diff --git a/README.md b/README.md
index 0f7d24c45..90931141d 100644
--- a/README.md
+++ b/README.md
@@ -14,7 +14,7 @@ This ui will let you design and execute advanced stable diffusion pipelines usin
- Many optimizations: Only re-executes the parts of the workflow that changes between executions.
- Command line option: ```--lowvram``` to make it work on GPUs with less than 3GB vram (enabled automatically on GPUs with low vram)
- Works even if you don't have a GPU with: ```--cpu``` (slow)
-- Can load both ckpt and safetensors models/checkpoints. Standalone VAEs and CLIP models.
+- Can load ckpt, safetensors and diffusers models/checkpoints. Standalone VAEs and CLIP models.
- Embeddings/Textual inversion
- [Loras (regular, locon and loha)](https://comfyanonymous.github.io/ComfyUI_examples/lora/)
- Loading full workflows (with seeds) from generated PNG files.
diff --git a/comfy/cli_args.py b/comfy/cli_args.py
new file mode 100644
index 000000000..b24054ce0
--- /dev/null
+++ b/comfy/cli_args.py
@@ -0,0 +1,31 @@
+import argparse
+
+parser = argparse.ArgumentParser()
+
+parser.add_argument("--listen", type=str, default="127.0.0.1", metavar="IP", nargs="?", const="0.0.0.0", help="Specify the IP address to listen on (default: 127.0.0.1). If --listen is provided without an argument, it defaults to 0.0.0.0. (listens on all)")
+parser.add_argument("--port", type=int, default=8188, help="Set the listen port.")
+parser.add_argument("--enable-cors-header", type=str, default=None, metavar="ORIGIN", nargs="?", const="*", help="Enable CORS (Cross-Origin Resource Sharing) with optional origin or allow all with default '*'.")
+parser.add_argument("--extra-model-paths-config", type=str, default=None, metavar="PATH", nargs='+', action='append', help="Load one or more extra_model_paths.yaml files.")
+parser.add_argument("--output-directory", type=str, default=None, help="Set the ComfyUI output directory.")
+parser.add_argument("--cuda-device", type=int, default=None, metavar="DEVICE_ID", help="Set the id of the cuda device this instance will use.")
+parser.add_argument("--dont-upcast-attention", action="store_true", help="Disable upcasting of attention. Can boost speed but increase the chances of black images.")
+parser.add_argument("--force-fp32", action="store_true", help="Force fp32 (If this makes your GPU work better please report it).")
+
+attn_group = parser.add_mutually_exclusive_group()
+attn_group.add_argument("--use-split-cross-attention", action="store_true", help="Use the split cross attention optimization instead of the sub-quadratic one. Ignored when xformers is used.")
+attn_group.add_argument("--use-pytorch-cross-attention", action="store_true", help="Use the new pytorch 2.0 cross attention function.")
+
+parser.add_argument("--disable-xformers", action="store_true", help="Disable xformers.")
+
+vram_group = parser.add_mutually_exclusive_group()
+vram_group.add_argument("--highvram", action="store_true", help="By default models will be unloaded to CPU memory after being used. This option keeps them in GPU memory.")
+vram_group.add_argument("--normalvram", action="store_true", help="Used to force normal vram use if lowvram gets automatically enabled.")
+vram_group.add_argument("--lowvram", action="store_true", help="Split the unet in parts to use less vram.")
+vram_group.add_argument("--novram", action="store_true", help="When lowvram isn't enough.")
+vram_group.add_argument("--cpu", action="store_true", help="To use the CPU for everything (slow).")
+
+parser.add_argument("--dont-print-server", action="store_true", help="Don't print server output.")
+parser.add_argument("--quick-test-for-ci", action="store_true", help="Quick test for CI.")
+parser.add_argument("--windows-standalone-build", action="store_true", help="Windows standalone build: Enable convenient things that most people using the standalone windows build will probably enjoy (like auto opening the page on startup).")
+
+args = parser.parse_args()
diff --git a/comfy/diffusers_convert.py b/comfy/diffusers_convert.py
new file mode 100644
index 000000000..ceca80305
--- /dev/null
+++ b/comfy/diffusers_convert.py
@@ -0,0 +1,362 @@
+import json
+import os
+import yaml
+
+import folder_paths
+from comfy.ldm.util import instantiate_from_config
+from comfy.sd import ModelPatcher, load_model_weights, CLIP, VAE
+import os.path as osp
+import re
+import torch
+from safetensors.torch import load_file, save_file
+
+# conversion code from https://github.com/huggingface/diffusers/blob/main/scripts/convert_diffusers_to_original_stable_diffusion.py
+
+# =================#
+# UNet Conversion #
+# =================#
+
+unet_conversion_map = [
+ # (stable-diffusion, HF Diffusers)
+ ("time_embed.0.weight", "time_embedding.linear_1.weight"),
+ ("time_embed.0.bias", "time_embedding.linear_1.bias"),
+ ("time_embed.2.weight", "time_embedding.linear_2.weight"),
+ ("time_embed.2.bias", "time_embedding.linear_2.bias"),
+ ("input_blocks.0.0.weight", "conv_in.weight"),
+ ("input_blocks.0.0.bias", "conv_in.bias"),
+ ("out.0.weight", "conv_norm_out.weight"),
+ ("out.0.bias", "conv_norm_out.bias"),
+ ("out.2.weight", "conv_out.weight"),
+ ("out.2.bias", "conv_out.bias"),
+]
+
+unet_conversion_map_resnet = [
+ # (stable-diffusion, HF Diffusers)
+ ("in_layers.0", "norm1"),
+ ("in_layers.2", "conv1"),
+ ("out_layers.0", "norm2"),
+ ("out_layers.3", "conv2"),
+ ("emb_layers.1", "time_emb_proj"),
+ ("skip_connection", "conv_shortcut"),
+]
+
+unet_conversion_map_layer = []
+# hardcoded number of downblocks and resnets/attentions...
+# would need smarter logic for other networks.
+for i in range(4):
+ # loop over downblocks/upblocks
+
+ for j in range(2):
+ # loop over resnets/attentions for downblocks
+ hf_down_res_prefix = f"down_blocks.{i}.resnets.{j}."
+ sd_down_res_prefix = f"input_blocks.{3 * i + j + 1}.0."
+ unet_conversion_map_layer.append((sd_down_res_prefix, hf_down_res_prefix))
+
+ if i < 3:
+ # no attention layers in down_blocks.3
+ hf_down_atn_prefix = f"down_blocks.{i}.attentions.{j}."
+ sd_down_atn_prefix = f"input_blocks.{3 * i + j + 1}.1."
+ unet_conversion_map_layer.append((sd_down_atn_prefix, hf_down_atn_prefix))
+
+ for j in range(3):
+ # loop over resnets/attentions for upblocks
+ hf_up_res_prefix = f"up_blocks.{i}.resnets.{j}."
+ sd_up_res_prefix = f"output_blocks.{3 * i + j}.0."
+ unet_conversion_map_layer.append((sd_up_res_prefix, hf_up_res_prefix))
+
+ if i > 0:
+ # no attention layers in up_blocks.0
+ hf_up_atn_prefix = f"up_blocks.{i}.attentions.{j}."
+ sd_up_atn_prefix = f"output_blocks.{3 * i + j}.1."
+ unet_conversion_map_layer.append((sd_up_atn_prefix, hf_up_atn_prefix))
+
+ if i < 3:
+ # no downsample in down_blocks.3
+ hf_downsample_prefix = f"down_blocks.{i}.downsamplers.0.conv."
+ sd_downsample_prefix = f"input_blocks.{3 * (i + 1)}.0.op."
+ unet_conversion_map_layer.append((sd_downsample_prefix, hf_downsample_prefix))
+
+ # no upsample in up_blocks.3
+ hf_upsample_prefix = f"up_blocks.{i}.upsamplers.0."
+ sd_upsample_prefix = f"output_blocks.{3 * i + 2}.{1 if i == 0 else 2}."
+ unet_conversion_map_layer.append((sd_upsample_prefix, hf_upsample_prefix))
+
+hf_mid_atn_prefix = "mid_block.attentions.0."
+sd_mid_atn_prefix = "middle_block.1."
+unet_conversion_map_layer.append((sd_mid_atn_prefix, hf_mid_atn_prefix))
+
+for j in range(2):
+ hf_mid_res_prefix = f"mid_block.resnets.{j}."
+ sd_mid_res_prefix = f"middle_block.{2 * j}."
+ unet_conversion_map_layer.append((sd_mid_res_prefix, hf_mid_res_prefix))
+
+
+def convert_unet_state_dict(unet_state_dict):
+ # buyer beware: this is a *brittle* function,
+ # and correct output requires that all of these pieces interact in
+ # the exact order in which I have arranged them.
