You can deploy this model and start earning money today!

KnoxDevelopers/
English-to-Meru-language-translation-LoRA

Name: English-to-Meru-language-translation-LoRA
Availability: InStock

Views

No views yet

English-to-Meru-language-translation-LoRA

This model is a fine-tuned version of facebook/nllb-200-distilled-600M on the English-Meru dataset. It achieves the following results on the evaluation set:

A LoRA adapter for translating text from English to Kimeru.

More information needed

The following hyperparameters were used during training:

learning_rate: 0.0002
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 16
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 7
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss
5.2921	1.0	1742	2.3882
4.6466	2.0	3484	2.1125
4.3495	3.0	5226	1.9624
4.1860	4.0	6968	1.8715
4.0373	5.0	8710	1.8175
3.8794	6.0	10452	1.7889
3.9127	7.0	12194	1.7783

Metric	Results	Explained
sacrebleu	7.824	Evaluation matches exact word sequence referencing our ground truth and adapter output.
		0-20= translation is literal or broken, 20-50=good, >60=very good or identical training and test data.
chrF++	37.682	Eval matches character sequence instead of whole words.
		Model adapter is getting the root words and grammar concepts right but using different word variations or spellings than our ground truth, this is concluded due to the low sacrebleu results.