[Misc] Reduce LoRA-related static variable (#13166)

2025-02-22 16:21:30 +08:00
parent 2cb8c1540e
commit 105b8ce4c0
41 changed files with 120 additions and 395 deletions
--- a/vllm/model_executor/models/gpt_bigcode.py
+++ b/vllm/model_executor/models/gpt_bigcode.py
@@ -261,15 +261,12 @@ class GPTBigCodeModel(nn.Module):
 class GPTBigCodeForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
    packed_modules_mapping = {"c_attn": ["c_attn"]}

-    supported_lora_modules = ["c_fc", "c_proj", "wte", "c_attn"]
-
+    # LoRA specific attributes
    embedding_modules = {
        "wte": "input_embeddings",
        "lm_head": "output_embeddings",
    }

-    embedding_padding_modules = []
-
    def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
        super().__init__()
        config = vllm_config.model_config.hf_config