Make old scaled fp8 format use the new mixed quant ops system. (#11000)
This commit is contained in:
@@ -34,12 +34,9 @@ class ZImageTEModel(sd1_clip.SD1ClipModel):
|
||||
super().__init__(device=device, dtype=dtype, name="qwen3_4b", clip_model=Qwen3_4BModel, model_options=model_options)
|
||||
|
||||
|
||||
def te(dtype_llama=None, llama_scaled_fp8=None, llama_quantization_metadata=None):
|
||||
def te(dtype_llama=None, llama_quantization_metadata=None):
|
||||
class ZImageTEModel_(ZImageTEModel):
|
||||
def __init__(self, device="cpu", dtype=None, model_options={}):
|
||||
if llama_scaled_fp8 is not None and "scaled_fp8" not in model_options:
|
||||
model_options = model_options.copy()
|
||||
model_options["scaled_fp8"] = llama_scaled_fp8
|
||||
if dtype_llama is not None:
|
||||
dtype = dtype_llama
|
||||
if llama_quantization_metadata is not None:
|
||||
|
||||
Reference in New Issue
Block a user