cds_sh1_fr_30

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

• Loss: 3.3099

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: 0.0001
• train_batch_size: 256
• eval_batch_size: 256
• seed: 30
• optimizer: Use adamw_torch_fused with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
6.0017	1.0	145	4.5165
4.1747	2.0	290	4.0281
3.8777	3.0	435	3.7704
3.6723	4.0	580	3.6043
3.5366	5.0	725	3.5114
3.4452	6.0	870	3.4480
3.3743	7.0	1015	3.4009
3.3153	8.0	1160	3.3667
3.2657	9.0	1305	3.3414
3.2193	10.0	1450	3.3189
3.1773	11.0	1595	3.3084
3.1366	12.0	1740	3.2952
3.0985	13.0	1885	3.2936
3.061	14.0	2030	3.2873
3.0252	15.0	2175	3.2956
2.9897	16.0	2320	3.2938
2.9574	17.0	2465	3.2979
2.9302	18.0	2610	3.3015
2.9063	19.0	2755	3.3071
2.8895	20.0	2900	3.3099

Framework versions

• Transformers 4.56.1
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.0