5b3ca441014dd534338575e20fa7f09d

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

• Loss: 1.9342
• Data Size: 1.0
• Epoch Runtime: 18.8284
• Bleu: 8.8303

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	17.4501	0	2.2271	0.0055
No log	1	77	17.0538	0.0078	2.3920	0.0053
No log	2	154	16.7341	0.0156	2.9546	0.0058
No log	3	231	16.2059	0.0312	3.7517	0.0065
No log	4	308	14.5850	0.0625	4.1794	0.0031
No log	5	385	12.9984	0.125	6.5324	0.0057
1.4625	6	462	10.0731	0.25	8.9588	0.0042
4.6161	7	539	6.6948	0.5	13.1154	0.0066
5.8813	8.0	616	2.9076	1.0	21.2510	0.8848
4.0464	9.0	693	2.4169	1.0	18.8442	4.5801
3.2073	10.0	770	2.2892	1.0	19.0267	5.1261
3.0516	11.0	847	2.2187	1.0	18.8727	5.5731
2.8243	12.0	924	2.1611	1.0	18.5039	6.1641
2.655	13.0	1001	2.1078	1.0	19.3308	6.9869
2.577	14.0	1078	2.0727	1.0	19.5830	7.1556
2.476	15.0	1155	2.0449	1.0	20.3377	7.3569
2.4089	16.0	1232	2.0242	1.0	20.6235	7.5293
2.3497	17.0	1309	2.0138	1.0	19.2595	7.6576
2.2899	18.0	1386	2.0005	1.0	20.2983	7.8332
2.2052	19.0	1463	1.9870	1.0	21.0967	7.9906
2.1891	20.0	1540	1.9979	1.0	20.7798	7.8306
2.1244	21.0	1617	1.9645	1.0	20.9778	8.0652
2.087	22.0	1694	1.9574	1.0	19.0645	8.0635
2.0356	23.0	1771	1.9536	1.0	19.2787	8.2359
1.9882	24.0	1848	1.9505	1.0	19.5822	8.3249
1.9266	25.0	1925	1.9414	1.0	20.2196	8.3972
1.8875	26.0	2002	1.9496	1.0	20.5014	8.3262
1.859	27.0	2079	1.9328	1.0	18.8038	8.4741
1.8142	28.0	2156	1.9398	1.0	19.6716	8.4424
1.7913	29.0	2233	1.9282	1.0	19.1268	8.6150
1.7503	30.0	2310	1.9331	1.0	19.9942	8.4541
1.7261	31.0	2387	1.9364	1.0	19.1797	8.5984
1.6748	32.0	2464	1.9290	1.0	19.1266	8.8069
1.6499	33.0	2541	1.9342	1.0	18.8284	8.8303

Framework versions

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