In the README.md, you mentioned that

The data preparation code for self-distillation can be found in data_generation folder of the current repo.

In that folder, it says

python generate.py --data_path YOUR_DATA_PATH --output_path YOUR_OUTPUT_PATH --num_threads NUM_THREADS --max_tokens YOUR_MAX_TOKENS --temperature YOUR_TEMPERATURE

Which data/tokens/temperature should I use to reproduce existing Medusa-2 results? Should the --chat format be applied for reproduction? Could you list the full recipe for us?