95abbf185563e55eb1f0cd8f2e44da1c

This model is a fine-tuned version of google/mt5-large on the Helsinki-NLP/opus_books [en-fr] dataset. It achieves the following results on the evaluation set:

• Loss: 1.0984
• Data Size: 1.0
• Epoch Runtime: 1267.8508
• Bleu: 13.8876

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	23.9495	0	93.2618	0.0191
No log	1	3177	8.1473	0.0078	104.2668	0.0240
0.2377	2	6354	2.6256	0.0156	115.5180	0.5802
2.788	3	9531	1.8819	0.0312	137.5417	1.2068
2.2383	4	12708	1.6328	0.0625	174.4671	6.9639
1.974	5	15885	1.5148	0.125	248.4360	8.3588
1.7815	6	19062	1.4145	0.25	390.9339	9.9341
1.6135	7	22239	1.3232	0.5	681.1867	10.8747
1.4783	8.0	25416	1.2319	1.0	1268.3256	11.9094
1.3401	9.0	28593	1.1818	1.0	1263.0211	12.2181
1.2957	10.0	31770	1.1490	1.0	1265.3100	12.6773
1.2081	11.0	34947	1.1215	1.0	1260.4574	12.9597
1.1387	12.0	38124	1.1098	1.0	1257.1675	13.1244
1.0896	13.0	41301	1.0948	1.0	1260.0131	13.2040
1.0435	14.0	44478	1.0902	1.0	1260.3279	13.4506
1.0154	15.0	47655	1.0821	1.0	1258.6829	13.5941
0.9567	16.0	50832	1.0839	1.0	1263.6322	13.5352
0.9349	17.0	54009	1.0857	1.0	1264.0652	13.6297
0.9039	18.0	57186	1.0766	1.0	1264.8995	13.8477
0.8632	19.0	60363	1.0834	1.0	1259.8701	13.8033
0.8297	20.0	63540	1.0864	1.0	1257.9190	13.7898
0.7992	21.0	66717	1.0951	1.0	1257.7156	13.9414
0.7704	22.0	69894	1.0984	1.0	1267.8508	13.8876

Framework versions

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