e227683f38876549100b7a5ec987ce9a

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

• Loss: 1.1320
• Data Size: 1.0
• Epoch Runtime: 246.5735
• Bleu: 8.1993

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	3.7715	0	18.1804	0.5469
No log	1	1000	3.4010	0.0078	19.7679	0.8079
No log	2	2000	3.1686	0.0156	22.0671	1.1143
No log	3	3000	2.9896	0.0312	25.4138	1.0735
0.116	4	4000	2.8012	0.0625	32.0606	1.2803
2.9647	5	5000	2.6079	0.125	46.2747	1.6995
0.1672	6	6000	2.3794	0.25	74.7671	2.3617
0.2171	7	7000	2.1321	0.5	124.6736	3.0751
2.0913	8.0	8000	1.8615	1.0	229.3251	4.0645
1.9139	9.0	9000	1.7106	1.0	246.3512	4.7160
1.7825	10.0	10000	1.6044	1.0	234.0660	5.2285
1.6874	11.0	11000	1.5273	1.0	257.2715	5.6300
1.6142	12.0	12000	1.4681	1.0	233.0404	5.9439
1.5426	13.0	13000	1.4208	1.0	233.1306	6.1486
1.4938	14.0	14000	1.3815	1.0	233.9394	6.3779
1.4244	15.0	15000	1.3555	1.0	230.5803	6.6075
1.3899	16.0	16000	1.3250	1.0	232.0901	6.7548
1.3261	17.0	17000	1.3030	1.0	236.8272	6.8750
1.318	18.0	18000	1.2778	1.0	242.0059	7.0348
1.2823	19.0	19000	1.2580	1.0	225.9980	7.1927
1.2327	20.0	20000	1.2459	1.0	225.6653	7.2674
1.221	21.0	21000	1.2248	1.0	229.3647	7.3795
1.1973	22.0	22000	1.2157	1.0	226.9892	7.5226
1.1598	23.0	23000	1.2060	1.0	233.4547	7.5880
1.1326	24.0	24000	1.1892	1.0	233.8419	7.6334
1.1368	25.0	25000	1.1863	1.0	231.8198	7.7023
1.0797	26.0	26000	1.1801	1.0	231.9607	7.7432
1.0741	27.0	27000	1.1717	1.0	231.4938	7.8516
1.0421	28.0	28000	1.1633	1.0	236.6683	7.8662
1.0371	29.0	29000	1.1614	1.0	229.6369	7.9111
1.0067	30.0	30000	1.1589	1.0	235.2480	7.8860
1.0046	31.0	31000	1.1500	1.0	238.6989	7.9480
0.9796	32.0	32000	1.1466	1.0	231.9186	8.0116
0.9428	33.0	33000	1.1442	1.0	240.4297	8.0983
0.9312	34.0	34000	1.1413	1.0	243.4610	8.0653
0.9176	35.0	35000	1.1344	1.0	246.0070	8.1290
0.9046	36.0	36000	1.1382	1.0	238.9377	8.1121
0.8904	37.0	37000	1.1300	1.0	246.0509	8.1359
0.8863	38.0	38000	1.1341	1.0	240.4348	8.2015
0.8637	39.0	39000	1.1340	1.0	242.6820	8.1498
0.8548	40.0	40000	1.1349	1.0	240.7980	8.2200
0.8398	41.0	41000	1.1320	1.0	246.5735	8.1993

Framework versions

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