Skip to content

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.

Only enable fp16 on z image models that actually support it. (#12065)

comfyanonymous635406e283e9
diff --git a/comfy/ldm/lumina/model.py b/comfy/ldm/lumina/model.py
index b114d9e31..77d1abc97 100644
--- a/comfy/ldm/lumina/model.py
+++ b/comfy/ldm/lumina/model.py
@@ -451,6 +451,7 @@ class NextDiT(nn.Module):
device=None,
dtype=None,
operations=None,
+ **kwargs,
) -> None:
super().__init__()
self.dtype = dtype
diff --git a/comfy/model_detection.py b/comfy/model_detection.py
index b29a033cc..8cea16e50 100644
--- a/comfy/model_detection.py
+++ b/comfy/model_detection.py
@@ -444,6 +444,10 @@ def detect_unet_config(state_dict, key_prefix, metadata=None):
dit_config["ffn_dim_multiplier"] = (8.0 / 3.0)
dit_config["z_image_modulation"] = True
dit_config["time_scale"] = 1000.0
+ try:
+ dit_config["allow_fp16"] = torch.std(state_dict['{}layers.{}.ffn_norm1.weight'.format(key_prefix, dit_config["n_layers"] - 2)], unbiased=False).item() < 0.42
+ except Exception:
+ pass
if '{}cap_pad_token'.format(key_prefix) in state_dict_keys:
dit_config["pad_tokens_multiple"] = 32
sig_weight = state_dict.get('{}siglip_embedder.0.weight'.format(key_prefix), None)
diff --git a/comfy/supported_models.py b/comfy/supported_models.py
index 45d913fa6..d25271d6e 100644
--- a/comfy/supported_models.py
+++ b/comfy/supported_models.py
@@ -1093,7 +1093,7 @@ class ZImage(Lumina2):
def __init__(self, unet_config):
super().__init__(unet_config)
- if comfy.model_management.extended_fp16_support():
+ if comfy.model_management.extended_fp16_support() and unet_config.get("allow_fp16", False):
self.supported_inference_dtypes = self.supported_inference_dtypes.copy()
self.supported_inference_dtypes.insert(1, torch.float16)