This model is a instruct-tuned poylglot-ko-12.8b model, using 10% [Kullm, OIG, KoAlpaca] Instruction dataset. len10_k100_mrand_n0.01.json -> 29step
Training hyperparameters
- learning_rate: 5e-5
- seed: 42
- distributed_type: multi-GPU (A100 40G) + CPU offloading (512GB)
- num_devices: 1
- train_batch_size: 4
- gradient_accumulation_steps: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2.0
Framework versions
- Transformers 4.35.0
- Pytorch 2.0.1+cu117
- Datasets 2.14.6
- deepspeed 0.11.1
- accelerate 0.24.1
- Downloads last month
- 1,182