al-wmt14

This model is a fine-tuned version of google-t5/t5-small on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7427
• Bleu: 51.2627
• Gen Len: 26.713

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: 0.0005
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 100000

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
0.8503	0.2778	10000	0.9749	41.49	27.2803
0.7082	0.5556	20000	0.8905	44.7235	26.6963
0.6407	0.8333	30000	0.8530	46.6914	27.019
0.573	1.1111	40000	0.8260	47.4882	26.6827
0.5438	1.3889	50000	0.8017	48.472	26.8617
0.5263	1.6667	60000	0.7810	49.0812	26.8817
0.5091	1.9444	70000	0.7654	49.9355	26.7853
0.4699	2.2222	80000	0.7605	50.3601	26.72
0.4597	2.5	90000	0.7488	50.92	26.8803
0.454	2.7778	100000	0.7427	51.2627	26.713

Framework versions

• Transformers 4.40.0
• Pytorch 2.2.2+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1