[Model] roberta-base-rte

config.json

@@ -0,0 +1,37 @@
+{
+  "_name_or_path": "/mnt/lustre/zhangyunchen/transformers/roberta-base",
+  "architectures": [
+    "RobertaForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_id": 2,
+  "finetuning_task": "rte",
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "entailment",
+    "1": "not_entailment"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "entailment": 0,
+    "not_entailment": 1
+  },
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "roberta",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "torch_dtype": "float32",
+  "transformers_version": "4.16.2",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 50265
+}

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e15c1e57b0d22bb0bbcf55345061135bedf378bc4101933c524aaacecdac8e2
+size 997304355

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6834dc699a39051f9c5f7a547613201de3b43cfec84e79c77942a8c52cb1db8
+size 498679699

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6b6e8f5df33beb4fa46b8e006bbe8caeba063c97699d506031e45a1a9aeaf22
+size 14659

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e625bd50edc4a882fbd82c92833af2ec43afd0065387850debe1807891f58c54
+size 623

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": false}}

tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "add_prefix_space": false, "errors": "replace", "sep_token": "</s>", "cls_token": "<s>", "pad_token": "<pad>", "mask_token": "<mask>", "trim_offsets": true, "special_tokens_map_file": null, "name_or_path": "/mnt/lustre/zhangyunchen/transformers/roberta-base", "tokenizer_class": "RobertaTokenizer"}

