[Model] roberta-base-sst2

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": "sst2",
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "negative",
+    "1": "positive"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "negative": 0,
+    "positive": 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:e35fcab13ef84b0ef597e4f55ce4ad924be748df5a326624f8adc3d310db8d05
+size 997304355

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b9afe4a71de317b22919517eb9c369786dad608a31a7e5f9c5faedf7592b2b77
+size 498679699

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c893fcf4140352f1cc9787a4deecbb1bfd2590635200d93726863a9a8e4a5243
+size 14659

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13694d279ea42ce8f3ea6441084b51dd198d9a31e21cbb70cf02cc92c213461a
+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 +1,301 @@
+{
+  "best_metric": 0.9518348623853211,
+  "best_model_checkpoint": "./fp32/models/sst2-roberta-base/checkpoint-18000",
+  "epoch": 9.026128266033254,
+  "global_step": 19000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.48,
+      "learning_rate": 7.917656373713381e-06,
+      "loss": 0.4065,
+      "step": 1000
+    },
+    {
+      "epoch": 0.48,
+      "eval_accuracy": 0.9220183486238532,
+      "eval_loss": 0.2038680613040924,
+      "eval_runtime": 4.1534,
+      "eval_samples_per_second": 209.949,
+      "eval_steps_per_second": 6.741,
+      "step": 1000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 9.627533228887655e-06,
+      "loss": 0.2172,
+      "step": 2000
+    },
+    {
+      "epoch": 0.95,
+      "eval_accuracy": 0.9323394495412844,
+      "eval_loss": 0.19802363216876984,
+      "eval_runtime": 5.7226,
+      "eval_samples_per_second": 152.378,
+      "eval_steps_per_second": 4.893,
+      "step": 2000
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 9.122150907161269e-06,
+      "loss": 0.1695,
+      "step": 3000
+    },
+    {
+      "epoch": 1.43,
+      "eval_accuracy": 0.9426605504587156,
+      "eval_loss": 0.19059622287750244,
+      "eval_runtime": 6.285,
+      "eval_samples_per_second": 138.743,
+      "eval_steps_per_second": 4.455,
+      "step": 3000
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 8.616768585434883e-06,
+      "loss": 0.1624,
+      "step": 4000
+    },
+    {
+      "epoch": 1.9,
+      "eval_accuracy": 0.9403669724770642,
+      "eval_loss": 0.21938160061836243,
+      "eval_runtime": 5.6855,
+      "eval_samples_per_second": 153.373,
+      "eval_steps_per_second": 4.925,
+      "step": 4000
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 8.111386263708495e-06,
+      "loss": 0.1308,
+      "step": 5000
+    },
+    {
+      "epoch": 2.38,
+      "eval_accuracy": 0.9243119266055045,
+      "eval_loss": 0.2738950848579407,
+      "eval_runtime": 5.7111,
+      "eval_samples_per_second": 152.686,
+      "eval_steps_per_second": 4.903,
+      "step": 5000
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 7.60600394198211e-06,
+      "loss": 0.1305,
+      "step": 6000
+    },
+    {
+      "epoch": 2.85,
+      "eval_accuracy": 0.9415137614678899,
+      "eval_loss": 0.2265118807554245,
+      "eval_runtime": 6.4711,
+      "eval_samples_per_second": 134.753,
+      "eval_steps_per_second": 4.327,
+      "step": 6000
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 7.100621620255725e-06,
+      "loss": 0.1046,
+      "step": 7000
+    },
+    {
+      "epoch": 3.33,
+      "eval_accuracy": 0.9438073394495413,
+      "eval_loss": 0.22793179750442505,
+      "eval_runtime": 5.0992,
+      "eval_samples_per_second": 171.007,
+      "eval_steps_per_second": 5.491,
+      "step": 7000
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 6.595239298529338e-06,
+      "loss": 0.1076,
+      "step": 8000
+    },
+    {
+      "epoch": 3.8,
+      "eval_accuracy": 0.9392201834862385,
+      "eval_loss": 0.2734154760837555,
+      "eval_runtime": 6.5088,
+      "eval_samples_per_second": 133.973,
+      "eval_steps_per_second": 4.302,
+      "step": 8000
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 6.089856976802951e-06,
+      "loss": 0.0974,
+      "step": 9000
+    },
+    {
+      "epoch": 4.28,
