e93d50ea-7420-4148-8bc1-2c4ecbdc2835

This model is a fine-tuned version of MLP-KTLim/llama-3-Korean-Bllossom-8B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3223

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: 0.000212
• train_batch_size: 4
• eval_batch_size: 4
• seed: 120
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 50
• training_steps: 500

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0009	1	1.9441
0.895	0.0456	50	0.9801
0.7945	0.0912	100	0.8731
0.6886	0.1369	150	0.8619
0.7819	0.1825	200	0.7114
0.6048	0.2281	250	0.5683
0.503	0.2737	300	0.5383
0.4986	0.3193	350	0.4003
0.5244	0.3650	400	0.3518
0.5655	0.4106	450	0.3261
0.4691	0.4562	500	0.3223

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.5.0+cu124
• Datasets 3.0.1
• Tokenizers 0.20.1