Meta-Llama-3-8B-Instruct

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6482

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: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• total_eval_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
0.739	0.0714	100	0.8381
0.6745	0.1429	200	0.7258
0.644	0.2143	300	0.6986
0.6136	0.2857	400	0.6825
0.6225	0.3571	500	0.6725
0.6245	0.4286	600	0.6659
0.6006	0.5	700	0.6612
0.601	0.5714	800	0.6572
0.6225	0.6429	900	0.6543
0.6038	0.7143	1000	0.6520
0.5911	0.7857	1100	0.6504
0.6123	0.8571	1200	0.6491
0.5635	0.9286	1300	0.6484
0.5974	1.0	1400	0.6482

Framework versions

• PEFT 0.11.2.dev0
• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1