[Misc]Further reduce BNB static variable (#10597)

Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
2024-11-27 14:54:12 +08:00
parent e85250b1d1
commit 15cc2a9f1a
14 changed files with 131 additions and 219 deletions
--- a/vllm/model_executor/models/falcon.py
+++ b/vllm/model_executor/models/falcon.py
@@ -412,12 +412,6 @@ class FalconForCausalLM(nn.Module, SupportsPP):

    # BitandBytes specific attributes
    bitsandbytes_stacked_params_mapping = {}
-    default_bitsandbytes_target_modules = [
-        ".query_key_value.",
-        ".dense.",
-        ".dense_h_to_4h.",
-        ".dense_4h_to_h.",
-    ]

    def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
        super().__init__()