train_sst2_1752826677

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: 0.1472
• Num Input Tokens Seen: 33869824

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
• 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: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.4049	0.5	7577	0.3706	1694048
0.3186	1.0	15154	0.3073	3385616
0.2869	1.5	22731	0.2524	5082864
0.2487	2.0	30308	0.2290	6774096
0.1956	2.5	37885	0.2120	8467152
0.1281	3.0	45462	0.1964	10161824
0.1389	3.5	53039	0.1876	11856000
0.1675	4.0	60616	0.1777	13549104
0.1593	4.5	68193	0.1701	15241168
0.0673	5.0	75770	0.1639	16935568
0.2066	5.5	83347	0.1590	18626160
0.1403	6.0	90924	0.1558	20320896
0.1004	6.5	98501	0.1542	22013696
0.1858	7.0	106078	0.1506	23709008
0.1576	7.5	113655	0.1496	25400400
0.0563	8.0	121232	0.1485	27099520
0.1171	8.5	128809	0.1478	28792480
0.0912	9.0	136386	0.1475	30484864
0.0944	9.5	143963	0.1476	32173664
0.0799	10.0	151540	0.1472	33869824

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.7.1+cu126
• Datasets 3.6.0
• Tokenizers 0.21.1