zhangyiss/mlx-examples
1
0
Fork 0
mirror of https://github.com/ml-explore/mlx-examples.git synced 2025-09-01 04:14:38 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

YAML configuration for mlx_lm.lora (#503)

Browse Source 
* Convert mlx_lm.lora to use YAML configuration

* pre-commit run fixes

* Fix loading of config file

* Remove invalid YAML from doc

* Update command-line options and YAML parameter overriding, per feedback in #503

* Minor wording change

* Positional argument

* Moved config to a (-c/--config) flag

* Removed CLI option defaults (since CLI options take precedence and their defaults are in CONFIG_DEFAULTS)

* pre-commit format updates

* Fix handling of CLI option defaults

* Prevent None values of unspecified CLI options from overwriting values from CONFIG_DEFAULTS

* nits

---------

Co-authored-by: Awni Hannun <awni@apple.com>
This commit is contained in:
Chime Ogbuji
2024-03-08 10:57:52 -05:00
committed by GitHub
parent 8b05bb6d18
commit 8c2cf665ed
3 changed files with 129 additions and 28 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
llms/mlx_lm/LORA.md
View File

@@ -23,14 +23,24 @@ LoRA (QLoRA).[^qlora] LoRA fine-tuning works with the following model families:
## Run
The main command is `mlx_lm.lora`. To see a full list of options run:
The main command is `mlx_lm.lora`. To see a full list of command-line options run:
```shell
python -m mlx_lm.lora --help
```
Note, in the following the `--model` argument can be any compatible Hugging
Face repo or a local path to a converted model.
Face repo or a local path to a converted model.
You can also specify a YAML config with `-c`/`--config`. For more on the format see the
[example YAML](examples/lora_config.yaml). For example:
```shell
python -m mlx_lm.lora --config /path/to/config.yaml
```
If command-line flags are also used, they will override the corresponding
values in the config.
### Fine-tune
@@ -74,7 +84,7 @@ python -m mlx_lm.lora \
### Generate
For generation use mlx_lm.generate:
For generation use `mlx_lm.generate`:
```shell
python -m mlx_lm.generate \