+ mapping = {k: k for k in unet_state_dict.keys()}
+ for sd_name, hf_name in unet_conversion_map:
+ mapping[hf_name] = sd_name
+ for k, v in mapping.items():
+ if "resnets" in k:
+ for sd_part, hf_part in unet_conversion_map_resnet:
+ v = v.replace(hf_part, sd_part)
+ mapping[k] = v
+ for k, v in mapping.items():
+ for sd_part, hf_part in unet_conversion_map_layer:
+ v = v.replace(hf_part, sd_part)
+ mapping[k] = v
+ new_state_dict = {v: unet_state_dict[k] for k, v in mapping.items()}
+ return new_state_dict
+
+
+# ================#
+# VAE Conversion #
+# ================#
+
+vae_conversion_map = [
+ # (stable-diffusion, HF Diffusers)
+ ("nin_shortcut", "conv_shortcut"),
+ ("norm_out", "conv_norm_out"),
+ ("mid.attn_1.", "mid_block.attentions.0."),
+]
+
+for i in range(4):
+ # down_blocks have two resnets
+ for j in range(2):
+ hf_down_prefix = f"encoder.down_blocks.{i}.resnets.{j}."
+ sd_down_prefix = f"encoder.down.{i}.block.{j}."
+ vae_conversion_map.append((sd_down_prefix, hf_down_prefix))
+
+ if i < 3:
+ hf_downsample_prefix = f"down_blocks.{i}.downsamplers.0."
+ sd_downsample_prefix = f"down.{i}.downsample."
+ vae_conversion_map.append((sd_downsample_prefix, hf_downsample_prefix))
+
+ hf_upsample_prefix = f"up_blocks.{i}.upsamplers.0."
+ sd_upsample_prefix = f"up.{3 - i}.upsample."
+ vae_conversion_map.append((sd_upsample_prefix, hf_upsample_prefix))
+
+ # up_blocks have three resnets
+ # also, up blocks in hf are numbered in reverse from sd
+ for j in range(3):
+ hf_up_prefix = f"decoder.up_blocks.{i}.resnets.{j}."
+ sd_up_prefix = f"decoder.up.{3 - i}.block.{j}."
+ vae_conversion_map.append((sd_up_prefix, hf_up_prefix))
+
+# this part accounts for mid blocks in both the encoder and the decoder
+for i in range(2):
+ hf_mid_res_prefix = f"mid_block.resnets.{i}."
+ sd_mid_res_prefix = f"mid.block_{i + 1}."
+ vae_conversion_map.append((sd_mid_res_prefix, hf_mid_res_prefix))
+
+vae_conversion_map_attn = [
+ # (stable-diffusion, HF Diffusers)
+ ("norm.", "group_norm."),
+ ("q.", "query."),
+ ("k.", "key."),
+ ("v.", "value."),
+ ("proj_out.", "proj_attn."),
+]
+
+
+def reshape_weight_for_sd(w):
+ # convert HF linear weights to SD conv2d weights
+ return w.reshape(*w.shape, 1, 1)
+
+
+def convert_vae_state_dict(vae_state_dict):
+ mapping = {k: k for k in vae_state_dict.keys()}
+ for k, v in mapping.items():
+ for sd_part, hf_part in vae_conversion_map:
+ v = v.replace(hf_part, sd_part)
+ mapping[k] = v
+ for k, v in mapping.items():
+ if "attentions" in k:
+ for sd_part, hf_part in vae_conversion_map_attn:
+ v = v.replace(hf_part, sd_part)
+ mapping[k] = v
+ new_state_dict = {v: vae_state_dict[k] for k, v in mapping.items()}
+ weights_to_convert = ["q", "k", "v", "proj_out"]
+ for k, v in new_state_dict.items():
+ for weight_name in weights_to_convert:
+ if f"mid.attn_1.{weight_name}.weight" in k:
+ print(f"Reshaping {k} for SD format")
+ new_state_dict[k] = reshape_weight_for_sd(v)
+ return new_state_dict
+
+
+# =========================#
+# Text Encoder Conversion #
+# =========================#
+
+
+textenc_conversion_lst = [
+ # (stable-diffusion, HF Diffusers)
+ ("resblocks.", "text_model.encoder.layers."),
+ ("ln_1", "layer_norm1"),
+ ("ln_2", "layer_norm2"),
+ (".c_fc.", ".fc1."),
+ (".c_proj.", ".fc2."),
+ (".attn", ".self_attn"),
+ ("ln_final.", "transformer.text_model.final_layer_norm."),
+ ("token_embedding.weight", "transformer.text_model.embeddings.token_embedding.weight"),
+ ("positional_embedding", "transformer.text_model.embeddings.position_embedding.weight"),
+]
+protected = {re.escape(x[1]): x[0] for x in textenc_conversion_lst}
+textenc_pattern = re.compile("|".join(protected.keys()))
+
+# Ordering is from https://github.com/pytorch/pytorch/blob/master/test/cpp/api/modules.cpp
+code2idx = {"q": 0, "k": 1, "v": 2}
+
+
+def convert_text_enc_state_dict_v20(text_enc_dict):
+ new_state_dict = {}
+ capture_qkv_weight = {}
+ capture_qkv_bias = {}
+ for k, v in text_enc_dict.items():
+ if (
+ k.endswith(".self_attn.q_proj.weight")
+ or k.endswith(".self_attn.k_proj.weight")
+ or k.endswith(".self_attn.v_proj.weight")
+ ):
+ k_pre = k[: -len(".q_proj.weight")]
+ k_code = k[-len("q_proj.weight")]
+ if k_pre not in capture_qkv_weight:
+ capture_qkv_weight[k_pre] = [None, None, None]
+ capture_qkv_weight[k_pre][code2idx[k_code]] = v
+ continue
+
+ if (
+ k.endswith(".self_attn.q_proj.bias")
+ or k.endswith(".self_attn.k_proj.bias")
+ or k.endswith(".self_attn.v_proj.bias")
+ ):
+ k_pre = k[: -len(".q_proj.bias")]
+ k_code = k[-len("q_proj.bias")]
+ if k_pre not in capture_qkv_bias:
+ capture_qkv_bias[k_pre] = [None, None, None]
+ capture_qkv_bias[k_pre][code2idx[k_code]] = v
+ continue
+
+ relabelled_key = textenc_pattern.sub(lambda m: protected[re.escape(m.group(0))], k)
+ new_state_dict[relabelled_key] = v
+
+ for k_pre, tensors in capture_qkv_weight.items():
+ if None in tensors:
+ raise Exception("CORRUPTED MODEL: one of the q-k-v values for the text encoder was missing")
+ relabelled_key = textenc_pattern.sub(lambda m: protected[re.escape(m.group(0))], k_pre)
+ new_state_dict[relabelled_key + ".in_proj_weight"] = torch.cat(tensors)
+
+ for k_pre, tensors in capture_qkv_bias.items():
+ if None in tensors:
+ raise Exception("CORRUPTED MODEL: one of the q-k-v values for the text encoder was missing")
+ relabelled_key = textenc_pattern.sub(lambda m: protected[re.escape(m.group(0))], k_pre)
+ new_state_dict[relabelled_key + ".in_proj_bias"] = torch.cat(tensors)
+
+ return new_state_dict
+
+
+def convert_text_enc_state_dict(text_enc_dict):
+ return text_enc_dict
+
+
+def load_diffusers(model_path, fp16=True, output_vae=True, output_clip=True, embedding_directory=None):
+ diffusers_unet_conf = json.load(open(osp.join(model_path, "unet/config.json")))
+ diffusers_scheduler_conf = json.load(open(osp.join(model_path, "scheduler/scheduler_config.json")))
+
+ # magic
+ v2 = diffusers_unet_conf["sample_size"] == 96
+ if 'prediction_type' in diffusers_scheduler_conf:
+ v_pred = diffusers_scheduler_conf['prediction_type'] == 'v_prediction'
+
+ if v2:
+ if v_pred:
+ config_path = folder_paths.get_full_path("configs", 'v2-inference-v.yaml')
+ else:
+ config_path = folder_paths.get_full_path("configs", 'v2-inference.yaml')
+ else:
+ config_path = folder_paths.get_full_path("configs", 'v1-inference.yaml')
+
+ with open(config_path, 'r') as stream:
+ config = yaml.safe_load(stream)
+
+ model_config_params = config['model']['params']
+ clip_config = model_config_params['cond_stage_config']
+ scale_factor = model_config_params['scale_factor']
+ vae_config = model_config_params['first_stage_config']
+ vae_config['scale_factor'] = scale_factor
