train_wsc_42_1760606567

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: 0.4219
• 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
0.3601	3.0	333	0.3483	130808
0.336	6.0	666	0.3607	260688
0.3504	9.0	999	0.3498	390936
0.3623	12.0	1332	0.3519	521840
0.3341	15.0	1665	0.3620	653216
0.3467	18.0	1998	0.3577	784328
0.3688	21.0	2331	0.3584	916192
0.3212	24.0	2664	0.3775	1046952
0.3438	27.0	2997	0.4106	1177504
0.3288	30.0	3330	0.4219	1308280

Framework versions

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