powerinfer-seq-cls-ywko_e5

This model is a fine-tuned version of PowerInfer/SmallThinker-3B-Preview on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0727

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: 64
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 64
• total_train_batch_size: 4096
• 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: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.4779	0.9143	30	0.1895
0.1332	1.8533	60	0.1015
0.0797	2.7924	90	0.0780
0.0811	3.7314	120	0.0743
0.0669	4.6705	150	0.0727

Framework versions

• Transformers 4.50.3
• Pytorch 2.5.1+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1