+ model_config_params["unet_config"]["params"]["use_fp16"] = fp16
+
+ unet_path = osp.join(model_path, "unet", "diffusion_pytorch_model.safetensors")
+ vae_path = osp.join(model_path, "vae", "diffusion_pytorch_model.safetensors")
+ text_enc_path = osp.join(model_path, "text_encoder", "model.safetensors")
+
+ # Load models from safetensors if it exists, if it doesn't pytorch
+ if osp.exists(unet_path):
+ unet_state_dict = load_file(unet_path, device="cpu")
+ else:
+ unet_path = osp.join(model_path, "unet", "diffusion_pytorch_model.bin")
+ unet_state_dict = torch.load(unet_path, map_location="cpu")
+
+ if osp.exists(vae_path):
+ vae_state_dict = load_file(vae_path, device="cpu")
+ else:
+ vae_path = osp.join(model_path, "vae", "diffusion_pytorch_model.bin")
+ vae_state_dict = torch.load(vae_path, map_location="cpu")
+
+ if osp.exists(text_enc_path):
+ text_enc_dict = load_file(text_enc_path, device="cpu")
+ else:
+ text_enc_path = osp.join(model_path, "text_encoder", "pytorch_model.bin")
+ text_enc_dict = torch.load(text_enc_path, map_location="cpu")
+
+ # Convert the UNet model
+ unet_state_dict = convert_unet_state_dict(unet_state_dict)
+ unet_state_dict = {"model.diffusion_model." + k: v for k, v in unet_state_dict.items()}
+
+ # Convert the VAE model
+ vae_state_dict = convert_vae_state_dict(vae_state_dict)
+ vae_state_dict = {"first_stage_model." + k: v for k, v in vae_state_dict.items()}
+
+ # Easiest way to identify v2.0 model seems to be that the text encoder (OpenCLIP) is deeper
+ is_v20_model = "text_model.encoder.layers.22.layer_norm2.bias" in text_enc_dict
+
+ if is_v20_model:
+ # Need to add the tag 'transformer' in advance so we can knock it out from the final layer-norm
+ text_enc_dict = {"transformer." + k: v for k, v in text_enc_dict.items()}
+ text_enc_dict = convert_text_enc_state_dict_v20(text_enc_dict)
+ text_enc_dict = {"cond_stage_model.model." + k: v for k, v in text_enc_dict.items()}
+ else:
+ text_enc_dict = convert_text_enc_state_dict(text_enc_dict)
+ text_enc_dict = {"cond_stage_model.transformer." + k: v for k, v in text_enc_dict.items()}
+
+ # Put together new checkpoint
+ sd = {**unet_state_dict, **vae_state_dict, **text_enc_dict}
+
+ clip = None
+ vae = None
+
+ class WeightsLoader(torch.nn.Module):
+ pass
+
+ w = WeightsLoader()
+ load_state_dict_to = []
+ if output_vae:
+ vae = VAE(scale_factor=scale_factor, config=vae_config)
+ w.first_stage_model = vae.first_stage_model
+ load_state_dict_to = [w]
+
+ if output_clip:
+ clip = CLIP(config=clip_config, embedding_directory=embedding_directory)
+ w.cond_stage_model = clip.cond_stage_model
+ load_state_dict_to = [w]
+
+ model = instantiate_from_config(config["model"])
+ model = load_model_weights(model, sd, verbose=False, load_state_dict_to=load_state_dict_to)
+
+ if fp16:
+ model = model.half()
+
+ return ModelPatcher(model), clip, vae
diff --git a/comfy/ldm/modules/attention.py b/comfy/ldm/modules/attention.py
index 07553627c..92b3eca7c 100644
--- a/comfy/ldm/modules/attention.py
+++ b/comfy/ldm/modules/attention.py
@@ -21,6 +21,8 @@ if model_management.xformers_enabled():
import os
_ATTN_PRECISION = os.environ.get("ATTN_PRECISION", "fp32")
+from cli_args import args
+
def exists(val):
return val is not None
@@ -474,7 +476,6 @@ class CrossAttentionPytorch(nn.Module):
return self.to_out(out)
-import sys
if model_management.xformers_enabled():
print("Using xformers cross attention")
CrossAttention = MemoryEfficientCrossAttention
@@ -482,7 +483,7 @@ elif model_management.pytorch_attention_enabled():
print("Using pytorch cross attention")
CrossAttention = CrossAttentionPytorch
else:
- if "--use-split-cross-attention" in sys.argv:
+ if args.use_split_cross_attention:
print("Using split optimization for cross attention")
CrossAttention = CrossAttentionDoggettx
else:
diff --git a/comfy/model_management.py b/comfy/model_management.py
index 052dfb775..2407140fd 100644
--- a/comfy/model_management.py
+++ b/comfy/model_management.py
@@ -1,36 +1,42 @@
+import psutil
+from enum import Enum
+from cli_args import args
-CPU = 0
-NO_VRAM = 1
-LOW_VRAM = 2
-NORMAL_VRAM = 3
-HIGH_VRAM = 4
-MPS = 5
+class VRAMState(Enum):
+ CPU = 0
+ NO_VRAM = 1
+ LOW_VRAM = 2
+ NORMAL_VRAM = 3
+ HIGH_VRAM = 4
+ MPS = 5
-accelerate_enabled = False
-vram_state = NORMAL_VRAM
+# Determine VRAM State
+vram_state = VRAMState.NORMAL_VRAM
+set_vram_to = VRAMState.NORMAL_VRAM
total_vram = 0
total_vram_available_mb = -1
-import sys
-import psutil
-
-forced_cpu = "--cpu" in sys.argv
-
-set_vram_to = NORMAL_VRAM
+accelerate_enabled = False
+xpu_available = False
try:
import torch
- total_vram = torch.cuda.mem_get_info(torch.cuda.current_device())[1] / (1024 * 1024)
+ try:
+ import intel_extension_for_pytorch as ipex
+ if torch.xpu.is_available():
+ xpu_available = True
+ total_vram = torch.xpu.get_device_properties(torch.xpu.current_device()).total_memory / (1024 * 1024)
+ except:
+ total_vram = torch.cuda.mem_get_info(torch.cuda.current_device())[1] / (1024 * 1024)
total_ram = psutil.virtual_memory().total / (1024 * 1024)
- forced_normal_vram = "--normalvram" in sys.argv
- if not forced_normal_vram and not forced_cpu:
+ if not args.normalvram and not args.cpu:
if total_vram <= 4096:
print("Trying to enable lowvram mode because your GPU seems to have 4GB or less. If you don't want this use: --normalvram")
- set_vram_to = LOW_VRAM
+ set_vram_to = VRAMState.LOW_VRAM
elif total_vram > total_ram * 1.1 and total_vram > 14336:
print("Enabling highvram mode because your GPU has more vram than your computer has ram. If you don't want this use: --normalvram")
- vram_state = HIGH_VRAM
+ vram_state = VRAMState.HIGH_VRAM
except:
pass
@@ -39,34 +45,37 @@ try:
except:
OOM_EXCEPTION = Exception
-if "--disable-xformers" in sys.argv:
- XFORMERS_IS_AVAILBLE = False
+if args.disable_xformers:
+ XFORMERS_IS_AVAILABLE = False
else:
try:
import xformers
import xformers.ops
- XFORMERS_IS_AVAILBLE = True
+ XFORMERS_IS_AVAILABLE = True
except:
- XFORMERS_IS_AVAILBLE = False
+ XFORMERS_IS_AVAILABLE = False
-ENABLE_PYTORCH_ATTENTION = False
-if "--use-pytorch-cross-attention" in sys.argv:
+ENABLE_PYTORCH_ATTENTION = args.use_pytorch_cross_attention
+if ENABLE_PYTORCH_ATTENTION:
torch.backends.cuda.enable_math_sdp(True)
torch.backends.cuda.enable_flash_sdp(True)
torch.backends.cuda.enable_mem_efficient_sdp(True)
- ENABLE_PYTORCH_ATTENTION = True
- XFORMERS_IS_AVAILBLE = False
+ XFORMERS_IS_AVAILABLE = False
+
+if args.lowvram:
+ set_vram_to = VRAMState.LOW_VRAM
+elif args.novram:
+ set_vram_to = VRAMState.NO_VRAM
+elif args.highvram:
+ vram_state = VRAMState.HIGH_VRAM
+
+FORCE_FP32 = False
+if args.force_fp32:
+ print("Forcing FP32, if this improves things please report it.")
