Gemma support (#474)

* gemma support * format * lora support for gemma
2025-10-24 06:28:07 +08:00 · 2024-02-21 08:47:13 -08:00
parent cfeef6d9d2
commit ab9172baac
4 changed files with 190 additions and 2 deletions
--- a/llms/mlx_lm/tuner/trainer.py
+++ b/llms/mlx_lm/tuner/trainer.py
@@ -233,7 +233,7 @@ def train(
                val_info = {
                    "iteration": it + 1,
                    "val_loss": val_loss,
-                    "val_time": val_time
+                    "val_time": val_time,
                }
                training_callback.on_val_loss_report(val_info)

--- a/llms/mlx_lm/tuner/utils.py
+++ b/llms/mlx_lm/tuner/utils.py
@@ -31,6 +31,7 @@ def linear_to_lora_layers(model: nn.Module, num_lora_layers: int):
        "mixtral",
        "stablelm_epoch",
        "qwen2",
+        "gemma",
    ]:
        check_lora_layers(len(model.model.layers))