The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Some improvements to the lowvram unloading.
comfyanonymous839ed3368efd
diff --git a/comfy/model_patcher.py b/comfy/model_patcher.pyindex fc2329543..f53f10749 100644--- a/comfy/model_patcher.py+++ b/comfy/model_patcher.py@@ -367,10 +367,7 @@ class ModelPatcher:else:set_func(out_weight, inplace_update=inplace_update, seed=string_to_seed(key))- def load(self, device_to=None, lowvram_model_memory=0, force_patch_weights=False, full_load=False):- mem_counter = 0- patch_counter = 0- lowvram_counter = 0+ def _load_list(self):loading = []for n, m in self.model.named_modules():params = []@@ -383,6 +380,13 @@ class ModelPatcher:breakif not skip and (hasattr(m, "comfy_cast_weights") or len(params) > 0):loading.append((comfy.model_management.module_size(m), n, m, params))+ return loading++ def load(self, device_to=None, lowvram_model_memory=0, force_patch_weights=False, full_load=False):+ mem_counter = 0+ patch_counter = 0+ lowvram_counter = 0+ loading = self._load_list()load_completely = []loading.sort(reverse=True)@@ -514,14 +518,7 @@ class ModelPatcher:def partially_unload(self, device_to, memory_to_free=0):memory_freed = 0patch_counter = 0- unload_list = []-- for n, m in self.model.named_modules():- shift_lowvram = False- if hasattr(m, "comfy_cast_weights"):- module_mem = comfy.model_management.module_size(m)- unload_list.append((module_mem, n, m))-+ unload_list = self._load_list()unload_list.sort()for unload in unload_list:if memory_to_free < memory_freed:@@ -529,32 +526,42 @@ class ModelPatcher:module_mem = unload[0]n = unload[1]m = unload[2]- weight_key = "{}.weight".format(n)- bias_key = "{}.bias".format(n)+ params = unload[3]+ lowvram_possible = hasattr(m, "comfy_cast_weights")if hasattr(m, "comfy_patched_weights") and m.comfy_patched_weights == True:- for key in [weight_key, bias_key]:+ move_weight = True+ for param in params:+ key = "{}.{}".format(n, param)bk = self.backup.get(key, None)if bk is not None:+ if not lowvram_possible:+ move_weight = False+ break+if bk.inplace_update:comfy.utils.copy_to_param(self.model, key, bk.weight)else:comfy.utils.set_attr_param(self.model, key, bk.weight)self.backup.pop(key)- m.to(device_to)- if weight_key in self.patches:- m.weight_function = LowVramPatch(weight_key, self.patches)- patch_counter += 1- if bias_key in self.patches:- m.bias_function = LowVramPatch(bias_key, self.patches)- patch_counter += 1-- m.prev_comfy_cast_weights = m.comfy_cast_weights- m.comfy_cast_weights = True- m.comfy_patched_weights = False- memory_freed += module_mem- logging.debug("freed {}".format(n))+ weight_key = "{}.weight".format(n)+ bias_key = "{}.bias".format(n)+ if move_weight:+ m.to(device_to)+ if lowvram_possible:+ if weight_key in self.patches:+ m.weight_function = LowVramPatch(weight_key, self.patches)+ patch_counter += 1+ if bias_key in self.patches:+ m.bias_function = LowVramPatch(bias_key, self.patches)+ patch_counter += 1++ m.prev_comfy_cast_weights = m.comfy_cast_weights+ m.comfy_cast_weights = True+ m.comfy_patched_weights = False+ memory_freed += module_mem+ logging.debug("freed {}".format(n))self.model.model_lowvram = Trueself.model.lowvram_patch_counter += patch_counter