En-Nso

This model is a fine-tuned version of kabelomalapane/en_nso_ukuxhumana_model on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.9067
• Bleu: 23.5436

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: 2e-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu
No log	1.0	14	3.7614	8.0360
No log	2.0	28	3.3181	20.7201
No log	3.0	42	3.1627	21.5932
No log	4.0	56	3.0935	22.0268
No log	5.0	70	3.0227	21.0859
No log	6.0	84	2.9740	21.6963
No log	7.0	98	2.9419	23.2214
No log	8.0	112	2.9227	24.4649
No log	9.0	126	2.9102	23.5293
No log	10.0	140	2.9067	23.5516

Framework versions

• Transformers 4.20.1
• Pytorch 1.11.0+cu113
• Datasets 2.3.2
• Tokenizers 0.12.1