train_boolq_456_1765342117

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

• Loss: 1.2676
• Num Input Tokens Seen: 37905784

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
• train_batch_size: 4
• eval_batch_size: 4
• seed: 456
• 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
0.2652	2.0	3772	0.3582	3787392
0.2954	4.0	7544	0.3361	7574416
0.2594	6.0	11316	0.3424	11359560
0.3434	8.0	15088	0.3840	15147312
0.1768	10.0	18860	0.4522	18943384
0.0876	12.0	22632	0.6568	22733808
0.0422	14.0	26404	0.9193	26523680
0.0309	16.0	30176	1.1920	30314480
0.0002	18.0	33948	1.2599	34105096
0.0002	20.0	37720	1.2676	37905784

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.8.0+cu128
• Datasets 3.6.0
• Tokenizers 0.21.1