Public/automatic
1
0
Fork 0
mirror of https://github.com/vladmandic/automatic synced 2026-09-19 09:14:35 +02:00
Code Issues Packages Projects Releases Wiki Activity

Fix Chroma quantization

Browse Source
This commit is contained in:
Disty0
2025-08-12 22:07:52 +03:00
parent a36676a5a9
commit 362ec0d412
2 changed files with 5 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
pipelines/generic.py
+4 -4
View File
@@ -8,10 +8,10 @@ from modules import shared, devices, errors, sd_models, model_quant
debug = os.environ.get('SD_LOAD_DEBUG', None) is not None
def load_transformer(repo_id, cls_name, load_config={}, subfolder="transformer", allow_quant=True, variant=None, dtype=None):
def load_transformer(repo_id, cls_name, load_config={}, subfolder="transformer", allow_quant=True, variant=None, dtype=None, modules_to_not_convert=[]):
transformer = None
try:
load_args, quant_args = model_quant.get_dit_args(load_config, module='Model', device_map=True, allow_quant=allow_quant)
load_args, quant_args = model_quant.get_dit_args(load_config, module='Model', device_map=True, allow_quant=allow_quant, modules_to_not_convert=modules_to_not_convert)
quant_type = model_quant.get_quant_type(quant_args)
dtype = dtype or devices.dtype
@@ -71,10 +71,10 @@ def load_transformer(repo_id, cls_name, load_config={}, subfolder="transformer",
return transformer
def load_text_encoder(repo_id, cls_name, load_config={}, subfolder="text_encoder", allow_quant=True, allow_shared=True, variant=None, dtype=None):
def load_text_encoder(repo_id, cls_name, load_config={}, subfolder="text_encoder", allow_quant=True, allow_shared=True, variant=None, dtype=None, modules_to_not_convert=[]):
text_encoder = None
try:
load_args, quant_args = model_quant.get_dit_args(load_config, module='TE', device_map=True, allow_quant=allow_quant)
load_args, quant_args = model_quant.get_dit_args(load_config, module='TE', device_map=True, allow_quant=allow_quant, modules_to_not_convert=modules_to_not_convert)
quant_type = model_quant.get_quant_type(quant_args)
dtype = dtype or devices.dtype
pipelines/model_chroma.py
+1 -1
View File
@@ -11,7 +11,7 @@ def load_chroma(checkpoint_info, diffusers_load_config={}):
load_args, _quant_args = model_quant.get_dit_args(diffusers_load_config, allow_quant=False)
shared.log.debug(f'Load model: type=Chroma repo="{repo_id}" config={diffusers_load_config} offload={shared.opts.diffusers_offload_mode} dtype={devices.dtype} args={load_args}')
transformer = generic.load_transformer(repo_id, cls_name=diffusers.ChromaTransformer2DModel, load_config=diffusers_load_config)
transformer = generic.load_transformer(repo_id, cls_name=diffusers.ChromaTransformer2DModel, load_config=diffusers_load_config, modules_to_not_convert=["distilled_guidance_layer"])
text_encoder = generic.load_text_encoder(repo_id, cls_name=transformers.T5EncoderModel, load_config=diffusers_load_config)
pipe = diffusers.ChromaPipeline.from_pretrained(