Merge branch 'comfyanonymous:master' into feature/cancel-queue-hotkey

This commit is contained in:
Miguel C 2024-01-01 09:01:44 -06:00 committed by GitHub
commit bf78495492
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
10 changed files with 80 additions and 59 deletions

View File

@ -1,9 +0,0 @@
{
"path-intellisense.mappings": {
"../": "${workspaceFolder}/web/extensions/core"
},
"[python]": {
"editor.defaultFormatter": "ms-python.autopep8"
},
"python.formatting.provider": "none"
}

View File

@ -66,6 +66,8 @@ fpvae_group.add_argument("--fp16-vae", action="store_true", help="Run the VAE in
fpvae_group.add_argument("--fp32-vae", action="store_true", help="Run the VAE in full precision fp32.")
fpvae_group.add_argument("--bf16-vae", action="store_true", help="Run the VAE in bf16.")
parser.add_argument("--cpu-vae", action="store_true", help="Run the VAE on the CPU.")
fpte_group = parser.add_mutually_exclusive_group()
fpte_group.add_argument("--fp8_e4m3fn-text-enc", action="store_true", help="Store text encoder weights in fp8 (e4m3fn variant).")
fpte_group.add_argument("--fp8_e5m2-text-enc", action="store_true", help="Store text encoder weights in fp8 (e5m2 variant).")

View File

@ -151,7 +151,7 @@ class CLIPVisionEmbeddings(torch.nn.Module):
def forward(self, pixel_values):
embeds = self.patch_embedding(pixel_values).flatten(2).transpose(1, 2)
return torch.cat([self.class_embedding.expand(pixel_values.shape[0], 1, -1), embeds], dim=1) + self.position_embedding.weight
return torch.cat([self.class_embedding.to(embeds.device).expand(pixel_values.shape[0], 1, -1), embeds], dim=1) + self.position_embedding.weight.to(embeds.device)
class CLIPVision(torch.nn.Module):

View File

@ -186,6 +186,9 @@ except:
if is_intel_xpu():
VAE_DTYPE = torch.bfloat16
if args.cpu_vae:
VAE_DTYPE = torch.float32
if args.fp16_vae:
VAE_DTYPE = torch.float16
elif args.bf16_vae:
@ -259,6 +262,14 @@ print("VAE dtype:", VAE_DTYPE)
current_loaded_models = []
def module_size(module):
module_mem = 0
sd = module.state_dict()
for k in sd:
t = sd[k]
module_mem += t.nelement() * t.element_size()
return module_mem
class LoadedModel:
def __init__(self, model):
self.model = model
@ -296,14 +307,14 @@ class LoadedModel:
if hasattr(m, "comfy_cast_weights"):
m.prev_comfy_cast_weights = m.comfy_cast_weights
m.comfy_cast_weights = True
module_mem = 0
sd = m.state_dict()
for k in sd:
t = sd[k]
module_mem += t.nelement() * t.element_size()
module_mem = module_size(m)
if mem_counter + module_mem < lowvram_model_memory:
m.to(self.device)
mem_counter += module_mem
elif hasattr(m, "weight"): #only modules with comfy_cast_weights can be set to lowvram mode
m.to(self.device)
mem_counter += module_size(m)
print("lowvram: loaded module regularly", m)
self.model_accelerated = True
@ -547,6 +558,8 @@ def intermediate_device():
return torch.device("cpu")
def vae_device():
if args.cpu_vae:
return torch.device("cpu")
return get_torch_device()
def vae_offload_device():

View File

@ -28,13 +28,9 @@ class ModelPatcher:
if self.size > 0:
return self.size
model_sd = self.model.state_dict()
size = 0
for k in model_sd:
t = model_sd[k]
size += t.nelement() * t.element_size()
self.size = size
self.size = comfy.model_management.module_size(self.model)
self.model_keys = set(model_sd.keys())
return size
return self.size
def clone(self):
n = ModelPatcher(self.model, self.load_device, self.offload_device, self.size, self.current_device, weight_inplace_update=self.weight_inplace_update)