+ FORCE_FP32 = True
-if "--lowvram" in sys.argv:
- set_vram_to = LOW_VRAM
-if "--novram" in sys.argv:
- set_vram_to = NO_VRAM
-if "--highvram" in sys.argv:
- vram_state = HIGH_VRAM
-
-
-if set_vram_to == LOW_VRAM or set_vram_to == NO_VRAM:
+if set_vram_to in (VRAMState.LOW_VRAM, VRAMState.NO_VRAM):
try:
import accelerate
accelerate_enabled = True
@@ -81,14 +90,14 @@ if set_vram_to == LOW_VRAM or set_vram_to == NO_VRAM:
try:
if torch.backends.mps.is_available():
- vram_state = MPS
+ vram_state = VRAMState.MPS
except:
pass
-if forced_cpu:
- vram_state = CPU
+if args.cpu:
+ vram_state = VRAMState.CPU
-print("Set vram state to:", ["CPU", "NO VRAM", "LOW VRAM", "NORMAL VRAM", "HIGH VRAM", "MPS"][vram_state])
+print(f"Set vram state to: {vram_state.name}")
current_loaded_model = None
@@ -109,12 +118,12 @@ def unload_model():
model_accelerated = False
#never unload models from GPU on high vram
- if vram_state != HIGH_VRAM:
+ if vram_state != VRAMState.HIGH_VRAM:
current_loaded_model.model.cpu()
current_loaded_model.unpatch_model()
current_loaded_model = None
- if vram_state != HIGH_VRAM:
+ if vram_state != VRAMState.HIGH_VRAM:
if len(current_gpu_controlnets) > 0:
for n in current_gpu_controlnets:
n.cpu()
@@ -135,32 +144,32 @@ def load_model_gpu(model):
model.unpatch_model()
raise e
current_loaded_model = model
- if vram_state == CPU:
+ if vram_state == VRAMState.CPU:
pass
- elif vram_state == MPS:
+ elif vram_state == VRAMState.MPS:
mps_device = torch.device("mps")
real_model.to(mps_device)
pass
- elif vram_state == NORMAL_VRAM or vram_state == HIGH_VRAM:
+ elif vram_state == VRAMState.NORMAL_VRAM or vram_state == VRAMState.HIGH_VRAM:
model_accelerated = False
- real_model.cuda()
+ real_model.to(get_torch_device())
else:
- if vram_state == NO_VRAM:
+ if vram_state == VRAMState.NO_VRAM:
device_map = accelerate.infer_auto_device_map(real_model, max_memory={0: "256MiB", "cpu": "16GiB"})
- elif vram_state == LOW_VRAM:
+ elif vram_state == VRAMState.LOW_VRAM:
device_map = accelerate.infer_auto_device_map(real_model, max_memory={0: "{}MiB".format(total_vram_available_mb), "cpu": "16GiB"})
- accelerate.dispatch_model(real_model, device_map=device_map, main_device="cuda")
+ accelerate.dispatch_model(real_model, device_map=device_map, main_device=get_torch_device())
model_accelerated = True
return current_loaded_model
def load_controlnet_gpu(models):
global current_gpu_controlnets
global vram_state
- if vram_state == CPU:
+ if vram_state == VRAMState.CPU:
return
- if vram_state == LOW_VRAM or vram_state == NO_VRAM:
+ if vram_state == VRAMState.LOW_VRAM or vram_state == VRAMState.NO_VRAM:
#don't load controlnets like this if low vram because they will be loaded right before running and unloaded right after
return
@@ -176,23 +185,27 @@ def load_controlnet_gpu(models):
def load_if_low_vram(model):
global vram_state
- if vram_state == LOW_VRAM or vram_state == NO_VRAM:
- return model.cuda()
+ if vram_state == VRAMState.LOW_VRAM or vram_state == VRAMState.NO_VRAM:
+ return model.to(get_torch_device())
return model
def unload_if_low_vram(model):
global vram_state
- if vram_state == LOW_VRAM or vram_state == NO_VRAM:
+ if vram_state == VRAMState.LOW_VRAM or vram_state == VRAMState.NO_VRAM:
return model.cpu()
return model
def get_torch_device():
- if vram_state == MPS:
+ global xpu_available
+ if vram_state == VRAMState.MPS:
return torch.device("mps")
- if vram_state == CPU:
+ if vram_state == VRAMState.CPU:
return torch.device("cpu")
else:
- return torch.cuda.current_device()
+ if xpu_available:
+ return torch.device("xpu")
+ else:
+ return torch.cuda.current_device()
def get_autocast_device(dev):
if hasattr(dev, 'type'):
@@ -201,9 +214,9 @@ def get_autocast_device(dev):
def xformers_enabled():
- if vram_state == CPU:
+ if vram_state == VRAMState.CPU:
return False
- return XFORMERS_IS_AVAILBLE
+ return XFORMERS_IS_AVAILABLE
def xformers_enabled_vae():
@@ -222,6 +235,7 @@ def pytorch_attention_enabled():
return ENABLE_PYTORCH_ATTENTION
def get_free_memory(dev=None, torch_free_too=False):
+ global xpu_available
if dev is None:
dev = get_torch_device()
@@ -229,12 +243,16 @@ def get_free_memory(dev=None, torch_free_too=False):
mem_free_total = psutil.virtual_memory().available
mem_free_torch = mem_free_total
else:
- stats = torch.cuda.memory_stats(dev)
- mem_active = stats['active_bytes.all.current']
- mem_reserved = stats['reserved_bytes.all.current']
- mem_free_cuda, _ = torch.cuda.mem_get_info(dev)
- mem_free_torch = mem_reserved - mem_active
- mem_free_total = mem_free_cuda + mem_free_torch
+ if xpu_available:
+ mem_free_total = torch.xpu.get_device_properties(dev).total_memory - torch.xpu.memory_allocated(dev)
+ mem_free_torch = mem_free_total
+ else:
+ stats = torch.cuda.memory_stats(dev)
+ mem_active = stats['active_bytes.all.current']
+ mem_reserved = stats['reserved_bytes.all.current']
+ mem_free_cuda, _ = torch.cuda.mem_get_info(dev)
+ mem_free_torch = mem_reserved - mem_active
+ mem_free_total = mem_free_cuda + mem_free_torch
if torch_free_too:
return (mem_free_total, mem_free_torch)
@@ -243,7 +261,7 @@ def get_free_memory(dev=None, torch_free_too=False):
def maximum_batch_area():
global vram_state
- if vram_state == NO_VRAM:
+ if vram_state == VRAMState.NO_VRAM:
return 0
memory_free = get_free_memory() / (1024 * 1024)
@@ -252,14 +270,18 @@ def maximum_batch_area():
def cpu_mode():
global vram_state
- return vram_state == CPU
+ return vram_state == VRAMState.CPU
def mps_mode():
global vram_state
- return vram_state == MPS
+ return vram_state == VRAMState.MPS
def should_use_fp16():
- if cpu_mode() or mps_mode():
+ global xpu_available
+ if FORCE_FP32:
+ return False
+
+ if cpu_mode() or mps_mode() or xpu_available:
return False #TODO ?
if torch.cuda.is_bf16_supported():
diff --git a/custom_nodes/example_node.py.example b/custom_nodes/example_node.py.example
index fb8172648..175202aeb 100644
--- a/custom_nodes/example_node.py.example
+++ b/custom_nodes/example_node.py.example
@@ -88,3 +88,8 @@ class Example:
NODE_CLASS_MAPPINGS = {
"Example": Example
}
+
+# A dictionary that contains the friendly/humanly readable titles for the nodes
+NODE_DISPLAY_NAME_MAPPINGS = {
+ "Example": "Example Node"
+}
diff --git a/folder_paths.py b/folder_paths.py
index f13e4895f..ab3359347 100644
--- a/folder_paths.py
+++ b/folder_paths.py
@@ -23,6 +23,7 @@ folder_names_and_paths["clip"] = ([os.path.join(models_dir, "clip")], supported_
folder_names_and_paths["clip_vision"] = ([os.path.join(models_dir, "clip_vision")], supported_pt_extensions)
folder_names_and_paths["style_models"] = ([os.path.join(models_dir, "style_models")], supported_pt_extensions)
folder_names_and_paths["embeddings"] = ([os.path.join(models_dir, "embeddings")], supported_pt_extensions)
+folder_names_and_paths["diffusers"] = ([os.path.join(models_dir, "diffusers")], ["folder"])
folder_names_and_paths["controlnet"] = ([os.path.join(models_dir, "controlnet"), os.path.join(models_dir, "t2i_adapter")], supported_pt_extensions)
folder_names_and_paths["upscale_models"] = ([os.path.join(models_dir, "upscale_models")], supported_pt_extensions)
diff --git a/main.py b/main.py
index a3549b86f..9c0a3d8a1 100644
--- a/main.py
+++ b/main.py
@@ -1,57 +1,33 @@
-import os
-import sys
-import shutil
-
-import threading
import asyncio
+import itertools
+import os
+import shutil
+import threading
+
+from comfy.cli_args import args
if os.name == "nt":
import logging
logging.getLogger("xformers").addFilter(lambda record: 'A matching Triton is not available' not in record.getMessage())
if __name__ == "__main__":
- if '--help' in sys.argv:
- print()
- print("Valid Command line Arguments:")
- print("\t--listen [ip]\t\t\tListen on ip or 0.0.0.0 if none given so the UI can be accessed from other computers.")
- print("\t--port 8188\t\t\tSet the listen port.")
- print()
- print("\t--extra-model-paths-config file.yaml\tload an extra_model_paths.yaml file.")
- print("\t--output-directory path/to/output\tSet the ComfyUI output directory.")
- print()
- print()
- print("\t--dont-upcast-attention\t\tDisable upcasting of attention \n\t\t\t\t\tcan boost speed but increase the chances of black images.\n")
- print("\t--use-split-cross-attention\tUse the split cross attention optimization instead of the sub-quadratic one.\n\t\t\t\t\tIgnored when xformers is used.")
- print("\t--use-pytorch-cross-attention\tUse the new pytorch 2.0 cross attention function.")
- print("\t--disable-xformers\t\tdisables xformers")
- print("\t--cuda-device 1\t\tSet the id of the cuda device this instance will use.")
- print()
- print("\t--highvram\t\t\tBy default models will be unloaded to CPU memory after being used.\n\t\t\t\t\tThis option keeps them in GPU memory.\n")
- print("\t--normalvram\t\t\tUsed to force normal vram use if lowvram gets automatically enabled.")
- print("\t--lowvram\t\t\tSplit the unet in parts to use less vram.")
- print("\t--novram\t\t\tWhen lowvram isn't enough.")
- print()
- print("\t--cpu\t\t\tTo use the CPU for everything (slow).")
