Modelpublic

Donny-Guo/distrilbert_full_parameter_finetune_crisismmd

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face

0likes2downloads

Model Card

distrilbertfullparameterfinetunecrisismmd

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

—Loss: 0.3472
—Accuracy: 84.87%
—Precision: 84.73%
—Recall: 84.87%
—F1: 84.42%

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
—trainbatchsize: 8
—evalbatchsize: 8
—seed: 42
—optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.5759	0.0482	50	0.4439	80.61%	80.68%	80.61%	80.64%
0.4797	0.0964	100	0.4242	81.50%	81.28%	81.50%	80.67%
0.4362	0.1446	150	0.4136	82.77%	82.97%	82.77%	82.86%
0.4631	0.1929	200	0.4900	81.25%	82.43%	81.25%	79.54%
0.4326	0.2411	250	0.4333	83.28%	83.86%	83.28%	82.20%
0.3911	0.2893	300	0.4526	82.45%	83.21%	82.45%	81.17%
0.4426	0.3375	350	0.4395	83.22%	84.14%	83.22%	83.48%
0.4022	0.3857	400	0.3950	83.47%	83.21%	83.47%	83.25%
0.3937	0.4339	450	0.3667	84.55%	84.36%	84.55%	84.42%
0.3773	0.4822	500	0.3843	84.74%	84.53%	84.74%	84.38%
0.451	0.5304	550	0.3858	82.52%	83.49%	82.52%	81.15%
0.3488	0.5786	600	0.3840	84.68%	84.54%	84.68%	84.21%
0.3773	0.6268	650	0.3851	84.23%	84.27%	84.23%	83.57%
0.3442	0.6750	700	0.3492	85.31%	85.10%	85.31%	85.08%
0.3348	0.7232	750	0.4122	84.11%	84.64%	84.11%	83.17%
0.4304	0.7715	800	0.3499	84.81%	84.80%	84.81%	84.24%
0.3724	0.8197	850	0.3569	84.87%	84.81%	84.87%	84.34%
0.3212	0.8679	900	0.3706	84.62%	84.86%	84.62%	83.88%
0.4108	0.9161	950	0.3835	84.23%	84.65%	84.23%	83.36%
0.3334	0.9643	1000	0.3472	84.87%	84.73%	84.87%	84.42%

Framework versions

—Transformers 4.50.3
—Pytorch 2.6.0+cu124
—Datasets 3.5.0
—Tokenizers 0.21.1