tiny_bert_km_50_v1_mnli

This model is a fine-tuned version of Hartunka/tiny_bert_km_50_v1 on the GLUE MNLI dataset. It achieves the following results on the evaluation set:

• Loss: 0.8205
• Accuracy: 0.6414

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: 256
• eval_batch_size: 256
• seed: 10
• 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
1.0082	1.0	1534	0.9426	0.5438
0.9147	2.0	3068	0.8890	0.5845
0.8604	3.0	4602	0.8646	0.6063
0.8145	4.0	6136	0.8447	0.6156
0.7652	5.0	7670	0.8321	0.6299
0.7172	6.0	9204	0.8490	0.6406
0.6722	7.0	10738	0.8362	0.6512
0.6289	8.0	12272	0.8635	0.6460
0.5887	9.0	13806	0.9014	0.6464
0.5489	10.0	15340	0.9343	0.6449

Framework versions

• Transformers 4.40.0
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.19.1