t5-finetuned-ar-to-arsl2

This model is a fine-tuned version of PRAli22/arat5-arabic-dialects-translation on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3244
• Rouge1: 0.0
• Rouge2: 0.0
• Rougel: 0.0
• Rougelsum: 0.0
• Gen Len: 5.2892

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	0.99	78	0.4098	0.0	0.0	0.0	0.0	5.2132
No log	1.99	157	0.3022	0.0	0.0	0.0	0.0	5.248
No log	2.99	236	0.2753	0.0	0.0	0.0	0.0	5.2655
No log	3.99	315	0.2742	0.0	0.0	0.0	0.0	5.275
No log	5.0	394	0.2730	0.0	0.0	0.0	0.0	5.2845
No log	6.0	473	0.2821	0.0	0.0	0.0	0.0	5.2805
0.4176	7.0	552	0.2923	0.0	0.0	0.0	0.0	5.2868
0.4176	8.0	631	0.2977	0.0	0.0	0.0	0.0	5.2892
0.4176	8.99	709	0.2937	0.0	0.0	0.0	0.0	5.2964
0.4176	9.99	788	0.3020	0.0	0.0	0.0	0.0	5.2916
0.4176	10.99	867	0.3177	0.0	0.0	0.0	0.0	5.294
0.4176	11.99	946	0.3186	0.0	0.0	0.0	0.0	5.2916
0.0879	13.0	1025	0.3255	0.0	0.0	0.0	0.0	5.2884
0.0879	14.0	1104	0.3241	0.0	0.0	0.0	0.0	5.2868
0.0879	14.83	1170	0.3244	0.0	0.0	0.0	0.0	5.2892

Framework versions

• Transformers 4.39.3
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.15.2