m2m100_418M_pt_informal

This model is a fine-tuned version of facebook/m2m100_418M on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2567
• Bleu: 33.4128

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: 32
• eval_batch_size: 8
• seed: 42
• 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_steps: 500
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu
4.3504	0.4149	500	0.3221	29.1340
0.3117	0.8299	1000	0.2714	31.5481
0.2688	1.2448	1500	0.2658	32.3248
0.259	1.6598	2000	0.2612	32.9338
0.2528	2.0747	2500	0.2606	32.9240
0.2311	2.4896	3000	0.2573	33.2674
0.2298	2.9046	3500	0.2567	33.4128

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0