e14dfac9d43f01218619eac3fca7ca9c

This model is a fine-tuned version of google-t5/t5-base on the Helsinki-NLP/opus_books [de-fr] dataset. It achieves the following results on the evaluation set:

• Loss: 1.4196
• Data Size: 1.0
• Epoch Runtime: 205.8362
• Bleu: 9.3789

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	2.3367	0	16.6961	4.0341
No log	1	872	2.2019	0.0078	17.1484	4.6135
No log	2	1744	2.1132	0.0156	18.7268	4.8180
0.0377	3	2616	2.0425	0.0312	21.9222	4.8406
0.1431	4	3488	1.9730	0.0625	28.4409	5.2560
2.1816	5	4360	1.9060	0.125	40.2414	5.1929
2.0629	6	5232	1.8339	0.25	62.2581	5.8377
1.9233	7	6104	1.7445	0.5	109.0056	6.3272
1.8188	8.0	6976	1.6562	1.0	203.7760	6.9404
1.7321	9.0	7848	1.6029	1.0	205.9968	7.5055
1.6903	10.0	8720	1.5617	1.0	215.5350	7.8362
1.626	11.0	9592	1.5304	1.0	211.5992	7.9350
1.5501	12.0	10464	1.5096	1.0	217.1961	8.2531
1.5287	13.0	11336	1.4904	1.0	216.2585	8.4558
1.4598	14.0	12208	1.4758	1.0	210.5595	8.5746
1.4347	15.0	13080	1.4656	1.0	203.0477	8.5464
1.4006	16.0	13952	1.4533	1.0	198.8969	8.9832
1.3355	17.0	14824	1.4463	1.0	206.0515	9.0535
1.3292	18.0	15696	1.4374	1.0	205.1314	8.9567
1.3087	19.0	16568	1.4313	1.0	212.4399	9.1311
1.2654	20.0	17440	1.4274	1.0	200.9783	9.2130
1.2788	21.0	18312	1.4221	1.0	202.8125	9.2160
1.241	22.0	19184	1.4193	1.0	208.9738	9.3522
1.1823	23.0	20056	1.4198	1.0	207.6572	9.2450
1.1673	24.0	20928	1.4205	1.0	202.4762	9.2619
1.1537	25.0	21800	1.4252	1.0	211.7916	9.3895
1.1538	26.0	22672	1.4196	1.0	205.8362	9.3789

Framework versions

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