--- language: - ja base_model: - microsoft/Phi-3.5-mini-instruct --- ## Base Model - [microsoft/Phi-3.5-mini-instruct](https://huggingface.co/microsoft/Phi-3.5-mini-instruct) ## Dataset - [amazon-science/esci-data](https://github.com/amazon-science/esci-data) - shopping_qeries_dataset_products.parquetの日本語データを使用(339,059件) ## Parameter - learning_rate: 1e-6 - num_train_epochs: 10 - per_device_train_batch_size: 64 - per_device_eval_batch_size: 64 - gradient_accumulation_steps: 1