- exit()
-
- if '--dont-upcast-attention' in sys.argv:
+ if args.dont_upcast_attention:
print("disabling upcasting of attention")
os.environ['ATTN_PRECISION'] = "fp16"
- try:
- index = sys.argv.index('--cuda-device')
- device = sys.argv[index + 1]
- os.environ['CUDA_VISIBLE_DEVICES'] = device
- print("Set cuda device to:", device)
- except:
- pass
+ if args.cuda_device is not None:
+ os.environ['CUDA_VISIBLE_DEVICES'] = str(args.cuda_device)
+ print("Set cuda device to:", args.cuda_device)
+
-from nodes import init_custom_nodes
-import execution
-import server
-import folder_paths
import yaml
+import execution
+import folder_paths
+import server
+from nodes import init_custom_nodes
+
+
def prompt_worker(q, server):
e = execution.PromptExecutor(server)
while True:
@@ -110,51 +86,31 @@ if __name__ == "__main__":
hijack_progress(server)
threading.Thread(target=prompt_worker, daemon=True, args=(q,server,)).start()
- try:
- address = '0.0.0.0'
- p_index = sys.argv.index('--listen')
- try:
- ip = sys.argv[p_index + 1]
- if ip[:2] != '--':
- address = ip
- except:
- pass
- except:
- address = '127.0.0.1'
- dont_print = False
- if '--dont-print-server' in sys.argv:
- dont_print = True
+ address = args.listen
+
+ dont_print = args.dont_print_server
extra_model_paths_config_path = os.path.join(os.path.dirname(os.path.realpath(__file__)), "extra_model_paths.yaml")
if os.path.isfile(extra_model_paths_config_path):
load_extra_path_config(extra_model_paths_config_path)
- if '--extra-model-paths-config' in sys.argv:
- indices = [(i + 1) for i in range(len(sys.argv) - 1) if sys.argv[i] == '--extra-model-paths-config']
- for i in indices:
- load_extra_path_config(sys.argv[i])
+ if args.extra_model_paths_config:
+ for config_path in itertools.chain(*args.extra_model_paths_config):
+ load_extra_path_config(config_path)
- try:
- output_dir = sys.argv[sys.argv.index('--output-directory') + 1]
- output_dir = os.path.abspath(output_dir)
- print("setting output directory to:", output_dir)
+ if args.output_directory:
+ output_dir = os.path.abspath(args.output_directory)
+ print(f"Setting output directory to: {output_dir}")
folder_paths.set_output_directory(output_dir)
- except:
- pass
- port = 8188
- try:
- p_index = sys.argv.index('--port')
- port = int(sys.argv[p_index + 1])
- except:
- pass
+ port = args.port
- if '--quick-test-for-ci' in sys.argv:
+ if args.quick_test_for_ci:
exit(0)
call_on_start = None
- if "--windows-standalone-build" in sys.argv:
+ if args.windows_standalone_build:
def startup_server(address, port):
import webbrowser
webbrowser.open("http://{}:{}".format(address, port))
diff --git a/models/diffusers/put_diffusers_models_here b/models/diffusers/put_diffusers_models_here
new file mode 100644
index 000000000..e69de29bb
diff --git a/nodes.py b/nodes.py
index 187d54a11..14a73bcd7 100644
--- a/nodes.py
+++ b/nodes.py
@@ -4,16 +4,17 @@ import os
import sys
import json
import hashlib
-import copy
import traceback
from PIL import Image
from PIL.PngImagePlugin import PngInfo
import numpy as np
+
sys.path.insert(0, os.path.join(os.path.dirname(os.path.realpath(__file__)), "comfy"))
+import comfy.diffusers_convert
import comfy.samplers
import comfy.sd
import comfy.utils
@@ -219,6 +220,30 @@ class CheckpointLoaderSimple:
out = comfy.sd.load_checkpoint_guess_config(ckpt_path, output_vae=True, output_clip=True, embedding_directory=folder_paths.get_folder_paths("embeddings"))
return out
+class DiffusersLoader:
+ @classmethod
+ def INPUT_TYPES(cls):
+ paths = []
+ for search_path in folder_paths.get_folder_paths("diffusers"):
+ if os.path.exists(search_path):
+ paths += next(os.walk(search_path))[1]
+ return {"required": {"model_path": (paths,), }}
+ RETURN_TYPES = ("MODEL", "CLIP", "VAE")
+ FUNCTION = "load_checkpoint"
+
+ CATEGORY = "advanced/loaders"
+
+ def load_checkpoint(self, model_path, output_vae=True, output_clip=True):
+ for search_path in folder_paths.get_folder_paths("diffusers"):
+ if os.path.exists(search_path):
+ paths = next(os.walk(search_path))[1]
+ if model_path in paths:
+ model_path = os.path.join(search_path, model_path)
+ break
+
+ return comfy.diffusers_convert.load_diffusers(model_path, fp16=model_management.should_use_fp16(), output_vae=output_vae, output_clip=output_clip, embedding_directory=folder_paths.get_folder_paths("embeddings"))
+
+
class unCLIPCheckpointLoader:
@classmethod
def INPUT_TYPES(s):
@@ -1076,6 +1101,55 @@ NODE_CLASS_MAPPINGS = {
"TomePatchModel": TomePatchModel,
"unCLIPCheckpointLoader": unCLIPCheckpointLoader,
"CheckpointLoader": CheckpointLoader,
+ "DiffusersLoader": DiffusersLoader,
+}
+
+NODE_DISPLAY_NAME_MAPPINGS = {
+ # Sampling
+ "KSampler": "KSampler",
+ "KSamplerAdvanced": "KSampler (Advanced)",
+ # Loaders
+ "CheckpointLoader": "Load Checkpoint (With Config)",
+ "CheckpointLoaderSimple": "Load Checkpoint",
+ "VAELoader": "Load VAE",
+ "LoraLoader": "Load LoRA",
+ "CLIPLoader": "Load CLIP",
+ "ControlNetLoader": "Load ControlNet Model",
+ "DiffControlNetLoader": "Load ControlNet Model (diff)",
+ "StyleModelLoader": "Load Style Model",
+ "CLIPVisionLoader": "Load CLIP Vision",
+ "UpscaleModelLoader": "Load Upscale Model",
+ # Conditioning
+ "CLIPVisionEncode": "CLIP Vision Encode",
+ "StyleModelApply": "Apply Style Model",
+ "CLIPTextEncode": "CLIP Text Encode (Prompt)",
+ "CLIPSetLastLayer": "CLIP Set Last Layer",
+ "ConditioningCombine": "Conditioning (Combine)",
+ "ConditioningSetArea": "Conditioning (Set Area)",
+ "ControlNetApply": "Apply ControlNet",
+ # Latent
+ "VAEEncodeForInpaint": "VAE Encode (for Inpainting)",
+ "SetLatentNoiseMask": "Set Latent Noise Mask",
+ "VAEDecode": "VAE Decode",
+ "VAEEncode": "VAE Encode",
+ "LatentRotate": "Rotate Latent",
+ "LatentFlip": "Flip Latent",
+ "LatentCrop": "Crop Latent",
+ "EmptyLatentImage": "Empty Latent Image",
+ "LatentUpscale": "Upscale Latent",
+ "LatentComposite": "Latent Composite",
+ # Image
+ "SaveImage": "Save Image",
+ "PreviewImage": "Preview Image",
+ "LoadImage": "Load Image",
+ "LoadImageMask": "Load Image (as Mask)",
+ "ImageScale": "Upscale Image",
+ "ImageUpscaleWithModel": "Upscale Image (using Model)",
+ "ImageInvert": "Invert Image",
+ "ImagePadForOutpaint": "Pad Image for Outpainting",
+ # _for_testing
+ "VAEDecodeTiled": "VAE Decode (Tiled)",
+ "VAEEncodeTiled": "VAE Encode (Tiled)",
}
def load_custom_node(module_path):
@@ -1093,6 +1167,8 @@ def load_custom_node(module_path):
module_spec.loader.exec_module(module)
if hasattr(module, "NODE_CLASS_MAPPINGS") and getattr(module, "NODE_CLASS_MAPPINGS") is not None:
NODE_CLASS_MAPPINGS.update(module.NODE_CLASS_MAPPINGS)
+ if hasattr(module, "NODE_DISPLAY_NAME_MAPPINGS") and getattr(module, "NODE_DISPLAY_NAME_MAPPINGS") is not None:
+ NODE_DISPLAY_NAME_MAPPINGS.update(module.NODE_DISPLAY_NAME_MAPPINGS)
else:
print(f"Skip {module_path} module for custom nodes due to the lack of NODE_CLASS_MAPPINGS.")
