distilbert-base-uncased-finetuned-squad-clos-v4

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

• Loss: 5.6212

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss
4.5833	1.0	1384	4.5272
4.4156	2.0	2768	4.4612
4.2998	3.0	4152	4.4308
4.1939	4.0	5536	4.4414
4.1197	5.0	6920	4.4658
4.0254	6.0	8304	4.4956
3.9493	7.0	9688	4.5488
3.8801	8.0	11072	4.5977
3.7998	9.0	12456	4.6510
3.741	10.0	13840	4.6884
3.6751	11.0	15224	4.7773
3.6182	12.0	16608	4.8803
3.5358	13.0	17992	4.8822
3.4968	14.0	19376	4.9722
3.4541	15.0	20760	5.0355
3.4186	16.0	22144	5.0545
3.3653	17.0	23528	5.1249
3.3134	18.0	24912	5.2119
3.2808	19.0	26296	5.2069
3.2592	20.0	27680	5.3077
3.2279	21.0	29064	5.3592
3.184	22.0	30448	5.4193
3.1597	23.0	31832	5.4387
3.1417	24.0	33216	5.4711
3.1307	25.0	34600	5.4601
3.098	26.0	35984	5.5668
3.0761	27.0	37368	5.5658
3.0785	28.0	38752	5.5651
3.0635	29.0	40136	5.6047
3.0508	30.0	41520	5.6212

Framework versions

• Transformers 4.42.3
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1