+      "eval_accuracy": 0.9415137614678899,
+      "eval_loss": 0.29046502709388733,
+      "eval_runtime": 5.7978,
+      "eval_samples_per_second": 150.403,
+      "eval_steps_per_second": 4.829,
+      "step": 9000
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 5.584474655076566e-06,
+      "loss": 0.092,
+      "step": 10000
+    },
+    {
+      "epoch": 4.75,
+      "eval_accuracy": 0.948394495412844,
+      "eval_loss": 0.22252324223518372,
+      "eval_runtime": 6.2131,
+      "eval_samples_per_second": 140.348,
+      "eval_steps_per_second": 4.507,
+      "step": 10000
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 5.07909233335018e-06,
+      "loss": 0.0836,
+      "step": 11000
+    },
+    {
+      "epoch": 5.23,
+      "eval_accuracy": 0.9461009174311926,
+      "eval_loss": 0.2741754353046417,
+      "eval_runtime": 5.948,
+      "eval_samples_per_second": 146.605,
+      "eval_steps_per_second": 4.708,
+      "step": 11000
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 4.573710011623794e-06,
+      "loss": 0.0802,
+      "step": 12000
+    },
+    {
+      "epoch": 5.7,
+      "eval_accuracy": 0.9472477064220184,
+      "eval_loss": 0.2603645324707031,
+      "eval_runtime": 5.8251,
+      "eval_samples_per_second": 149.696,
+      "eval_steps_per_second": 4.807,
+      "step": 12000
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 4.068327689897408e-06,
+      "loss": 0.0702,
+      "step": 13000
+    },
+    {
+      "epoch": 6.18,
+      "eval_accuracy": 0.9461009174311926,
+      "eval_loss": 0.3199061155319214,
+      "eval_runtime": 5.8055,
+      "eval_samples_per_second": 150.202,
+      "eval_steps_per_second": 4.823,
+      "step": 13000
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 3.562945368171022e-06,
+      "loss": 0.0673,
+      "step": 14000
+    },
+    {
+      "epoch": 6.65,
+      "eval_accuracy": 0.9438073394495413,
+      "eval_loss": 0.29321831464767456,
+      "eval_runtime": 5.6513,
+      "eval_samples_per_second": 154.302,
+      "eval_steps_per_second": 4.955,
+      "step": 14000
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 3.0575630464446353e-06,
+      "loss": 0.0683,
+      "step": 15000
+    },
+    {
+      "epoch": 7.13,
+      "eval_accuracy": 0.9506880733944955,
+      "eval_loss": 0.28904497623443604,
+      "eval_runtime": 5.7574,
+      "eval_samples_per_second": 151.459,
+      "eval_steps_per_second": 4.863,
+      "step": 15000
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 2.5521807247182495e-06,
+      "loss": 0.0608,
+      "step": 16000
+    },
+    {
+      "epoch": 7.6,
+      "eval_accuracy": 0.948394495412844,
+      "eval_loss": 0.3019457757472992,
+      "eval_runtime": 6.5808,
+      "eval_samples_per_second": 132.506,
+      "eval_steps_per_second": 4.255,
+      "step": 16000
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 2.0467984029918637e-06,
+      "loss": 0.0553,
+      "step": 17000
+    },
+    {
+      "epoch": 8.08,
+      "eval_accuracy": 0.9426605504587156,
+      "eval_loss": 0.33538028597831726,
+      "eval_runtime": 5.9627,
+      "eval_samples_per_second": 146.243,
+      "eval_steps_per_second": 4.696,
+      "step": 17000
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 1.5414160812654775e-06,
+      "loss": 0.0494,
+      "step": 18000
+    },
+    {
+      "epoch": 8.55,
+      "eval_accuracy": 0.9518348623853211,
+      "eval_loss": 0.3174615502357483,
+      "eval_runtime": 9.642,
+      "eval_samples_per_second": 90.438,
+      "eval_steps_per_second": 2.904,
+      "step": 18000
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 1.0360337595390914e-06,
+      "loss": 0.0507,
+      "step": 19000
+    },
+    {
+      "epoch": 9.03,
+      "eval_accuracy": 0.9518348623853211,
+      "eval_loss": 0.3315552771091461,
+      "eval_runtime": 5.8184,
+      "eval_samples_per_second": 149.869,
+      "eval_steps_per_second": 4.812,
+      "step": 19000
+    }
+  ],
+  "max_steps": 21050,
+  "num_train_epochs": 10,
+  "total_flos": 3.998636841609984e+16,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7e440150530807bf13dcb780c05b33c0df5ee5dc9ba1bcf062637a89e3b7523
+size 3055