zhangyiss/mlx-examples
1
0
Fork 0
mirror of https://github.com/ml-explore/mlx-examples.git synced 2025-09-01 12:49:50 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

Additional comment about YAML file format

Browse Source
This commit is contained in:
Chime Ogbuji
2024-01-05 11:59:37 -05:00
parent 4f7023ce19
commit 0dbc78ca65
1 changed files with 9 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
lora/supervised_lora.py
View File

@@ -2,6 +2,15 @@
Based on lora and performs supervised instruction finetuning
However, it but breaks out parameters into a YAML file and allows arbitrary training data (and prompt) formats
The configuration .yaml file is expected to be in the following format:
parameters:
model: "..."
num_tokens: 100
[..]
Where each entry under parameters is the argparse version of the argumens originally provided to lora.py plus new ones
A module for a particular prompt syntax or training dataset format just needs to overide TrainingRecordHandler,
provide an instance of it to main, and run the script with a single argument which is a path to a YAML file with the
configuration parameters that were originally command-line arguments in lora.py