except Exception as e:
diff --git a/requirements.txt b/requirements.txt
index 3b4040a29..0527b31df 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -4,7 +4,7 @@ torchsde
einops
open-clip-torch
transformers>=4.25.1
-safetensors
+safetensors>=0.3.0
pytorch_lightning
aiohttp
accelerate
diff --git a/server.py b/server.py
index 840d9a4e7..b5403670f 100644
--- a/server.py
+++ b/server.py
@@ -18,6 +18,7 @@ except ImportError:
sys.exit()
import mimetypes
+from comfy.cli_args import args
@web.middleware
@@ -27,6 +28,23 @@ async def cache_control(request: web.Request, handler):
response.headers.setdefault('Cache-Control', 'no-cache')
return response
+def create_cors_middleware(allowed_origin: str):
+ @web.middleware
+ async def cors_middleware(request: web.Request, handler):
+ if request.method == "OPTIONS":
+ # Pre-flight request. Reply successfully:
+ response = web.Response()
+ else:
+ response = await handler(request)
+
+ response.headers['Access-Control-Allow-Origin'] = allowed_origin
+ response.headers['Access-Control-Allow-Methods'] = 'POST, GET, DELETE, PUT, OPTIONS'
+ response.headers['Access-Control-Allow-Headers'] = 'Content-Type, Authorization'
+ response.headers['Access-Control-Allow-Credentials'] = 'true'
+ return response
+
+ return cors_middleware
+
class PromptServer():
def __init__(self, loop):
PromptServer.instance = self
@@ -37,7 +55,12 @@ class PromptServer():
self.loop = loop
self.messages = asyncio.Queue()
self.number = 0
- self.app = web.Application(client_max_size=20971520, middlewares=[cache_control])
+
+ middlewares = [cache_control]
+ if args.enable_cors_header:
+ middlewares.append(create_cors_middleware(args.enable_cors_header))
+
+ self.app = web.Application(client_max_size=20971520, middlewares=middlewares)
self.sockets = dict()
self.web_root = os.path.join(os.path.dirname(
os.path.realpath(__file__)), "web")
@@ -154,7 +177,8 @@ class PromptServer():
info['input'] = obj_class.INPUT_TYPES()
info['output'] = obj_class.RETURN_TYPES
info['output_name'] = obj_class.RETURN_NAMES if hasattr(obj_class, 'RETURN_NAMES') else info['output']
- info['name'] = x #TODO
+ info['name'] = x
+ info['display_name'] = nodes.NODE_DISPLAY_NAME_MAPPINGS[x] if x in nodes.NODE_DISPLAY_NAME_MAPPINGS.keys() else x
info['description'] = ''
info['category'] = 'sd'
if hasattr(obj_class, 'CATEGORY'):
diff --git a/web/extensions/core/colorPalette.js b/web/extensions/core/colorPalette.js
index e54bc2a38..a08d46684 100644
--- a/web/extensions/core/colorPalette.js
+++ b/web/extensions/core/colorPalette.js
@@ -21,28 +21,74 @@ const colorPalettes = {
"MODEL": "#B39DDB", // light lavender-purple
"STYLE_MODEL": "#C2FFAE", // light green-yellow
"VAE": "#FF6E6E", // bright red
- }
- }
+ },
+ "litegraph_base": {
+ "NODE_TITLE_COLOR": "#999",
+ "NODE_SELECTED_TITLE_COLOR": "#FFF",
+ "NODE_TEXT_SIZE": 14,
+ "NODE_TEXT_COLOR": "#AAA",
+ "NODE_SUBTEXT_SIZE": 12,
+ "NODE_DEFAULT_COLOR": "#333",
+ "NODE_DEFAULT_BGCOLOR": "#353535",
+ "NODE_DEFAULT_BOXCOLOR": "#666",
+ "NODE_DEFAULT_SHAPE": "box",
+ "NODE_BOX_OUTLINE_COLOR": "#FFF",
+ "DEFAULT_SHADOW_COLOR": "rgba(0,0,0,0.5)",
+ "DEFAULT_GROUP_FONT": 24,
+
+ "WIDGET_BGCOLOR": "#222",
+ "WIDGET_OUTLINE_COLOR": "#666",
+ "WIDGET_TEXT_COLOR": "#DDD",
+ "WIDGET_SECONDARY_TEXT_COLOR": "#999",
+
+ "LINK_COLOR": "#9A9",
+ "EVENT_LINK_COLOR": "#A86",
+ "CONNECTING_LINK_COLOR": "#AFA",
+ },
+ },
},
- "palette_2": {
- "id": "palette_2",
- "name": "Palette 2",
+ "solarized": {
+ "id": "solarized",
+ "name": "Solarized",
"colors": {
"node_slot": {
- "CLIP": "#556B2F", // Dark Olive Green
- "CLIP_VISION": "#4B0082", // Indigo
- "CLIP_VISION_OUTPUT": "#006400", // Green
- "CONDITIONING": "#FF1493", // Deep Pink
- "CONTROL_NET": "#8B4513", // Saddle Brown
- "IMAGE": "#8B0000", // Dark Red
- "LATENT": "#00008B", // Dark Blue
- "MASK": "#2F4F4F", // Dark Slate Grey
- "MODEL": "#FF8C00", // Dark Orange
- "STYLE_MODEL": "#004A4A", // Sherpa Blue
- "UPSCALE_MODEL": "#4A004A", // Tyrian Purple
- "VAE": "#4F394F", // Loulou
- }
- }
+ "CLIP": "#859900", // Green
+ "CLIP_VISION": "#6c71c4", // Indigo
+ "CLIP_VISION_OUTPUT": "#859900", // Green
+ "CONDITIONING": "#d33682", // Magenta
+ "CONTROL_NET": "#cb4b16", // Orange
+ "IMAGE": "#dc322f", // Red
+ "LATENT": "#268bd2", // Blue
+ "MASK": "#073642", // Base02
+ "MODEL": "#cb4b16", // Orange
+ "STYLE_MODEL": "#073642", // Base02
+ "UPSCALE_MODEL": "#6c71c4", // Indigo
+ "VAE": "#586e75", // Base1
+ },
+ "litegraph_base": {
+ "NODE_TITLE_COLOR": "#fdf6e3",
+ "NODE_SELECTED_TITLE_COLOR": "#b58900",
+ "NODE_TEXT_SIZE": 14,
+ "NODE_TEXT_COLOR": "#657b83",
+ "NODE_SUBTEXT_SIZE": 12,
+ "NODE_DEFAULT_COLOR": "#586e75",
+ "NODE_DEFAULT_BGCOLOR": "#073642",
+ "NODE_DEFAULT_BOXCOLOR": "#839496",
+ "NODE_DEFAULT_SHAPE": "box",
+ "NODE_BOX_OUTLINE_COLOR": "#fdf6e3",
+ "DEFAULT_SHADOW_COLOR": "rgba(0,0,0,0.5)",
+ "DEFAULT_GROUP_FONT": 24,
+
+ "WIDGET_BGCOLOR": "#002b36",
+ "WIDGET_OUTLINE_COLOR": "#839496",
+ "WIDGET_TEXT_COLOR": "#fdf6e3",
+ "WIDGET_SECONDARY_TEXT_COLOR": "#93a1a1",
+
+ "LINK_COLOR": "#2aa198",
+ "EVENT_LINK_COLOR": "#268bd2",
+ "CONNECTING_LINK_COLOR": "#859900",
+ },
+ },
}
};
@@ -192,8 +238,20 @@ app.registerExtension({
if (colorPalette.colors) {
if (colorPalette.colors.node_slot) {
Object.assign(app.canvas.default_connection_color_byType, colorPalette.colors.node_slot);
- app.canvas.draw(true, true);
+ Object.assign(LGraphCanvas.link_type_colors, colorPalette.colors.node_slot);
}
+ if (colorPalette.colors.litegraph_base) {
+ // Everything updates correctly in the loop, except the Node Title and Link Color for some reason
+ app.canvas.node_title_color = colorPalette.colors.litegraph_base.NODE_TITLE_COLOR;
+ app.canvas.default_link_color = colorPalette.colors.litegraph_base.LINK_COLOR;
+
+ for (const key in colorPalette.colors.litegraph_base) {
+ if (colorPalette.colors.litegraph_base.hasOwnProperty(key) && LiteGraph.hasOwnProperty(key)) {
+ LiteGraph[key] = colorPalette.colors.litegraph_base[key];
+ }
+ }
+ }
+ app.canvas.draw(true, true);
}
};
diff --git a/web/index.html b/web/index.html
index 86156a7f8..bb79433ce 100644
--- a/web/index.html
+++ b/web/index.html
@@ -2,6 +2,7 @@
+ ComfyUI
diff --git a/web/lib/litegraph.core.js b/web/lib/litegraph.core.js
index 862d59067..c3efa22a9 100644
--- a/web/lib/litegraph.core.js
+++ b/web/lib/litegraph.core.js
@@ -89,6 +89,7 @@
NO_TITLE: 1,
TRANSPARENT_TITLE: 2,
AUTOHIDE_TITLE: 3,
+ VERTICAL_LAYOUT: "vertical", // arrange nodes vertically
proxy: null, //used to redirect calls
node_images_path: "",
@@ -125,14 +126,14 @@
registered_slot_out_types: {}, // slot types for nodeclass
slot_types_in: [], // slot types IN
slot_types_out: [], // slot types OUT
- slot_types_default_in: [], // specify for each IN slot type a(/many) deafult node(s), use single string, array, or object (with node, title, parameters, ..) like for search
- slot_types_default_out: [], // specify for each OUT slot type a(/many) deafult node(s), use single string, array, or object (with node, title, parameters, ..) like for search
+ slot_types_default_in: [], // specify for each IN slot type a(/many) default node(s), use single string, array, or object (with node, title, parameters, ..) like for search
+ slot_types_default_out: [], // specify for each OUT slot type a(/many) default node(s), use single string, array, or object (with node, title, parameters, ..) like for search
alt_drag_do_clone_nodes: false, // [true!] very handy, ALT click to clone and drag the new node
do_add_triggers_slots: false, // [true!] will create and connect event slots when using action/events connections, !WILL CHANGE node mode when using onTrigger (enable mode colors), onExecuted does not need this
- allow_multi_output_for_events: true, // [false!] being events, it is strongly reccomended to use them sequentually, one by one
