From e1b214c62b24594465d756452008e83f5bd82c68 Mon Sep 17 00:00:00 2001 From: Casper Date: Sat, 14 Oct 2023 15:28:12 +0200 Subject: [PATCH] Clarify custom format example (#729) * Clarify custom prompt format * Simplify format --- README.md | 23 +++++++++++------------ 1 file changed, 11 insertions(+), 12 deletions(-) diff --git a/README.md b/README.md index 57447b36de..c70abf6480 100644 --- a/README.md +++ b/README.md @@ -297,25 +297,24 @@ Have dataset(s) in one of the following format (JSONL recommended): #### How to add custom prompts -Using yaml. Example: +For a dataset that is preprocessed for instruction purposes: + +```json +{"instruction": "...", "output": "..."} +``` + +You can use this example in your YAML config: + ```yaml datasets: - path: repo type: system_prompt: "" - no_input_format: |- - User: {instruction}<|end_of_turn|> - Assistant: - format: |- - User: {instruction} - {input}<|end_of_turn|> - Assistant: + field_system: system + format: "[INST] {instruction} [/INST]" + no_input_format: "[INST] {instruction} [/INST]" ``` -Using file: -1. Add your method to a file in [prompt_strategies](src/axolotl/prompt_strategies). Please see other files as example. -2. Use your custom file name as the dataset type `.load_`. - #### How to use your custom pretokenized dataset - Do not pass a `type:`