Modelpublic

juancopi81/js-fake-bach-epochs50

sourceHugging Facemitupdated 2y agoView on Hugging Face

0likes104downloads

Model Card

js-fake-bach-epochs50

This model is a fine-tuned version of gpt2 on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 0.9888
—Accuracy: 0.0005

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: 0.0006058454513356471
—trainbatchsize: 16
—evalbatchsize: 32
—seed: 1
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: cosine
—lrschedulerwarmup_ratio: 0.01
—num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.3512	1.25	315	0.8371	0.0003
0.8149	2.51	630	0.7684	0.0006
0.7601	3.76	945	0.7187	0.0004
0.7186	5.02	1260	0.6903	0.0002
0.679	6.27	1575	0.6563	0.0005
0.6419	7.53	1890	0.6292	0.0001
0.6073	8.78	2205	0.5949	0.0006
0.575	10.04	2520	0.5828	0.0001
0.5425	11.29	2835	0.5696	0.0003
0.5174	12.55	3150	0.5609	0.0007
0.4933	13.8	3465	0.5576	0.0004
0.4696	15.06	3780	0.5661	0.0002
0.4423	16.31	4095	0.5708	0.0007
0.4196	17.57	4410	0.5780	0.0006
0.398	18.82	4725	0.5820	0.0009
0.374	20.08	5040	0.6099	0.0003
0.3452	21.33	5355	0.6230	0.0006
0.3256	22.59	5670	0.6386	0.0005
0.3047	23.84	5985	0.6462	0.0003
0.2812	25.1	6300	0.6789	0.0003
0.2582	26.35	6615	0.7053	0.0007
0.2406	27.61	6930	0.7199	0.0006
0.2237	28.86	7245	0.7399	0.0006
0.204	30.12	7560	0.7729	0.0006
0.1873	31.37	7875	0.7960	0.0004
0.1725	32.63	8190	0.8231	0.0005
0.1609	33.88	8505	0.8493	0.0004
0.1479	35.14	8820	0.8707	0.0003
0.1361	36.39	9135	0.8931	0.0003
0.1273	37.65	9450	0.9095	0.0003
0.12	38.9	9765	0.9339	0.0005
0.1129	40.16	10080	0.9444	0.0004
0.1062	41.41	10395	0.9626	0.0006
0.1027	42.67	10710	0.9669	0.0006
0.0994	43.92	11025	0.9713	0.0005
0.0955	45.18	11340	0.9830	0.0005
0.0939	46.43	11655	0.9855	0.0005
0.0924	47.69	11970	0.9884	0.0005
0.0916	48.94	12285	0.9888	0.0005

Framework versions

—Transformers 4.29.1
—Pytorch 2.0.0+cu118
—Datasets 2.12.0
—Tokenizers 0.13.3