llama_3_8b_lora_completion_only

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

• Loss: 0.0924

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: 0.0001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 8
• total_train_batch_size: 16
• total_eval_batch_size: 2
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 700
• num_epochs: 5.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.105	0.3690	250	0.0762
0.0716	0.7380	500	0.0897
0.0652	1.1070	750	0.0832
0.061	1.4760	1000	0.0640
0.0373	1.8450	1250	0.0813
0.0344	2.2140	1500	0.0686
0.0207	2.5830	1750	0.0662
0.0351	2.9520	2000	0.0669
0.0028	3.3210	2250	0.0996
0.0101	3.6900	2500	0.0718
0.0044	4.0590	2750	0.0825
0.0123	4.4280	3000	0.0969
0.0031	4.7970	3250	0.0924

Framework versions

• PEFT 0.10.0
• Transformers 4.40.0
• Pytorch 2.2.2+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1