llama_3_sky_o1_4_full

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

• Loss: 0.7143

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: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• total_eval_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: 1

Training results

Training Loss	Epoch	Step	Validation Loss
0.804	0.0979	100	0.8166
0.7468	0.1959	200	0.8032
0.7245	0.2938	300	0.7886
0.6915	0.3918	400	0.7813
0.6794	0.4897	500	0.7666
0.6498	0.5877	600	0.7470
0.63	0.6856	700	0.7334
0.6123	0.7835	800	0.7229
0.5961	0.8815	900	0.7159
0.6015	0.9794	1000	0.7142

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1