README.md

metadata

library_name: transformers
base_model: huawei-noah/TinyBERT_General_4L_312D
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - precision
  - recall
model-index:
  - name: Structured-NF4-KD-NID
    results: []

Structured-NF4-KD-NID

This model is a fine-tuned version of huawei-noah/TinyBERT_General_4L_312D on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0384
• Accuracy: 0.9923
• Precision: 0.9796
• Recall: 0.9374
• F1 score: 0.9470

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: 650
• eval_batch_size: 650
• 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
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1 score
0.217	1.0	1828	0.1859	0.9775	0.7808	0.7811	0.7797
0.1295	2.0	3656	0.1142	0.9844	0.9142	0.8421	0.8404
0.0958	3.0	5484	0.0916	0.9864	0.9136	0.8662	0.8699
0.0846	4.0	7312	0.0779	0.9882	0.9207	0.8902	0.8953
0.0683	5.0	9140	0.0677	0.9894	0.9246	0.9034	0.9073
0.07	6.0	10968	0.0599	0.9899	0.9259	0.9103	0.9118
0.0574	7.0	12796	0.0555	0.9904	0.9284	0.9097	0.9156
0.0503	8.0	14624	0.0541	0.9906	0.9211	0.9185	0.9162
0.048	9.0	16452	0.0498	0.9911	0.9318	0.9185	0.9217
0.0479	10.0	18280	0.0483	0.9912	0.9353	0.9178	0.9232
0.0452	11.0	20108	0.0467	0.9914	0.9264	0.9233	0.9213
0.042	12.0	21936	0.0437	0.9917	0.9331	0.9211	0.9244
0.038	13.0	23764	0.0426	0.9918	0.9782	0.9282	0.9366
0.0358	14.0	25592	0.0414	0.9919	0.9359	0.9226	0.9253
0.0351	15.0	27420	0.0417	0.9920	0.9785	0.9288	0.9370
0.0319	16.0	29248	0.0404	0.9921	0.9768	0.9437	0.9518
0.0302	17.0	31076	0.0391	0.9922	0.9789	0.9319	0.9386
0.0319	18.0	32904	0.0383	0.9922	0.9801	0.9428	0.9533
0.0265	19.0	34732	0.0383	0.9923	0.9802	0.9436	0.9542
0.0277	20.0	36560	0.0384	0.9923	0.9796	0.9374	0.9470

Framework versions

• Transformers 4.50.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1