llama-3-full-data-changed

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0986
• Accuracy: 0.5904
• F1: 0.5855
• Precision: 0.5975
• Recall: 0.5904

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: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.4985	0.1935	2000	1.1392	0.5867	0.5795	0.5962	0.5867
0.4899	0.3870	4000	1.0967	0.5910	0.5827	0.6023	0.5910
0.4878	0.5805	6000	1.1262	0.5903	0.5826	0.6007	0.5903
0.478	0.7740	8000	1.0881	0.5806	0.5784	0.5840	0.5806
0.471	0.9675	10000	1.1098	0.5786	0.5764	0.5820	0.5786
0.4632	1.1610	12000	1.1000	0.5748	0.5743	0.5760	0.5748
0.4561	1.3546	14000	1.1171	0.5868	0.5823	0.5933	0.5868
0.4585	1.5481	16000	1.1054	0.5914	0.5836	0.6020	0.5914
0.4561	1.7416	18000	1.0993	0.5895	0.5848	0.5962	0.5895
0.4564	1.9351	20000	1.0986	0.5904	0.5855	0.5975	0.5904

Framework versions

• PEFT 0.10.0
• Transformers 4.40.0
• Pytorch 2.2.2+cu121
• Datasets 2.18.0
• Tokenizers 0.19.1