Modelpublic

pierreqi/peft-starcoder-qlora-r

sourceHugging Facebigcode-openrail-mupdated 2y agoView on Hugging Face

0likes5downloads

Model Card

peft-starcoder-qlora-r

This model is a fine-tuned version of bigcode/starcoderbase-1b on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 1.1974

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: 0.0005
—trainbatchsize: 16
—evalbatchsize: 16
—seed: 42
—optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: cosine
—lrschedulerwarmup_steps: 30
—training_steps: 2000

Training results

Training Loss	Epoch	Step	Validation Loss
1.0603	0.05	100	1.1678
2.2372	0.1	200	1.1616
1.0571	0.15	300	1.1632
1.2376	0.2	400	1.1639
1.1043	0.25	500	1.1664
1.0032	0.3	600	1.1703
1.168	0.35	700	1.1713
0.9762	0.4	800	1.1749
2.2509	0.45	900	1.1789
1.0519	0.5	1000	1.1873
2.2067	0.55	1100	1.1825
1.0048	0.6	1200	1.1884
1.9986	0.65	1300	1.1865
1.0117	0.7	1400	1.1950
1.0123	0.75	1500	1.1916
2.1783	0.8	1600	1.1949
0.9197	0.85	1700	1.1943
2.1317	0.9	1800	1.1942
0.8868	0.95	1900	1.1985
1.008	1.0	2000	1.1974

Framework versions

—PEFT 0.14.0
—Transformers 4.48.3
—Pytorch 2.5.1+cu124
—Datasets 3.3.0
—Tokenizers 0.21.0