train_wsc_42_1760346461

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: 0.3513
• Num Input Tokens Seen: 492304

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: 0.03
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• 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
0.4418	0.504	63	0.4744	24288
0.347	1.008	126	0.5146	49584
0.4193	1.512	189	0.4732	74512
0.414	2.016	252	0.3585	99264
0.4028	2.52	315	0.3567	123360
0.3931	3.024	378	0.3820	149120
0.4203	3.528	441	0.3545	174208
0.3707	4.032	504	0.3513	198016
0.3529	4.536	567	0.3557	223296
0.3528	5.04	630	0.4266	247344
0.3358	5.5440	693	0.3645	271856
0.3215	6.048	756	0.3555	297472
0.3793	6.552	819	0.3579	322272
0.3406	7.056	882	0.3583	347200
0.355	7.5600	945	0.3603	372576
0.3073	8.064	1008	0.3546	397008
0.332	8.568	1071	0.3553	421904
0.3455	9.072	1134	0.3530	446720
0.3336	9.576	1197	0.3529	471168

Framework versions

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