+ allow_multi_output_for_events: true, // [false!] being events, it is strongly reccomended to use them sequentially, one by one
middle_click_slot_add_default_node: false, //[true!] allows to create and connect a ndoe clicking with the third button (wheel)
@@ -158,80 +159,67 @@
console.log("Node registered: " + type);
}
- var categories = type.split("/");
- var classname = base_class.name;
+ const classname = base_class.name;
- var pos = type.lastIndexOf("/");
- base_class.category = type.substr(0, pos);
+ const pos = type.lastIndexOf("/");
+ base_class.category = type.substring(0, pos);
if (!base_class.title) {
base_class.title = classname;
}
- //info.name = name.substr(pos+1,name.length - pos);
//extend class
- if (base_class.prototype) {
- //is a class
- for (var i in LGraphNode.prototype) {
- if (!base_class.prototype[i]) {
- base_class.prototype[i] = LGraphNode.prototype[i];
- }
+ for (var i in LGraphNode.prototype) {
+ if (!base_class.prototype[i]) {
+ base_class.prototype[i] = LGraphNode.prototype[i];
}
}
- var prev = this.registered_node_types[type];
- if(prev)
- console.log("replacing node type: " + type);
- else
- {
- if( !Object.hasOwnProperty( base_class.prototype, "shape") )
- Object.defineProperty(base_class.prototype, "shape", {
- set: function(v) {
- switch (v) {
- case "default":
- delete this._shape;
- break;
- case "box":
- this._shape = LiteGraph.BOX_SHAPE;
- break;
- case "round":
- this._shape = LiteGraph.ROUND_SHAPE;
- break;
- case "circle":
- this._shape = LiteGraph.CIRCLE_SHAPE;
- break;
- case "card":
- this._shape = LiteGraph.CARD_SHAPE;
- break;
- default:
- this._shape = v;
- }
- },
- get: function(v) {
- return this._shape;
- },
- enumerable: true,
- configurable: true
- });
+ const prev = this.registered_node_types[type];
+ if(prev) {
+ console.log("replacing node type: " + type);
+ }
+ if( !Object.prototype.hasOwnProperty.call( base_class.prototype, "shape") ) {
+ Object.defineProperty(base_class.prototype, "shape", {
+ set: function(v) {
+ switch (v) {
+ case "default":
+ delete this._shape;
+ break;
+ case "box":
+ this._shape = LiteGraph.BOX_SHAPE;
+ break;
+ case "round":
+ this._shape = LiteGraph.ROUND_SHAPE;
+ break;
+ case "circle":
+ this._shape = LiteGraph.CIRCLE_SHAPE;
+ break;
+ case "card":
+ this._shape = LiteGraph.CARD_SHAPE;
+ break;
+ default:
+ this._shape = v;
+ }
+ },
+ get: function() {
+ return this._shape;
+ },
+ enumerable: true,
+ configurable: true
+ });
+
- //warnings
- if (base_class.prototype.onPropertyChange) {
- console.warn(
- "LiteGraph node class " +
- type +
- " has onPropertyChange method, it must be called onPropertyChanged with d at the end"
- );
- }
-
- //used to know which nodes create when dragging files to the canvas
- if (base_class.supported_extensions) {
- for (var i in base_class.supported_extensions) {
- var ext = base_class.supported_extensions[i];
- if(ext && ext.constructor === String)
- this.node_types_by_file_extension[ ext.toLowerCase() ] = base_class;
- }
- }
- }
+ //used to know which nodes to create when dragging files to the canvas
+ if (base_class.supported_extensions) {
+ for (let i in base_class.supported_extensions) {
+ const ext = base_class.supported_extensions[i];
+ if(ext && ext.constructor === String) {
+ this.node_types_by_file_extension[ ext.toLowerCase() ] = base_class;
+ }
+ }
+ }
+ }
this.registered_node_types[type] = base_class;
if (base_class.constructor.name) {
@@ -252,19 +240,11 @@
" has onPropertyChange method, it must be called onPropertyChanged with d at the end"
);
}
-
- //used to know which nodes create when dragging files to the canvas
- if (base_class.supported_extensions) {
- for (var i=0; i < base_class.supported_extensions.length; i++) {
- var ext = base_class.supported_extensions[i];
- if(ext && ext.constructor === String)
- this.node_types_by_file_extension[ ext.toLowerCase() ] = base_class;
- }
- }
- // TODO one would want to know input and ouput :: this would allow trought registerNodeAndSlotType to get all the slots types
- //console.debug("Registering "+type);
- if (this.auto_load_slot_types) nodeTmp = new base_class(base_class.title || "tmpnode");
+ // TODO one would want to know input and ouput :: this would allow through registerNodeAndSlotType to get all the slots types
+ if (this.auto_load_slot_types) {
+ new base_class(base_class.title || "tmpnode");
+ }
},
/**
@@ -1260,37 +1240,39 @@
* Positions every node in a more readable manner
* @method arrange
*/
- LGraph.prototype.arrange = function(margin) {
+ LGraph.prototype.arrange = function (margin, layout) {
margin = margin || 100;
- var nodes = this.computeExecutionOrder(false, true);
- var columns = [];
- for (var i = 0; i < nodes.length; ++i) {
- var node = nodes[i];
- var col = node._level || 1;
+ const nodes = this.computeExecutionOrder(false, true);
+ const columns = [];
+ for (let i = 0; i < nodes.length; ++i) {
+ const node = nodes[i];
+ const col = node._level || 1;
if (!columns[col]) {
columns[col] = [];
}
columns[col].push(node);
}
- var x = margin;
+ let x = margin;
- for (var i = 0; i < columns.length; ++i) {
- var column = columns[i];
+ for (let i = 0; i < columns.length; ++i) {
+ const column = columns[i];
if (!column) {
continue;
}
- var max_size = 100;
- var y = margin + LiteGraph.NODE_TITLE_HEIGHT;
- for (var j = 0; j < column.length; ++j) {
- var node = column[j];
- node.pos[0] = x;
- node.pos[1] = y;
- if (node.size[0] > max_size) {
- max_size = node.size[0];
+ let max_size = 100;
+ let y = margin + LiteGraph.NODE_TITLE_HEIGHT;
+ for (let j = 0; j < column.length; ++j) {
+ const node = column[j];
+ node.pos[0] = (layout == LiteGraph.VERTICAL_LAYOUT) ? y : x;
+ node.pos[1] = (layout == LiteGraph.VERTICAL_LAYOUT) ? x : y;
+ const max_size_index = (layout == LiteGraph.VERTICAL_LAYOUT) ? 1 : 0;
+ if (node.size[max_size_index] > max_size) {
+ max_size = node.size[max_size_index];
}
- y += node.size[1] + margin + LiteGraph.NODE_TITLE_HEIGHT;
+ const node_size_index = (layout == LiteGraph.VERTICAL_LAYOUT) ? 0 : 1;
+ y += node.size[node_size_index] + margin + LiteGraph.NODE_TITLE_HEIGHT;
}
x += max_size + margin;
}
@@ -2468,43 +2450,34 @@
this.title = this.constructor.title;
}
- if (this.onConnectionsChange) {
- if (this.inputs) {
- for (var i = 0; i < this.inputs.length; ++i) {
- var input = this.inputs[i];
- var link_info = this.graph
- ? this.graph.links[input.link]
- : null;
- this.onConnectionsChange(
- LiteGraph.INPUT,
- i,
- true,
- link_info,
- input
- ); //link_info has been created now, so its updated
- }
- }
+ if (this.inputs) {
+ for (var i = 0; i < this.inputs.length; ++i) {
+ var input = this.inputs[i];
+ var link_info = this.graph ? this.graph.links[input.link] : null;
+ if (this.onConnectionsChange)
+ this.onConnectionsChange( LiteGraph.INPUT, i, true, link_info, input ); //link_info has been created now, so its updated
- if (this.outputs) {
- for (var i = 0; i < this.outputs.length; ++i) {
- var output = this.outputs[i];
- if (!output.links) {
- continue;
- }
- for (var j = 0; j < output.links.length; ++j) {
- var link_info = this.graph
- ? this.graph.links[output.links[j]]
- : null;
- this.onConnectionsChange(
- LiteGraph.OUTPUT,
- i,
- true,
- link_info,
- output
- ); //link_info has been created now, so its updated
- }
- }
- }
+ if( this.onInputAdded )
+ this.onInputAdded(input);
+
+ }
+ }
+
+ if (this.outputs) {
+ for (var i = 0; i < this.outputs.length; ++i) {
+ var output = this.outputs[i];
+ if (!output.links) {
+ continue;
+ }
+ for (var j = 0; j < output.links.length; ++j) {
+ var link_info = this.graph ? this.graph.links[output.links[j]] : null;
+ if (this.onConnectionsChange)
+ this.onConnectionsChange( LiteGraph.OUTPUT, i, true, link_info, output ); //link_info has been created now, so its updated
+ }
+
+ if( this.onOutputAdded )
+ this.onOutputAdded(output);
+ }
}
if( this.widgets )
