train_wsc_42_1760611410

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: 0.3865
• Num Input Tokens Seen: 1308280

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: 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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.3637	3.0	333	0.3500	130808
0.3389	6.0	666	0.3582	260688
0.355	9.0	999	0.3499	390936
0.3659	12.0	1332	0.3540	521840
0.3409	15.0	1665	0.3616	653216
0.3468	18.0	1998	0.3560	784328
0.3672	21.0	2331	0.3557	916192
0.326	24.0	2664	0.3640	1046952
0.3398	27.0	2997	0.3820	1177504
0.3305	30.0	3330	0.3865	1308280

Framework versions

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