[Core] Refactor model loading code (#4097)

2024-04-16 11:34:39 -07:00
parent 05434764cd
commit 69e1d2fb69
67 changed files with 1054 additions and 963 deletions
--- a/tests/quantization/test_autogptq_marlin_configs.py
+++ b/tests/quantization/test_autogptq_marlin_configs.py
@@ -36,8 +36,6 @@ def test_auto_gptq(model_quant_type: str, ) -> None:
        model_path,
        tokenizer_mode="auto",
        trust_remote_code=False,
-        download_dir=None,
-        load_format="dummy",
        seed=0,
        dtype="float16",
        revision=None,
@@ -49,8 +47,6 @@ def test_auto_gptq(model_quant_type: str, ) -> None:
        model_path,
        tokenizer_mode="auto",
        trust_remote_code=False,
-        download_dir=None,
-        load_format="dummy",
        seed=0,
        dtype="float16",
        revision=None,