From 9aeadd40dc349ae3834cb4da4fc41713b5335462 Mon Sep 17 00:00:00 2001 From: Anchen Date: Sat, 30 Dec 2023 00:25:38 +1100 Subject: [PATCH] chore: fix the load quantization model for deepseek coder (#203) * chore: fix the load quantization model * change to explicitly check for quantization config --- llms/deepseek-coder/deepseek_coder.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/llms/deepseek-coder/deepseek_coder.py b/llms/deepseek-coder/deepseek_coder.py index 3f3928e7..9b8a8a3e 100644 --- a/llms/deepseek-coder/deepseek_coder.py +++ b/llms/deepseek-coder/deepseek_coder.py @@ -244,7 +244,7 @@ def load_model(model_path: str): model = DeepseekCoder(model_args) weights = mx.load(str(model_path / "weights.npz")) - if quantization := config.get("quantization", False): + if quantization is not None: nn.QuantizedLinear.quantize_module(model, **quantization) model.update(tree_unflatten(list(weights.items())))