recipe-distilbert-tis

This model is a fine-tuned version of distilbert-base-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.9886

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: 2e-05
• train_batch_size: 256
• eval_batch_size: 256
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.8792	1.0	1038	1.4680
1.4691	2.0	2076	1.3122
1.3471	3.0	3114	1.2343
1.2798	4.0	4152	1.1829
1.2284	5.0	5190	1.1437
1.1956	6.0	6228	1.1120
1.1683	7.0	7266	1.0905
1.1458	8.0	8304	1.0731
1.126	9.0	9342	1.0645
1.113	10.0	10380	1.0471
1.0962	11.0	11418	1.0348
1.0864	12.0	12456	1.0256
1.0772	13.0	13494	1.0173
1.0692	14.0	14532	1.0130
1.0622	15.0	15570	1.0111
1.0577	16.0	16608	1.0021
1.0537	17.0	17646	1.0001
1.0474	18.0	18684	0.9928
1.0471	19.0	19722	0.9908
1.0458	20.0	20760	0.9886

Framework versions

• Transformers 4.19.0.dev0
• Pytorch 1.11.0+cu102
• Datasets 2.3.2
• Tokenizers 0.12.1