View File

@ -13,6 +13,7 @@ class BasicScheduler:
{"model": ("MODEL",),
"scheduler": (comfy.samplers.SCHEDULER_NAMES, ),
"steps": ("INT", {"default": 20, "min": 1, "max": 10000}),
"denoise": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.01}),
}
}
RETURN_TYPES = ("SIGMAS",)
@ -20,8 +21,13 @@ class BasicScheduler:
FUNCTION = "get_sigmas"
def get_sigmas(self, model, scheduler, steps):
sigmas = comfy.samplers.calculate_sigmas_scheduler(model.model, scheduler, steps).cpu()
def get_sigmas(self, model, scheduler, steps, denoise):
total_steps = steps
if denoise < 1.0:
total_steps = int(steps/denoise)
sigmas = comfy.samplers.calculate_sigmas_scheduler(model.model, scheduler, total_steps).cpu()
sigmas = sigmas[-(steps + 1):]
return (sigmas, )

View File

@ -7,6 +7,7 @@ import threading
import heapq
import traceback
import gc
import inspect
import torch
import nodes
@ -402,6 +403,10 @@ def validate_inputs(prompt, item, validated):
errors = []
valid = True
validate_function_inputs = []
if hasattr(obj_class, "VALIDATE_INPUTS"):
validate_function_inputs = inspect.getfullargspec(obj_class.VALIDATE_INPUTS).args
for x in required_inputs:
if x not in inputs:
error = {
@ -531,29 +536,7 @@ def validate_inputs(prompt, item, validated):
errors.append(error)
continue
if hasattr(obj_class, "VALIDATE_INPUTS"):
input_data_all = get_input_data(inputs, obj_class, unique_id)
#ret = obj_class.VALIDATE_INPUTS(**input_data_all)
ret = map_node_over_list(obj_class, input_data_all, "VALIDATE_INPUTS")
for i, r in enumerate(ret):
if r is not True:
details = f"{x}"
if r is not False:
details += f" - {str(r)}"
error = {
"type": "custom_validation_failed",
"message": "Custom validation failed for node",
"details": details,
"extra_info": {
"input_name": x,
"input_config": info,
"received_value": val,
}
}
errors.append(error)
continue
else:
if x not in validate_function_inputs:
if isinstance(type_input, list):
if val not in type_input:
input_config = info
@ -580,6 +563,35 @@ def validate_inputs(prompt, item, validated):
errors.append(error)
continue
if len(validate_function_inputs) > 0:
input_data_all = get_input_data(inputs, obj_class, unique_id)
input_filtered = {}
for x in input_data_all:
if x in validate_function_inputs:
input_filtered[x] = input_data_all[x]
#ret = obj_class.VALIDATE_INPUTS(**input_filtered)
ret = map_node_over_list(obj_class, input_filtered, "VALIDATE_INPUTS")
for x in input_filtered:
for i, r in enumerate(ret):
if r is not True:
details = f"{x}"
if r is not False:
details += f" - {str(r)}"
error = {
"type": "custom_validation_failed",
"message": "Custom validation failed for node",
"details": details,
"extra_info": {
"input_name": x,
"input_config": info,
"received_value": val,
}
}
errors.append(error)
continue
if len(errors) > 0 or valid is not True:
ret = (False, errors, unique_id)
else:

View File

@ -184,8 +184,7 @@ def cached_filename_list_(folder_name):
if folder_name not in filename_list_cache:
return None
out = filename_list_cache[folder_name]
if time.perf_counter() < (out[2] + 0.5):
return out
for x in out[1]:
time_modified = out[1][x]
folder = x

View File

@ -1491,13 +1491,10 @@ class LoadImageMask:
return m.digest().hex()
@classmethod
def VALIDATE_INPUTS(s, image, channel):
def VALIDATE_INPUTS(s, image):
if not folder_paths.exists_annotated_filepath(image):
return "Invalid image file: {}".format(image)
if channel not in s._color_channels:
return "Invalid color channel: {}".format(channel)
return True
class ImageScale:

View File

@ -1781,10 +1781,19 @@ export class ComfyApp {
}
}
output[String(node.id)] = {
let node_data = {
inputs,
class_type: node.comfyClass,
};
if (this.ui.settings.getSettingValue("Comfy.DevMode")) {
// Ignored by the backend.
node_data["_meta"] = {
title: node.title,
}
}
output[String(node.id)] = node_data;
}
}
@ -2011,12 +2020,8 @@ export class ComfyApp {
async refreshComboInNodes() {
const defs = await api.getNodeDefs();
for(const nodeId in LiteGraph.registered_node_types) {
const node = LiteGraph.registered_node_types[nodeId];
const nodeDef = defs[nodeId];
if(!nodeDef) continue;
node.nodeData = nodeDef;
for (const nodeId in defs) {
this.registerNodeDef(nodeId, defs[nodeId]);
}
for(let nodeNum in this.graph._nodes) {