train_wsc_1753094173

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: 490000

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
0.491	0.504	63	0.4640	25504
0.6266	1.008	126	0.5576	49696
0.3411	1.512	189	0.4053	74112
0.301	2.016	252	0.3891	99136
0.3604	2.52	315	0.3701	123904
0.3626	3.024	378	0.3648	148736
0.3679	3.528	441	0.3597	174432
0.3354	4.032	504	0.3662	198656
0.3301	4.536	567	0.3645	224032
0.3524	5.04	630	0.3640	247424
0.3559	5.5440	693	0.3592	271232
0.3673	6.048	756	0.3600	295728
0.3548	6.552	819	0.3643	320464
0.2933	7.056	882	0.3873	345856
0.3339	7.5600	945	0.3879	371040
0.3441	8.064	1008	0.3660	395216
0.3314	8.568	1071	0.3669	419184
0.3611	9.072	1134	0.3654	444560
0.3369	9.576	1197	0.3668	469104

Framework versions

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