tocmo0nlord/axolotl
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

WIP large refactor to make finetune script a little more manageable (#3)

Browse Source
This commit is contained in:
Wing Lian
2023-04-18 14:01:38 -04:00
committed by GitHub
parent 81de0efc18
commit 6045345d6b
9 changed files with 524 additions and 398 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/axolotl/prompters.py
View File

@@ -35,6 +35,10 @@ class GPTeacherPrompter(AlpacaPrompter):
...
class NomicGPT4AllPrompter(AlpacaPrompter):
...
class ReflectAlpacaPrompter:
prompt_input = "Below is an instruction that describes a task, paired with an input that provides further context. You, the Assistant, should generate a response as if it were an abstract for an academic or technical paper on the query along with a methodology. Then generate an Agent Reflection where you create a long form response as if from subject matter expert, be verbose, diligent, and creative in your application of knowledge, apply it through the lens of the response generated by the assistant. Look for flawed reasoning, faulty logic, or other mistakes in the method. Finally, generate a final response and method for the user with the Assistant abstract and Reflection analysis as augmentations to the generation\n\n### Instruction:\n{instruction}\n\n### Input:\n{input}\n\n### Response:\n"
prompt_no_input = "Below is an instruction that describes a task. You, the Assistant, should generate a response as if it were an abstract for an academic or technical paper on the query along with a methodology. Then generate an Agent Reflection where you create a long form response as if from subject matter expert, be verbose, diligent, and creative in your application of knowledge, apply it through the lens of the response generated by the assistant. Look for flawed reasoning, faulty logic, or other mistakes in the method. Finally, generate a final response and method for the user with the Assistant abstract and Reflection analysis as augmentations to the generation\n\n### Instruction:\n{instruction}\n\n### Response:\n"