Llama-3.1-8B-Instruct-EI2-2ep-sft-bs

This model is a fine-tuned version of qfq/Llama-3.1-8B-Instruct-EI1-2ep-sft on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1866

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: 6e-06
• train_batch_size: 1
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 16
• total_train_batch_size: 16
• total_eval_batch_size: 128
• optimizer: Adam with betas=(0.9,0.95) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 2.0

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.1456	100	0.2037
No log	0.2911	200	0.2081
No log	0.4367	300	0.2095
No log	0.5822	400	0.2072
0.1766	0.7278	500	0.2051
0.1766	0.8734	600	0.1987
0.1766	1.0189	700	0.2012
0.1766	1.1645	800	0.2008
0.1766	1.3100	900	0.1960
0.1264	1.4556	1000	0.1920
0.1264	1.6012	1100	0.1894
0.1264	1.7467	1200	0.1877
0.1264	1.8923	1300	0.1866

Framework versions

• Transformers 4.43.4
• Pytorch 2.4.0+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1