Llama-360M

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

  • Loss: 5.3269

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: 0.0003
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 128
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 50
  • num_epochs: 40
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
9.6295 0.57 1 9.6320
9.4685 1.71 3 9.4277
8.7308 2.86 5 8.9834
7.7978 4.0 7 8.3652
7.4895 4.57 8 8.1048
6.9772 5.71 10 7.7260
6.6117 6.86 12 7.4107
6.2461 8.0 14 7.1384
6.0376 8.57 15 6.9993
5.6415 9.71 17 6.7886
5.3502 10.86 19 6.6009
5.0627 12.0 21 6.4227
4.9292 12.57 22 6.3169
4.5619 13.71 24 6.1217
4.1745 14.86 26 5.9089
3.895 16.0 28 5.7244
3.7108 16.57 29 5.6837
3.4811 17.71 31 5.5533
3.3174 18.86 33 5.4525
3.0011 20.0 35 5.4535
2.8812 20.57 36 5.4168
2.6512 21.71 38 5.4168
2.3009 22.86 40 5.3269

Framework versions

  • Transformers 4.39.1
  • Pytorch 2.1.2+cu121
  • Datasets 2.16.1
  • Tokenizers 0.15.0
Downloads last month
1
Safetensors
Model size
0.4B params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support