bigbench_entailedpolarity-bert-base-uncased

This model is a fine-tuned version of bert-base-uncased on the bigbench dataset. It achieves the following results on the evaluation set:

• Loss: 0.3706
• Accuracy: 0.9583

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 4
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	24	0.5613	0.7917
No log	2.0	48	0.5337	0.7917
No log	3.0	72	0.3178	0.875
No log	4.0	96	0.1924	0.9167
No log	5.0	120	0.2007	0.9583
No log	6.0	144	0.2766	0.9583
No log	7.0	168	0.3031	0.9583
No log	8.0	192	0.3291	0.9583
No log	9.0	216	0.3381	0.9583
No log	10.0	240	0.3454	0.9583
No log	11.0	264	0.3508	0.9583
No log	12.0	288	0.3554	0.9583
No log	13.0	312	0.3592	0.9583
No log	14.0	336	0.3624	0.9583
No log	15.0	360	0.3650	0.9583
No log	16.0	384	0.3671	0.9583
No log	17.0	408	0.3686	0.9583
No log	18.0	432	0.3696	0.9583
No log	19.0	456	0.3704	0.9583
No log	20.0	480	0.3706	0.9583

Framework versions

• Transformers 4.35.2
• Pytorch 1.10.1+cu102
• Datasets 2.15.0
• Tokenizers 0.15.0