train_rte_1753094153

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

• Loss: 0.0824
• Num Input Tokens Seen: 3481336

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
0.387	0.5009	281	0.2907	176032
0.1348	1.0018	562	0.1241	349200
0.117	1.5027	843	0.1047	524208
0.0708	2.0036	1124	0.0999	699264
0.0739	2.5045	1405	0.0971	873600
0.0656	3.0053	1686	0.0924	1048184
0.105	3.5062	1967	0.0896	1223864
0.041	4.0071	2248	0.0880	1397624
0.0602	4.5080	2529	0.0871	1570936
0.102	5.0089	2810	0.0842	1746384
0.0558	5.5098	3091	0.0852	1922384
0.0511	6.0107	3372	0.0827	2092320
0.0312	6.5116	3653	0.0834	2267520
0.031	7.0125	3934	0.0845	2441688
0.0529	7.5134	4215	0.0831	2614936
0.0485	8.0143	4496	0.0827	2790832
0.0298	8.5152	4777	0.0824	2963888
0.0473	9.0160	5058	0.0828	3137352
0.0689	9.5169	5339	0.0824	3312648

Framework versions

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