trainer_state.json

@@ -0,0 +1,316 @@
+{
+  "best_metric": 0.8050541516245487,
+  "best_model_checkpoint": "./fp32_1e_5/models/rte-roberta-base/checkpoint-1000",
+  "epoch": 12.820512820512821,
+  "global_step": 1000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.319148936170213e-06,
+      "loss": 0.6932,
+      "step": 50
+    },
+    {
+      "epoch": 0.64,
+      "eval_accuracy": 0.4729241877256318,
+      "eval_loss": 0.6963141560554504,
+      "eval_runtime": 2.0901,
+      "eval_samples_per_second": 132.533,
+      "eval_steps_per_second": 4.306,
+      "step": 50
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 9.959072305593452e-06,
+      "loss": 0.6957,
+      "step": 100
+    },
+    {
+      "epoch": 1.28,
+      "eval_accuracy": 0.4729241877256318,
+      "eval_loss": 0.69448322057724,
+      "eval_runtime": 0.7491,
+      "eval_samples_per_second": 369.778,
+      "eval_steps_per_second": 12.014,
+      "step": 100
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 9.618008185538881e-06,
+      "loss": 0.6869,
+      "step": 150
+    },
+    {
+      "epoch": 1.92,
+      "eval_accuracy": 0.6245487364620939,
+      "eval_loss": 0.6610168218612671,
+      "eval_runtime": 0.7857,
+      "eval_samples_per_second": 352.54,
+      "eval_steps_per_second": 11.454,
+      "step": 150
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 9.276944065484312e-06,
+      "loss": 0.6292,
+      "step": 200
+    },
+    {
+      "epoch": 2.56,
+      "eval_accuracy": 0.6570397111913358,
+      "eval_loss": 0.6132365465164185,
+      "eval_runtime": 0.7843,
+      "eval_samples_per_second": 353.202,
+      "eval_steps_per_second": 11.476,
+      "step": 200
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 8.935879945429742e-06,
+      "loss": 0.5473,
+      "step": 250
+    },
+    {
+      "epoch": 3.21,
+      "eval_accuracy": 0.7075812274368231,
+      "eval_loss": 0.6135809421539307,
+      "eval_runtime": 0.8034,
+      "eval_samples_per_second": 344.792,
+      "eval_steps_per_second": 11.203,
+      "step": 250
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 8.59481582537517e-06,
+      "loss": 0.4867,
+      "step": 300
+    },
+    {
+      "epoch": 3.85,
+      "eval_accuracy": 0.7256317689530686,
+      "eval_loss": 0.5372898578643799,
+      "eval_runtime": 0.8424,
+      "eval_samples_per_second": 328.815,
+      "eval_steps_per_second": 10.684,
+      "step": 300
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 8.253751705320601e-06,
+      "loss": 0.3673,
+      "step": 350
+    },
+    {
+      "epoch": 4.49,
+      "eval_accuracy": 0.7292418772563177,
+      "eval_loss": 0.6152286529541016,
+      "eval_runtime": 0.7525,
+      "eval_samples_per_second": 368.117,
+      "eval_steps_per_second": 11.96,
+      "step": 350
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 7.912687585266032e-06,
+      "loss": 0.3246,
+      "step": 400
+    },
+    {
+      "epoch": 5.13,
+      "eval_accuracy": 0.7545126353790613,
+      "eval_loss": 0.6094390153884888,
+      "eval_runtime": 0.7684,
+      "eval_samples_per_second": 360.479,
+      "eval_steps_per_second": 11.712,
+      "step": 400
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 7.5716234652114605e-06,
+      "loss": 0.2585,
+      "step": 450
+    },
+    {
+      "epoch": 5.77,
+      "eval_accuracy": 0.7725631768953068,
+      "eval_loss": 0.6251689791679382,
+      "eval_runtime": 0.7559,
+      "eval_samples_per_second": 366.46,
+      "eval_steps_per_second": 11.907,
+      "step": 450
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 7.23055934515689e-06,
+      "loss": 0.1975,
+      "step": 500
+    },
+    {
+      "epoch": 6.41,
+      "eval_accuracy": 0.7436823104693141,
+      "eval_loss": 0.7674959301948547,
+      "eval_runtime": 0.766,
+      "eval_samples_per_second": 361.606,
+      "eval_steps_per_second": 11.749,
+      "step": 500
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 6.88949522510232e-06,
+      "loss": 0.19,
+      "step": 550
+    },
+    {
+      "epoch": 7.05,
+      "eval_accuracy": 0.7617328519855595,
+      "eval_loss": 0.7783448696136475,
+      "eval_runtime": 0.7892,
+      "eval_samples_per_second": 350.981,
+      "eval_steps_per_second": 11.404,
+      "step": 550
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 6.54843110504775e-06,
+      "loss": 0.1266,
+      "step": 600
+    },
+    {
+      "epoch": 7.69,
+      "eval_accuracy": 0.7545126353790613,
+      "eval_loss": 1.0359293222427368,
+      "eval_runtime": 0.7751,
+      "eval_samples_per_second": 357.395,
+      "eval_steps_per_second": 11.612,
+      "step": 600
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 6.20736698499318e-06,
+      "loss": 0.1222,
+      "step": 650
+    },
+    {
+      "epoch": 8.33,
+      "eval_accuracy": 0.7833935018050542,
+      "eval_loss": 0.8435311913490295,
+      "eval_runtime": 1.3166,
+      "eval_samples_per_second": 210.387,
+      "eval_steps_per_second": 6.836,
+      "step": 650
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 5.8663028649386085e-06,
+      "loss": 0.1146,
+      "step": 700
+    },
+    {
+      "epoch": 8.97,
+      "eval_accuracy": 0.7833935018050542,
+      "eval_loss": 0.941527247428894,
+      "eval_runtime": 0.749,
+      "eval_samples_per_second": 369.818,
+      "eval_steps_per_second": 12.016,
+      "step": 700
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 5.525238744884038e-06,
+      "loss": 0.1131,
+      "step": 750
+    },
+    {
+      "epoch": 9.62,
+      "eval_accuracy": 0.7833935018050542,
+      "eval_loss": 0.9660640358924866,
+      "eval_runtime": 0.7502,
+      "eval_samples_per_second": 369.243,
+      "eval_steps_per_second": 11.997,
+      "step": 750
+    },
+    {
+      "epoch": 10.26,
+      "learning_rate": 5.1841746248294686e-06,
+      "loss": 0.0719,
+      "step": 800
+    },
+    {
+      "epoch": 10.26,
+      "eval_accuracy": 0.7906137184115524,
+      "eval_loss": 1.0486385822296143,
+      "eval_runtime": 0.7951,
+      "eval_samples_per_second": 348.395,
+      "eval_steps_per_second": 11.32,
+      "step": 800
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 4.843110504774898e-06,
+      "loss": 0.1002,
+      "step": 850
+    },
+    {
+      "epoch": 10.9,
+      "eval_accuracy": 0.776173285198556,
+      "eval_loss": 1.1470500230789185,
+      "eval_runtime": 0.7467,
+      "eval_samples_per_second": 370.982,
+      "eval_steps_per_second": 12.054,
+      "step": 850
+    },
+    {
+      "epoch": 11.54,
+      "learning_rate": 4.502046384720328e-06,
+      "loss": 0.0515,
+      "step": 900
+    },
+    {
+      "epoch": 11.54,
+      "eval_accuracy": 0.7906137184115524,
+      "eval_loss": 1.0903133153915405,
+      "eval_runtime": 0.7707,
+      "eval_samples_per_second": 359.412,
+      "eval_steps_per_second": 11.678,
+      "step": 900
+    },
+    {
+      "epoch": 12.18,
+      "learning_rate": 4.160982264665757e-06,
+      "loss": 0.0732,
+      "step": 950
+    },
+    {
+      "epoch": 12.18,
+      "eval_accuracy": 0.7978339350180506,
+      "eval_loss": 1.0761293172836304,
+      "eval_runtime": 0.8013,
+      "eval_samples_per_second": 345.69,
+      "eval_steps_per_second": 11.232,
+      "step": 950
+    },
+    {
+      "epoch": 12.82,
+      "learning_rate": 3.819918144611187e-06,
+      "loss": 0.0644,
+      "step": 1000
+    },
+    {
+      "epoch": 12.82,
+      "eval_accuracy": 0.8050541516245487,
+      "eval_loss": 1.118359088897705,
+      "eval_runtime": 0.7549,
+      "eval_samples_per_second": 366.937,
+      "eval_steps_per_second": 11.922,
+      "step": 1000
+    }
+  ],
+  "max_steps": 1560,
+  "num_train_epochs": 20,
+  "total_flos": 2100152443883520.0,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:26d756f7d0d0350688dc0b99497cc50be5c0cb31e741e1545025d3ef34d7e623
+size 3055