jan-hq's picture
End of training
832f1d7 verified
metadata
license: llama2
library_name: peft
tags:
  - alignment-handbook
  - trl
  - sft
  - generated_from_trainer
  - trl
  - sft
  - generated_from_trainer
datasets:
  - jan-hq/indonesian_sft_binarized
base_model: Yellow-AI-NLP/komodo-7b-base
model-index:
  - name: komodo-7b-chat
    results: []

komodo-7b-chat

This model is a fine-tuned version of Yellow-AI-NLP/komodo-7b-base on the jan-hq/indonesian_sft_binarized dataset. It achieves the following results on the evaluation set:

  • Loss: 1.2932

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: 2e-05
  • train_batch_size: 2
  • eval_batch_size: 2
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 2
  • gradient_accumulation_steps: 16
  • total_train_batch_size: 64
  • total_eval_batch_size: 4
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 3

Training results

Training Loss Epoch Step Validation Loss
1.3204 0.99 67 1.3371
1.3083 1.99 134 1.2950
1.2921 2.98 201 1.2932

Framework versions

  • PEFT 0.8.2
  • Transformers 4.37.2
  • Pytorch 2.1.2+cu121
  • Datasets 2.14.6
  • Tokenizers 0.15.0