "<unk>"
+}

checkpoint-32000/trainer_state.json

@@ -0,0 +1,284 @@
+{
+  "best_metric": 0.6717793867023528,
+  "best_model_checkpoint": "output/pretraining/vihealthbert-w_unsup-SynPD/lr3e-5_wr0.1_wd0.0/checkpoint-30000",
+  "epoch": 5.51438910908151,
+  "eval_steps": 2000,
+  "global_step": 32000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.00017232465965879716,
+      "grad_norm": 35.608882904052734,
+      "learning_rate": 5.169739789763915e-09,
+      "loss": 10.7382,
+      "step": 1
+    },
+    {
+      "epoch": 0.34464931931759435,
+      "grad_norm": 4.884454250335693,
+      "learning_rate": 1.033947957952783e-05,
+      "loss": 5.8543,
+      "step": 2000
+    },
+    {
+      "epoch": 0.34464931931759435,
+      "eval_accuracy": 0.3950195480446223,
+      "eval_loss": 3.896709680557251,
+      "eval_runtime": 24.6709,
+      "eval_samples_per_second": 396.175,
+      "eval_steps_per_second": 24.766,
+      "step": 2000
+    },
+    {
+      "epoch": 0.6892986386351887,
+      "grad_norm": 5.278345108032227,
+      "learning_rate": 2.067895915905566e-05,
+      "loss": 3.4544,
+      "step": 4000
+    },
+    {
+      "epoch": 0.6892986386351887,
+      "eval_accuracy": 0.530626322082222,
+      "eval_loss": 2.811899423599243,
+      "eval_runtime": 24.6443,
+      "eval_samples_per_second": 396.602,
+      "eval_steps_per_second": 24.793,
+      "step": 4000
+    },
+    {
+      "epoch": 1.033947957952783,
+      "grad_norm": 4.731122016906738,
+      "learning_rate": 2.988684014015739e-05,
+      "loss": 2.8312,
+      "step": 6000
+    },
+    {
+      "epoch": 1.033947957952783,
+      "eval_accuracy": 0.5771132321637026,
+      "eval_loss": 2.4039628505706787,
+      "eval_runtime": 24.6266,
+      "eval_samples_per_second": 396.889,
+      "eval_steps_per_second": 24.811,
+      "step": 6000
+    },
+    {
+      "epoch": 1.3785972772703774,
+      "grad_norm": 12.213573455810547,
+      "learning_rate": 2.873800907576541e-05,
+      "loss": 2.5914,
+      "step": 8000
+    },
+    {
+      "epoch": 1.3785972772703774,
+      "eval_accuracy": 0.5349752225383133,
+      "eval_loss": 2.6481692790985107,
+      "eval_runtime": 24.6033,
+      "eval_samples_per_second": 397.264,
+      "eval_steps_per_second": 24.834,
+      "step": 8000
+    },
+    {
+      "epoch": 1.7232465965879717,
+      "grad_norm": 5.172228813171387,
+      "learning_rate": 2.758917801137343e-05,
+      "loss": 2.5649,
+      "step": 10000
+    },
+    {
+      "epoch": 1.7232465965879717,
+      "eval_accuracy": 0.6087076297937304,
+      "eval_loss": 2.13348388671875,
+      "eval_runtime": 24.6211,
+      "eval_samples_per_second": 396.976,
+      "eval_steps_per_second": 24.816,
+      "step": 10000
+    },
+    {
+      "epoch": 2.067895915905566,
+      "grad_norm": 4.871993064880371,
+      "learning_rate": 2.6440346946981447e-05,
+      "loss": 2.2749,
+      "step": 12000
+    },
+    {
+      "epoch": 2.067895915905566,
+      "eval_accuracy": 0.6281529792701336,
+      "eval_loss": 1.9895449876785278,
+      "eval_runtime": 24.453,
+      "eval_samples_per_second": 399.705,
+      "eval_steps_per_second": 24.987,
+      "step": 12000
+    },
+    {
+      "epoch": 2.4125452352231607,
+      "grad_norm": 8.037242889404297,
+      "learning_rate": 2.5291515882589467e-05,
+      "loss": 2.1572,
+      "step": 14000
+    },
+    {
+      "epoch": 2.4125452352231607,
+      "eval_accuracy": 0.6353159252095423,
+      "eval_loss": 1.9312896728515625,
+      "eval_runtime": 24.4658,
+      "eval_samples_per_second": 399.496,
+      "eval_steps_per_second": 24.974,
+      "step": 14000
+    },
+    {
+      "epoch": 2.757194554540755,
+      "grad_norm": 4.320974349975586,
+      "learning_rate": 2.4142684818197487e-05,
+      "loss": 2.1009,
+      "step": 16000
+    },
+    {
+      "epoch": 2.757194554540755,
+      "eval_accuracy": 0.6428579553856829,
+      "eval_loss": 1.864585280418396,
+      "eval_runtime": 24.4545,
+      "eval_samples_per_second": 399.68,
