492afd7004f78277b6d22c05796c196e

This model is a fine-tuned version of google/umt5-xl on the Helsinki-NLP/opus_books [en-fr] dataset. It achieves the following results on the evaluation set:

• Loss: 1.1946
• Data Size: 1.0
• Epoch Runtime: 1573.6064
• Bleu: 15.9065

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	Bleu
No log	0	0	4.9540	0	109.0665	2.6147
No log	1	3177	2.5131	0.0078	120.3525	17.7039
0.0599	2	6354	1.8654	0.0156	142.3404	24.2251
2.0544	3	9531	1.5550	0.0312	166.5225	10.5876
1.815	4	12708	1.4500	0.0625	209.0970	11.4431
1.6555	5	15885	1.3777	0.125	304.3525	12.2800
1.5215	6	19062	1.2914	0.25	487.2187	13.1397
1.3872	7	22239	1.2222	0.5	852.6690	14.0849
1.254	8.0	25416	1.1497	1.0	1578.2169	14.9966
1.1105	9.0	28593	1.1241	1.0	1574.6522	15.4534
1.0375	10.0	31770	1.1162	1.0	1577.5926	15.6800
0.9277	11.0	34947	1.1180	1.0	1581.7450	15.7313
0.8186	12.0	38124	1.1412	1.0	1586.2745	15.8256
0.7382	13.0	41301	1.1606	1.0	1581.3892	15.8825
0.6694	14.0	44478	1.1946	1.0	1573.6064	15.9065

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1