spow12/EEVE_ver_4.1_sft

ver_4.1_sft

This model is a fine-tuned version of yanolja/EEVE-Korean-10.8B-v1.0 on the Custom dataset.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• total_train_batch_size: 8
• total_eval_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 0

Training results

Training Loss	Epoch	Step
0.7739	0.2	1124
0.7214	0.4	2248
0.6832	0.6	3372
0.6935	0.8	4496

Framework versions

• PEFT 0.10.0
• Transformers 4.39.3
• Pytorch 2.0.1
• Datasets 2.18.0
• Tokenizers 0.15.2