train_wsc_456_1760637765

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

• Loss: 0.3592
• Num Input Tokens Seen: 875272

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
0.4044	2.0	222	0.3432	87200
0.3604	4.0	444	0.3508	174112
0.3466	6.0	666	0.3519	261352
0.3826	8.0	888	0.3522	348888
0.3377	10.0	1110	0.3515	436520
0.3469	12.0	1332	0.3579	524608
0.3416	14.0	1554	0.3550	611968
0.3389	16.0	1776	0.3580	699416
0.3391	18.0	1998	0.3589	787304
0.342	20.0	2220	0.3592	875272

Framework versions

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