+      "eval_steps_per_second": 24.985,
+      "step": 16000
+    },
+    {
+      "epoch": 3.101843873858349,
+      "grad_norm": 4.6416168212890625,
+      "learning_rate": 2.2993853753805504e-05,
+      "loss": 2.0609,
+      "step": 18000
+    },
+    {
+      "epoch": 3.101843873858349,
+      "eval_accuracy": 0.6450457483325106,
+      "eval_loss": 1.8571828603744507,
+      "eval_runtime": 24.4634,
+      "eval_samples_per_second": 399.535,
+      "eval_steps_per_second": 24.976,
+      "step": 18000
+    },
+    {
+      "epoch": 3.4464931931759435,
+      "grad_norm": 4.341440677642822,
+      "learning_rate": 2.184502268941352e-05,
+      "loss": 2.0885,
+      "step": 20000
+    },
+    {
+      "epoch": 3.4464931931759435,
+      "eval_accuracy": 0.6284792764252838,
+      "eval_loss": 1.9489394426345825,
+      "eval_runtime": 24.6433,
+      "eval_samples_per_second": 396.618,
+      "eval_steps_per_second": 24.794,
+      "step": 20000
+    },
+    {
+      "epoch": 3.791142512493538,
+      "grad_norm": 4.660053253173828,
+      "learning_rate": 2.069619162502154e-05,
+      "loss": 1.9891,
+      "step": 22000
+    },
+    {
+      "epoch": 3.791142512493538,
+      "eval_accuracy": 0.6582960564701575,
+      "eval_loss": 1.7700080871582031,
+      "eval_runtime": 24.616,
+      "eval_samples_per_second": 397.059,
+      "eval_steps_per_second": 24.821,
+      "step": 22000
+    },
+    {
+      "epoch": 4.135791831811132,
+      "grad_norm": 3.9445292949676514,
+      "learning_rate": 1.954736056062956e-05,
+      "loss": 1.9368,
+      "step": 24000
+    },
+    {
+      "epoch": 4.135791831811132,
+      "eval_accuracy": 0.6608816779512006,
+      "eval_loss": 1.7397598028182983,
+      "eval_runtime": 24.6251,
+      "eval_samples_per_second": 396.912,
+      "eval_steps_per_second": 24.812,
+      "step": 24000
+    },
+    {
+      "epoch": 4.480441151128726,
+      "grad_norm": 4.561331748962402,
+      "learning_rate": 1.839852949623758e-05,
+      "loss": 1.9003,
+      "step": 26000
+    },
+    {
+      "epoch": 4.480441151128726,
+      "eval_accuracy": 0.666368433126864,
+      "eval_loss": 1.7164973020553589,
+      "eval_runtime": 24.4665,
+      "eval_samples_per_second": 399.485,
+      "eval_steps_per_second": 24.973,
+      "step": 26000
+    },
+    {
+      "epoch": 4.825090470446321,
+      "grad_norm": 4.358778476715088,
+      "learning_rate": 1.72496984318456e-05,
+      "loss": 1.9058,
+      "step": 28000
+    },
+    {
+      "epoch": 4.825090470446321,
+      "eval_accuracy": 0.666961039556835,
+      "eval_loss": 1.703224539756775,
+      "eval_runtime": 24.4455,
+      "eval_samples_per_second": 399.829,
+      "eval_steps_per_second": 24.994,
+      "step": 28000
+    },
+    {
+      "epoch": 5.1697397897639155,
+      "grad_norm": 4.522342205047607,
+      "learning_rate": 1.6100867367453616e-05,
+      "loss": 1.859,
+      "step": 30000
+    },
+    {
+      "epoch": 5.1697397897639155,
+      "eval_accuracy": 0.6717793867023528,
+      "eval_loss": 1.677147626876831,
+      "eval_runtime": 24.4697,
+      "eval_samples_per_second": 399.432,
+      "eval_steps_per_second": 24.97,
+      "step": 30000
+    },
+    {
+      "epoch": 5.51438910908151,
+      "grad_norm": 4.162671089172363,
+      "learning_rate": 1.4952036303061636e-05,
+      "loss": 1.8401,
+      "step": 32000
+    },
+    {
+      "epoch": 5.51438910908151,
+      "eval_accuracy": 0.6709595439247201,
+      "eval_loss": 1.665189504623413,
+      "eval_runtime": 24.6261,
+      "eval_samples_per_second": 396.897,
+      "eval_steps_per_second": 24.811,
+      "step": 32000
+    }
+  ],
+  "logging_steps": 2000,
+  "max_steps": 58030,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 2000,
+  "total_flos": 0.0,
+  "train_batch_size": 32,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-32000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:828a11e9b75940b4120312d1450c3472eb18a2c7fa0179d8afafbcd0800beb75
+size 5304