Llama-3.1-8B-Instruct-SFT-900

This model is a fine-tuned version of meta-llama/Llama-3.1-8B-Instruct on the bct_non_cot_sft_900 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1053

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-06
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 8
• 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: 0.1
• num_epochs: 10.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.201	0.9877	50	1.0016
0.1407	1.9753	100	0.1513
0.0885	2.9630	150	0.1082
0.0743	3.9506	200	0.1068
0.0855	4.9383	250	0.1062
0.0571	5.9259	300	0.1058
0.063	6.9136	350	0.1054
0.0597	7.9012	400	0.1057
0.0694	8.8889	450	0.1053
0.0593	9.8765	500	0.1053

Framework versions

• PEFT 0.12.0
• Transformers 4.45.2
• Pytorch 2.3.0
• Datasets 2.19.0
• Tokenizers 0.20.0