Terjman-Large-v2

This model is a fine-tuned version of Helsinki-NLP/opus-mt-tc-big-en-ar on the None dataset. It achieves the following results on the evaluation set:

• Loss: 5.9483
• Bleu: 0.0891
• Gen Len: 511.0

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
12.4538	0.1449	1000	6.3599	0.0191	511.0
12.5456	0.2899	2000	6.7320	0.0026	499.9054
12.0592	0.4348	3000	6.2157	0.0097	511.0
11.8591	0.5798	4000	6.1470	0.0269	511.0
11.776	0.7247	5000	6.0090	0.0469	511.0
11.3534	0.8696	6000	5.9483	0.0891	511.0

Framework versions

• Transformers 4.48.0.dev0
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.21.0