distilroberta-base-ep20

This model is a fine-tuned version of distilroberta-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.2289

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: 1.9370859e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.17096
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.7361	1.0	14644	1.5825
1.6238	2.0	29288	1.4802
1.559	3.0	43932	1.4250
1.5033	4.0	58576	1.3855
1.4776	5.0	73220	1.3575
1.4485	6.0	87864	1.3329
1.4208	7.0	102508	1.3127
1.411	8.0	117152	1.2991
1.395	9.0	131796	1.2896
1.3923	10.0	146440	1.2799
1.3663	11.0	161084	1.2719
1.3592	12.0	175728	1.2630
1.3518	13.0	190372	1.2561
1.3478	14.0	205016	1.2511
1.3426	15.0	219660	1.2466
1.3296	16.0	234304	1.2389
1.3231	17.0	248948	1.2354
1.3305	18.0	263592	1.2343
1.3198	19.0	278236	1.2330
1.3177	20.0	292880	1.2300

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0