nllb-akkadian-normalized

This model is a fine-tuned version of facebook/nllb-200-distilled-600M on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 3.1974

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: 5e-05
  • train_batch_size: 4
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 32
  • 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: 3
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
45.2668 0.2097 250 5.0299
36.4053 0.4193 500 4.2545
32.7959 0.6290 750 3.8233
30.1804 0.8387 1000 3.5678
28.9688 1.0478 1250 3.4610
28.4473 1.2575 1500 3.3881
27.7909 1.4671 1750 3.3348
27.5623 1.6768 2000 3.2948
27.4065 1.8865 2250 3.2629
26.8964 2.0956 2500 3.2409
26.8113 2.3053 2750 3.2217
26.7997 2.5149 3000 3.2099
26.8521 2.7246 3250 3.2006
26.6507 2.9343 3500 3.1974

Framework versions

  • PEFT 0.18.1
  • Transformers 5.0.0
  • Pytorch 2.10.0+cu128
  • Datasets 4.0.0
  • Tokenizers 0.22.2
Downloads last month
77
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for NAQarabash/nllb-akkadian-normalized

Adapter
(69)
this model