d20536cf8a4db9bb7935c120ee347d92

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

• Loss: 1.7724
• Data Size: 1.0
• Epoch Runtime: 21.1339
• Bleu: 5.7456

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	3.8673	0	2.1007	0.7957
No log	1	29	3.7587	0.0078	2.3656	0.8667
No log	2	58	3.3359	0.0156	3.9018	1.2813
No log	3	87	3.1482	0.0312	5.7201	1.4520
No log	4	116	2.9528	0.0625	6.9152	1.6262
No log	5	145	2.5550	0.125	9.0873	3.0395
0.2895	6	174	2.2419	0.25	12.9129	3.6679
0.2895	7	203	2.1014	0.5	16.0726	3.4311
0.2895	8.0	232	1.9688	1.0	22.7370	3.8893
1.4182	9.0	261	1.8872	1.0	21.3748	4.9628
1.4182	10.0	290	1.8393	1.0	20.9844	4.8081
1.8728	11.0	319	1.8147	1.0	20.7618	5.0388
1.8728	12.0	348	1.7891	1.0	20.4942	5.2640
1.7288	13.0	377	1.7681	1.0	20.2467	5.3212
1.5867	14.0	406	1.7630	1.0	21.3503	5.4210
1.5867	15.0	435	1.7548	1.0	20.7495	5.5020
1.4762	16.0	464	1.7474	1.0	19.7304	5.5773
1.4762	17.0	493	1.7562	1.0	21.5750	5.6653
1.3933	18.0	522	1.7485	1.0	20.7114	5.7272
1.3097	19.0	551	1.7612	1.0	20.4159	5.6989
1.3097	20.0	580	1.7724	1.0	21.1339	5.7456

Framework versions

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