train_sst2_123_1760637735

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the sst2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.2809
• Num Input Tokens Seen: 67743008

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: 0.001
• train_batch_size: 4
• eval_batch_size: 4
• seed: 123
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.0427	1.0	15154	0.0658	3385616
0.102	2.0	30308	0.0607	6774096
0.0054	3.0	45462	0.0601	10161824
0.1286	4.0	60616	0.0586	13549104
0.0045	5.0	75770	0.0595	16935568
0.0179	6.0	90924	0.0594	20320896
0.09	7.0	106078	0.0578	23709008
0.004	8.0	121232	0.0583	27099520
0.0295	9.0	136386	0.0605	30484864
0.0162	10.0	151540	0.0662	33869824
0.0142	11.0	166694	0.0644	37256608
0.0082	12.0	181848	0.0678	40640592
0.004	13.0	197002	0.0755	44027424
0.0641	14.0	212156	0.0835	47415744
0.0047	15.0	227310	0.0855	50803600
0.0082	16.0	242464	0.0896	54189408
0.0231	17.0	257618	0.0919	57578368
0.002	18.0	272772	0.1025	60965904
0.1573	19.0	287926	0.1054	64353008
0.0006	20.0	303080	0.1050	67743008

Framework versions

• PEFT 0.17.1
• Transformers 4.51.3
• Pytorch 2.9.0+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4