@nielsr thanks for your response. I looked into it. I bit not sure about
# Data training arguments
chat_template: "{% for message in messages %}\n{% if message['role'] == 'user' %}\n{{ '<|user|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'system' %}\n{{ '<|system|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'assistant' %}\n{{ '<|assistant|>\n' + message['content'] + eos_token }}\n{% endif %}\n{% if loop.last and add_generation_prompt %}\n{{ '<|assistant|>' }}\n{% endif %}\n{% endfor %}"
dataset_mixer:
HuggingFaceH4/ultrachat_200k: 1.0
dataset_splits:
- train_sft
- test_sft
preprocessing_num_workers: 12
in Data training arguments, I couldnât understand chat_template . and âdataset_splitsâ. Because Iâm trying to do domain based fine-tuning, chunked my long text file and created huggingface dataset so it contains only train dataset. In this config, how can I pass the custom dataset? could you please give us some more examples?