train_wic_101112_1760638030

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

• Loss: 1.8323
• Num Input Tokens Seen: 7502512

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: 101112
• 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
0.349	2.0	2172	0.3450	750592
0.2779	4.0	4344	0.3794	1500800
0.305	6.0	6516	0.3511	2251232
0.3995	8.0	8688	0.3773	3001072
0.2226	10.0	10860	0.4986	3751408
0.2789	12.0	13032	0.8729	4501296
0.2892	14.0	15204	1.2637	5252112
0.0002	16.0	17376	1.7115	6002224
0.0001	18.0	19548	1.8162	6752432
0.0	20.0	21720	1.8323	7502512

Framework versions

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