Llama-3.1-8B-Instruct_sft_sg_values_p100_OA_gold

This model is a fine-tuned version of meta-llama/Llama-3.1-8B-Instruct on the sft_sg_values_p100_OA_gold dataset. It achieves the following results on the evaluation set:

• Loss: 0.1963

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-06
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss
1.4967	0.0935	250	0.4037
1.3228	0.1869	500	0.3101
1.1641	0.2804	750	0.2707
1.1352	0.3738	1000	0.2458
1.0849	0.4673	1250	0.2267
1.0522	0.5607	1500	0.2146
1.0489	0.6542	1750	0.2065
1.0515	0.7477	2000	0.2021
1.044	0.8411	2250	0.1979
1.0538	0.9346	2500	0.1966

Framework versions

• PEFT 0.15.2
• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 2.21.0
• Tokenizers 0.21.1