Enable native ModelOpt quantization support (2/3) (#9991)

Signed-off-by: Zhiyu Cheng <zhiyuc@nvidia.com>
2025-10-11 07:48:14 +00:00
parent 8b85926a6e
commit 129d299278
3 changed files with 127 additions and 26 deletions
@@ -86,6 +86,8 @@ class ModelConfig:
        dtype: str = "auto",
        quantization: Optional[str] = None,
        modelopt_quant: Optional[Union[str, Dict]] = None,
+        modelopt_checkpoint_restore_path: Optional[str] = None,
+        modelopt_checkpoint_save_path: Optional[str] = None,
        override_config_file: Optional[str] = None,
        is_draft_model: bool = False,
        hybrid_kvcache_ratio: Optional[float] = None,