Modelpublic

Heejindo/rationale_model_e10_save5000_eos

sourceHugging Facellama3.2updated 2y agoView on Hugging Face

0likes18downloads

Model Card

rationalemodele10save5000eos

This model is a fine-tuned version of meta-llama/Llama-3.2-1B on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 2.0876

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: 1e-05
—trainbatchsize: 4
—evalbatchsize: 4
—seed: 42
—optimizer: Use adamwtorch with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss
1.0611	0.4769	5000	2.0876
0.3274	0.9538	10000	2.7190
0.1866	1.4308	15000	2.9944
0.1541	1.9077	20000	3.1986
0.1317	2.3846	25000	3.2512
0.1239	2.8615	30000	3.4157
0.1082	3.3384	35000	3.4850
0.1058	3.8153	40000	3.5546
0.0903	4.2923	45000	3.6930
0.0888	4.7692	50000	3.7675
0.0705	5.2461	55000	3.8942
0.071	5.7230	60000	4.0015
0.0547	6.1999	65000	4.1481
0.0554	6.6768	70000	4.2596
0.0463	7.1538	75000	4.3939
0.0451	7.6307	80000	4.4723
0.0395	8.1076	85000	4.6841
0.0395	8.5845	90000	4.7395
0.0361	9.0614	95000	4.8502
0.0368	9.5383	100000	4.8884

Framework versions

—Transformers 4.46.3
—Pytorch 2.3.0
—Datasets 2.14.4
—Tokenizers 0.20.3