You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this model content.

moe_e_hom

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

  • Loss: 5.2265

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: 0.0001
  • train_batch_size: 32
  • eval_batch_size: 32
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 2
  • gradient_accumulation_steps: 16
  • total_train_batch_size: 1024
  • total_eval_batch_size: 64
  • optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-06 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • training_steps: 20871
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
No log 0 0 11.0172
7.6961 0.4791 1000 6.8598
6.4686 0.9582 2000 5.7974
6.0292 1.4370 3000 5.4299
5.8161 1.9161 4000 5.2142
5.5642 2.3948 5000 5.0939
5.4886 2.8739 6000 4.9927
5.2586 3.3526 7000 4.9522
5.2387 3.8317 8000 4.8943
5.0007 4.3105 9000 4.9058
5.0116 4.7896 10000 4.8711
4.7478 5.2683 11000 4.9165
4.7846 5.7474 12000 4.9028
4.509 6.2261 13000 4.9728
4.5599 6.7053 14000 4.9730
4.2942 7.1840 15000 5.0498
4.347 7.6631 16000 5.0659
4.3628 8.0 16704 5.0655
4.1082 8.1418 17000 5.1341
4.1578 8.6209 18000 5.1577
3.9673 9.0997 19000 5.2038
3.9913 9.5788 20000 5.2233

Framework versions

  • Transformers 4.53.1
  • Pytorch 2.7.0+cu126
  • Datasets 3.6.0
  • Tokenizers 0.21.1
Downloads last month
-
Safetensors
Model size
0.5B params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Evaluation results