From 243fe37a678930a0ef2ee9e899933730f4fa61ae Mon Sep 17 00:00:00 2001
From: Nehanth <nehanthnarendrula@gmail.com>
Date: Wed, 25 Jun 2025 16:40:37 -0400
Subject: [PATCH] fix: Add type field to DPOAlignmentConfig for remote provider
 serialization

Addresses review feedback to keep preference optimization APIs DPO-specific
rather than accepting all algorithm types.
---
 llama_stack/apis/post_training/post_training.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/llama_stack/apis/post_training/post_training.py b/llama_stack/apis/post_training/post_training.py
index 1a588b025..1c03cc1b2 100644
--- a/llama_stack/apis/post_training/post_training.py
+++ b/llama_stack/apis/post_training/post_training.py
@@ -125,7 +125,7 @@ class PostTrainingRLHFRequest(BaseModel):
     validation_dataset_id: str
 
     algorithm: RLHFAlgorithm
-    algorithm_config: AlgorithmConfig
+    algorithm_config: DPOAlignmentConfig
 
     optimizer_config: OptimizerConfig
     training_config: TrainingConfig
@@ -202,7 +202,7 @@ class PostTraining(Protocol):
         self,
         job_uuid: str,
         finetuned_model: str,
-        algorithm_config: AlgorithmConfig,
+        algorithm_config: DPOAlignmentConfig,
         training_config: TrainingConfig,
         hyperparam_search_config: dict[str, Any],
         logger_config: dict[str, Any],