run-llama

This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4304
• Accuracy: 0.7917

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: 32
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 200

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss
0.6426	0.1130	10	0.5625	0.6485
0.6074	0.2260	20	0.6319	0.5821
0.4355	0.3390	30	0.7153	0.4967
0.5957	0.4520	40	0.7569	0.4683
0.3984	0.5650	50	0.7431	0.4616
0.4795	0.9836	60	0.7361	0.4514
0.3154	1.1475	70	0.7292	0.4393
0.3086	1.3115	80	0.75	0.4339
0.499	1.4754	90	0.7569	0.4281
0.2642	1.6393	100	0.7847	0.4281
0.2173	1.8033	110	0.7708	0.4200
0.3281	1.9672	120	0.7847	0.4151
0.2329	3.4211	130	0.7847	0.4172
0.2891	3.6842	140	0.7847	0.4202
0.3906	3.9474	150	0.7708	0.4285
0.2373	4.1026	160	0.4338	0.7778
0.2144	4.3590	170	0.4312	0.7847
0.2715	4.6154	180	0.4300	0.7847
0.1855	4.8718	190	0.4327	0.7917
0.3506	5.1282	200	0.4304	0.7917

Framework versions

• PEFT 0.12.0
• Transformers 4.44.0
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1