demo_LID_ntu-spml_distilhubert

This model is a fine-tuned version of ntu-spml/distilhubert on the common_language dataset. It achieves the following results on the evaluation set:

• Loss: 2.2545
• Accuracy: 0.6554

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: 0.0003
• train_batch_size: 8
• eval_batch_size: 1
• seed: 0
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
9.6557	0.9989	693	2.6549	0.2614
6.1707	1.9989	1386	1.8478	0.4681
3.7871	2.9989	2079	1.6941	0.5474
2.7966	3.9989	2772	1.8580	0.5579
1.5871	4.9989	3465	1.6663	0.6140
0.7355	5.9989	4158	1.9491	0.6155
0.4492	6.9989	4851	2.0594	0.6379
0.1528	7.9989	5544	2.1739	0.6403
0.0468	8.9989	6237	2.3125	0.6505
0.0045	9.9989	6930	2.2545	0.6554

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0