mobilebert_add_GLUE_Experiment_logit_kd_mrpc_128

This model is a fine-tuned version of google/mobilebert-uncased on the GLUE MRPC dataset. It achieves the following results on the evaluation set:

• Loss: 0.5534
• Accuracy: 0.6838
• F1: 0.8122
• Combined Score: 0.7480

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: 128
• eval_batch_size: 128
• seed: 10
• distributed_type: multi-GPU
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Combined Score
0.6399	1.0	29	0.5562	0.6838	0.8122	0.7480
0.6101	2.0	58	0.5559	0.6838	0.8122	0.7480
0.6111	3.0	87	0.5557	0.6838	0.8122	0.7480
0.6104	4.0	116	0.5572	0.6838	0.8122	0.7480
0.6086	5.0	145	0.5550	0.6838	0.8122	0.7480
0.6058	6.0	174	0.5534	0.6838	0.8122	0.7480
0.6036	7.0	203	0.5745	0.6838	0.8122	0.7480
0.5969	8.0	232	0.5595	0.6838	0.8122	0.7480
0.5735	9.0	261	0.5699	0.6838	0.8122	0.7480
0.5597	10.0	290	0.5608	0.6838	0.8122	0.7480
0.5456	11.0	319	0.5714	0.6838	0.8122	0.7480

Framework versions

• Transformers 4.26.0
• Pytorch 1.14.0a0+410ce96
• Datasets 2.9.0
• Tokenizers 0.13.2