8d44004441c73212e62d462d7fe611bd

This model is a fine-tuned version of distilbert/distilbert-base-multilingual-cased on the nyu-mll/glue [mnli] dataset. It achieves the following results on the evaluation set:

• Loss: 0.7350
• Data Size: 1.0
• Epoch Runtime: 367.1770
• Accuracy: 0.7627
• F1 Macro: 0.7617
• Rouge1: 0.7626
• Rouge2: 0.0
• Rougel: 0.7626
• Rougelsum: 0.7630

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_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: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Accuracy	F1 Macro	Rouge1	Rouge2	Rougel	Rougelsum
No log	0	0	1.0969	0	3.0442	0.3526	0.1977	0.3525	0.0	0.3524	0.3525
1.0574	1	12271	0.9180	0.0078	6.3581	0.5770	0.5763	0.5769	0.0	0.5771	0.5774
0.9141	2	24542	0.8569	0.0156	9.1755	0.6140	0.6078	0.6143	0.0	0.6142	0.6143
0.8319	3	36813	0.8069	0.0312	14.8050	0.6450	0.6409	0.6451	0.0	0.6449	0.6449
0.8084	4	49084	0.7502	0.0625	26.0503	0.6713	0.6703	0.6712	0.0	0.6713	0.6714
0.7106	5	61355	0.7001	0.125	49.2368	0.7032	0.7022	0.7033	0.0	0.7032	0.7035
0.6739	6	73626	0.6723	0.25	92.9763	0.7158	0.7162	0.7160	0.0	0.7159	0.7157
0.5909	7	85897	0.6326	0.5	181.1791	0.7384	0.7380	0.7384	0.0	0.7384	0.7386
0.564	8.0	98168	0.6000	1.0	364.3526	0.7570	0.7567	0.7571	0.0	0.7570	0.7570
0.4927	9.0	110439	0.6232	1.0	359.4073	0.7602	0.7590	0.7602	0.0	0.7605	0.7604
0.4562	10.0	122710	0.6124	1.0	368.9395	0.7632	0.7626	0.7631	0.0	0.7633	0.7634
0.3814	11.0	134981	0.7018	1.0	372.2864	0.7623	0.7623	0.7622	0.0	0.7622	0.7627
0.3027	12.0	147252	0.7350	1.0	367.1770	0.7627	0.7617	0.7626	0.0	0.7626	0.7630

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1