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.

Fix int8 loras causing lower quality requant with wrong settings. (#14650)

comfyanonymous470ac36a0a80
diff --git a/comfy/ops.py b/comfy/ops.py
index 634610f1c..6a5090548 100644
--- a/comfy/ops.py
+++ b/comfy/ops.py
@@ -256,7 +256,7 @@ def resolve_cast_module_with_vbar(s, dtype, device, bias_dtype, compute_dtype, w
if (want_requant and len(fns) == 0 or update_weight):
seed = comfy.utils.string_to_seed(s.seed_key)
if isinstance(orig, QuantizedTensor):
- y = QuantizedTensor.from_float(x, s.layout_type, scale="recalculate", stochastic_rounding=seed)
+ y = orig.requantize_from_float(x, scale="recalculate", stochastic_rounding=seed)
else:
y = comfy.float.stochastic_rounding(x, orig.dtype, seed=seed)
if want_requant and len(fns) == 0:
@@ -1306,8 +1306,7 @@ def mixed_precision_ops(quant_config={}, compute_dtype=torch.bfloat16, full_prec
def set_weight(self, weight, inplace_update=False, seed=None, return_weight=False, **kwargs):
if getattr(self, 'layout_type', None) is not None:
- # dtype is now implicit in the layout class
- weight = QuantizedTensor.from_float(weight, self.layout_type, scale="recalculate", stochastic_rounding=seed, inplace_ops=True).to(self.weight.dtype)
+ weight = self.weight.requantize_from_float(weight, scale="recalculate", stochastic_rounding=seed, inplace_ops=True).to(self.weight.dtype)
else:
weight = weight.to(self.weight.dtype)
if return_weight:
diff --git a/requirements.txt b/requirements.txt
index eea7724f3..d7719178b 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -22,7 +22,7 @@ alembic
SQLAlchemy>=2.0.0
filelock
av>=16.0.0
-comfy-kitchen==0.2.12
+comfy-kitchen==0.2.13
comfy-aimdo==0.4.10
requests
simpleeval>=1.0.0