pritamdeka/mt5-base-en-bgc-2

tst-translation

This model is a fine-tuned version of google/mt5-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.1206
• Bleu: 7.9168
• Gen Len: 51.4945

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: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20.0

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
4.5126	1.5748	200	2.7474	1.0891	120.8315
2.4414	3.1496	400	2.5120	5.0172	54.7623
1.724	4.7244	600	2.4090	5.2115	61.985
1.2755	6.2992	800	2.5519	7.0203	46.1414
0.97	7.8740	1000	2.5975	7.005	56.5266
0.7251	9.4488	1200	2.7918	7.6235	52.6841
0.584	11.0236	1400	2.8952	7.3273	49.9659
0.4358	12.5984	1600	3.1206	7.9168	51.4945
0.3619	14.1732	1800	3.2645	7.9096	50.5517
0.2933	15.7480	2000	3.3971	7.9015	49.6169
0.2447	17.3228	2200	3.5039	7.8441	49.6911
0.2151	18.8976	2400	3.5556	7.8198	50.1153

Framework versions

• Transformers 4.43.0.dev0
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1