Merge pull request #142 from NanoCode012/feat/custom-prompt-readme
Browse filesFeat: Add custom prompt readme and add missing prompt strategies to Readme
README.md
CHANGED
@@ -159,8 +159,29 @@ Have dataset(s) in one of the following format (JSONL recommended):
|
|
159 |
```json
|
160 |
{"article": "...", "summary": "..."}
|
161 |
```
|
162 |
-
|
163 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
164 |
|
165 |
</details>
|
166 |
|
|
|
159 |
```json
|
160 |
{"article": "...", "summary": "..."}
|
161 |
```
|
162 |
+
- `alpaca_chat.load_qa`: question and answer for alpaca chat
|
163 |
+
```json
|
164 |
+
{"question": "...", "answer": "..."}
|
165 |
+
```
|
166 |
+
- `creative_acr.load_answer`: instruction and revision
|
167 |
+
```json
|
168 |
+
{"instruction": "...", "revision": "..."}
|
169 |
+
```
|
170 |
+
- `creative_acr.load_critique`: critique
|
171 |
+
```json
|
172 |
+
{"scores": "...", "critiques": "...", "instruction": "...", "answer": "..."}
|
173 |
+
```
|
174 |
+
- `creative_acr.load_revise`: critique and revise
|
175 |
+
```json
|
176 |
+
{"scores": "...", "critiques": "...", "instruction": "...", "answer": "...", "revision": "..."}
|
177 |
+
```
|
178 |
+
- `pygmalion`: pygmalion
|
179 |
+
```json
|
180 |
+
{"conversations": [{"role": "...", "value": "..."}]}
|
181 |
+
```
|
182 |
+
- custom prompts structure:
|
183 |
+
1. Add your method to a file in [prompt_strategies](src/axolotl/prompt_strategies). Please see other files as example.
|
184 |
+
2. Use your custom file name as the dataset type.
|
185 |
|
186 |
</details>
|
187 |
|