bert-base-uncased-rte

config.json

@@ -0,0 +1,21 @@
+{
+  "architectures": [
+    "BertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "finetuning_task": "glue:rte",
+  "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-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "type_vocab_size": 2,
+  "vocab_size": 30522
+}

eval_results_rte.txt

@@ -0,0 +1,3 @@
+eval_loss = 0.6541814684867859
+eval_acc = 0.6245487364620939
+epoch = 3.0

flax_model.msgpack

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc0293bd2736130884861f5891b9822e1e2109d2a8c4d6bf8db648b04b080ea5
+size 437942328

log.txt

@@ -0,0 +1,44 @@
+Writing logs to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/log.txt.
+Loading nlp dataset glue, subset rte, split train.
+Loading nlp dataset glue, subset rte, split validation.
+Loaded dataset. Found: 2 labels: ([0, 1])
+Loading transformers AutoModelForSequenceClassification: bert-base-uncased
+Tokenizing training data. (len: 2490)
+Tokenizing eval data (len: 277)
+Loaded data and tokenized in 14.295648097991943s
+Training model across 1 GPUs
+***** Running training *****
+	Num examples = 2490
+	Batch size = 128
+	Max sequence length = 128
+	Num steps = 95
+	Num epochs = 5
+	Learning rate = 3e-05
+Failed to predict with model <class 'transformers.modeling_bert.BertForSequenceClassification'>. Check tokenizer configuration.
+Writing logs to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/log.txt.
+Loading nlp dataset glue, subset rte, split train.
+Loading nlp dataset glue, subset rte, split validation.
+Loaded dataset. Found: 2 labels: ([0, 1])
+Loading transformers AutoModelForSequenceClassification: bert-base-uncased
+Tokenizing training data. (len: 2490)
+Tokenizing eval data (len: 277)
+Loaded data and tokenized in 13.395596742630005s
+Training model across 1 GPUs
+***** Running training *****
+	Num examples = 2490
+	Batch size = 8
+	Max sequence length = 128
+	Num steps = 1555
+	Num epochs = 5
+	Learning rate = 2e-05
+Eval accuracy: 68.23104693140795%
+Best acc found. Saved model to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/.
+Eval accuracy: 70.03610108303249%
+Best acc found. Saved model to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/.
+Eval accuracy: 72.56317689530685%
+Best acc found. Saved model to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/.
+Eval accuracy: 69.67509025270758%
+Eval accuracy: 69.31407942238266%
+Saved tokenizer <textattack.models.tokenizers.auto_tokenizer.AutoTokenizer object at 0x7fc688911c10> to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/.
+Wrote README to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/README.md.
+Wrote training args to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/train_args.json.

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:63733fb269828a14b20e371ee6ca3da62028da2831aaeaea00fe09df62e53b3a
+size 437985387

special_tokens_map.json

@@ -0,0 +1 @@
+{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"do_lower_case": true, "model_max_length": 512}

train_args.json

@@ -0,0 +1,25 @@
+{
+  "model": "bert-base-uncased",
+  "dataset": "glue:rte",
+  "dataset_train_split": "train",
+  "dataset_dev_split": "validation",
+  "tb_writer_step": 1000,
+  "checkpoint_steps": -1,
+  "checkpoint_every_epoch": false,
+  "num_train_epochs": 5,
+  "early_stopping_epochs": -1,
+  "batch_size": 8,
+  "max_length": 128,
+  "learning_rate": 2e-05,
+  "grad_accum_steps": 1,
+  "warmup_proportion": 0.1,
+  "config_name": "config.json",
+  "weights_name": "pytorch_model.bin",
+  "enable_wandb": false,
+  "output_dir": "/p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/",
+  "num_labels": 2,
+  "do_regression": false,
+  "best_eval_score": 0.7256317689530686,
+  "best_eval_score_epoch": 2,
+  "epochs_since_best_eval_score": 2
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24b2bb6c082efcc9b5a8c1dbf0f62d065ab5c72c4847cb8ebf42edd5efca94e7
+size 1051