small-vanilla-target-glue-mrpc-linear-probe

This model is a fine-tuned version of google/bert_uncased_L-4_H-512_A-8 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5860
• Accuracy: 0.7010
• F1: 0.8174

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: 3e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• training_steps: 5000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.6358	4.35	500	0.6136	0.6838	0.8111
0.6123	8.7	1000	0.6068	0.6863	0.8129
0.6054	13.04	1500	0.5990	0.6838	0.8095
0.6008	17.39	2000	0.5962	0.6912	0.8136
0.595	21.74	2500	0.5925	0.7059	0.8209
0.5916	26.09	3000	0.5898	0.7034	0.8191
0.5885	30.43	3500	0.5906	0.7010	0.8185
0.5915	34.78	4000	0.5860	0.7010	0.8174

Framework versions

• Transformers 4.26.0.dev0
• Pytorch 1.13.0+cu116
• Datasets 2.8.1.dev0
• Tokenizers 0.13.2