@@ -3200,6 +3173,15 @@
return;
}
+ if(slot == null)
+ {
+ console.error("slot must be a number");
+ return;
+ }
+
+ if(slot.constructor !== Number)
+ console.warn("slot must be a number, use node.trigger('name') if you want to use a string");
+
var output = this.outputs[slot];
if (!output) {
return;
@@ -3346,26 +3328,26 @@
* @param {Object} extra_info this can be used to have special properties of an output (label, special color, position, etc)
*/
LGraphNode.prototype.addOutput = function(name, type, extra_info) {
- var o = { name: name, type: type, links: null };
+ var output = { name: name, type: type, links: null };
if (extra_info) {
for (var i in extra_info) {
- o[i] = extra_info[i];
+ output[i] = extra_info[i];
}
}
if (!this.outputs) {
this.outputs = [];
}
- this.outputs.push(o);
+ this.outputs.push(output);
if (this.onOutputAdded) {
- this.onOutputAdded(o);
+ this.onOutputAdded(output);
}
if (LiteGraph.auto_load_slot_types) LiteGraph.registerNodeAndSlotType(this,type,true);
this.setSize( this.computeSize() );
this.setDirtyCanvas(true, true);
- return o;
+ return output;
};
/**
@@ -3437,10 +3419,10 @@
*/
LGraphNode.prototype.addInput = function(name, type, extra_info) {
type = type || 0;
- var o = { name: name, type: type, link: null };
+ var input = { name: name, type: type, link: null };
if (extra_info) {
for (var i in extra_info) {
- o[i] = extra_info[i];
+ input[i] = extra_info[i];
}
}
@@ -3448,17 +3430,17 @@
this.inputs = [];
}
- this.inputs.push(o);
+ this.inputs.push(input);
this.setSize( this.computeSize() );
if (this.onInputAdded) {
- this.onInputAdded(o);
+ this.onInputAdded(input);
}
LiteGraph.registerNodeAndSlotType(this,type);
this.setDirtyCanvas(true, true);
- return o;
+ return input;
};
/**
@@ -5210,6 +5192,7 @@ LGraphNode.prototype.executeAction = function(action)
this.allow_dragcanvas = true;
this.allow_dragnodes = true;
this.allow_interaction = true; //allow to control widgets, buttons, collapse, etc
+ this.multi_select = false; //allow selecting multi nodes without pressing extra keys
this.allow_searchbox = true;
this.allow_reconnect_links = true; //allows to change a connection with having to redo it again
this.align_to_grid = false; //snap to grid
@@ -5435,7 +5418,7 @@ LGraphNode.prototype.executeAction = function(action)
};
/**
- * returns the visualy active graph (in case there are more in the stack)
+ * returns the visually active graph (in case there are more in the stack)
* @method getCurrentGraph
* @return {LGraph} the active graph
*/
@@ -6060,9 +6043,13 @@ LGraphNode.prototype.executeAction = function(action)
this.graph.beforeChange();
this.node_dragged = node;
}
- if (!this.selected_nodes[node.id]) {
- this.processNodeSelected(node, e);
- }
+ this.processNodeSelected(node, e);
+ } else { // double-click
+ /**
+ * Don't call the function if the block is already selected.
+ * Otherwise, it could cause the block to be unselected while its panel is open.
+ */
+ if (!node.is_selected) this.processNodeSelected(node, e);
}
this.dirty_canvas = true;
@@ -6474,6 +6461,10 @@ LGraphNode.prototype.executeAction = function(action)
var n = this.selected_nodes[i];
n.pos[0] += delta[0] / this.ds.scale;
n.pos[1] += delta[1] / this.ds.scale;
+ if (!n.is_selected) this.processNodeSelected(n, e); /*
+ * Don't call the function if the block is already selected.
+ * Otherwise, it could cause the block to be unselected while dragging.
+ */
}
this.dirty_canvas = true;
@@ -7287,7 +7278,7 @@ LGraphNode.prototype.executeAction = function(action)
};
LGraphCanvas.prototype.processNodeSelected = function(node, e) {
- this.selectNode(node, e && (e.shiftKey||e.ctrlKey));
+ this.selectNode(node, e && (e.shiftKey || e.ctrlKey || this.multi_select));
if (this.onNodeSelected) {
this.onNodeSelected(node);
}
@@ -7323,6 +7314,7 @@ LGraphNode.prototype.executeAction = function(action)
for (var i in nodes) {
var node = nodes[i];
if (node.is_selected) {
+ this.deselectNode(node);
continue;
}
@@ -9742,13 +9734,17 @@ LGraphNode.prototype.executeAction = function(action)
ctx.fillRect(margin, y, widget_width - margin * 2, H);
var range = w.options.max - w.options.min;
var nvalue = (w.value - w.options.min) / range;
- ctx.fillStyle = active_widget == w ? "#89A" : "#678";
+ if(nvalue < 0.0) nvalue = 0.0;
+ if(nvalue > 1.0) nvalue = 1.0;
+ ctx.fillStyle = w.options.hasOwnProperty("slider_color") ? w.options.slider_color : (active_widget == w ? "#89A" : "#678");
ctx.fillRect(margin, y, nvalue * (widget_width - margin * 2), H);
if(show_text && !w.disabled)
ctx.strokeRect(margin, y, widget_width - margin * 2, H);
if (w.marker) {
var marker_nvalue = (w.marker - w.options.min) / range;
- ctx.fillStyle = "#AA9";
+ if(marker_nvalue < 0.0) marker_nvalue = 0.0;
+ if(marker_nvalue > 1.0) marker_nvalue = 1.0;
+ ctx.fillStyle = w.options.hasOwnProperty("marker_color") ? w.options.marker_color : "#AA9";
ctx.fillRect( margin + marker_nvalue * (widget_width - margin * 2), y, 2, H );
}
if (show_text) {
@@ -9915,6 +9911,7 @@ LGraphNode.prototype.executeAction = function(action)
case "slider":
var range = w.options.max - w.options.min;
var nvalue = Math.clamp((x - 15) / (widget_width - 30), 0, 1);
+ if(w.options.read_only) break;
w.value = w.options.min + (w.options.max - w.options.min) * nvalue;
if (w.callback) {
setTimeout(function() {
@@ -9926,8 +9923,16 @@ LGraphNode.prototype.executeAction = function(action)
case "number":
case "combo":
var old_value = w.value;
- if (event.type == LiteGraph.pointerevents_method+"move" && w.type == "number") {
- w.value += event.deltaX * 0.1 * (w.options.step || 1);
+ var delta = x < 40 ? -1 : x > widget_width - 40 ? 1 : 0;
+ var allow_scroll = true;
+ if (delta) {
+ if (x > -3 && x < widget_width + 3) {
+ allow_scroll = false;
+ }
+ }
+ if (allow_scroll && event.type == LiteGraph.pointerevents_method+"move" && w.type == "number") {
+ if(event.deltaX)
+ w.value += event.deltaX * 0.1 * (w.options.step || 1);
if ( w.options.min != null && w.value < w.options.min ) {
w.value = w.options.min;
}
@@ -9994,6 +9999,12 @@ LGraphNode.prototype.executeAction = function(action)
var delta = x < 40 ? -1 : x > widget_width - 40 ? 1 : 0;
if (event.click_time < 200 && delta == 0) {
this.prompt("Value",w.value,function(v) {
+ // check if v is a valid equation or a number
+ if (/^[0-9+\-*/()\s]+$/.test(v)) {
+ try {//solve the equation if possible
+ v = eval(v);
+ } catch (e) { }
+ }
this.value = Number(v);
inner_value_change(this, this.value);
}.bind(w),
@@ -10022,7 +10033,6 @@ LGraphNode.prototype.executeAction = function(action)
case "text":
if (event.type == LiteGraph.pointerevents_method+"down") {
this.prompt("Value",w.value,function(v) {
- this.value = v;
inner_value_change(this, v);
}.bind(w),
event,w.options ? w.options.multiline : false );
@@ -10047,6 +10057,9 @@ LGraphNode.prototype.executeAction = function(action)
}//end for
function inner_value_change(widget, value) {
+ if(widget.type == "number"){
+ value = Number(value);
+ }
widget.value = value;
if ( widget.options && widget.options.property && node.properties[widget.options.property] !== undefined ) {
node.setProperty( widget.options.property, value );
@@ -11165,7 +11178,7 @@ LGraphNode.prototype.executeAction = function(action)
LGraphCanvas.search_limit = -1;
LGraphCanvas.prototype.showSearchBox = function(event, options) {
// proposed defaults
- def_options = { slot_from: null
+ var def_options = { slot_from: null
,node_from: null
,node_to: null
,do_type_filter: LiteGraph.search_filter_enabled // TODO check for registered_slot_[in/out]_types not empty // this will be checked for functionality enabled : filter on slot type, in and out
@@ -11863,7 +11876,7 @@ LGraphNode.prototype.executeAction = function(action)
// TODO refactor, theer are different dialog, some uses createDialog, some dont
LGraphCanvas.prototype.createDialog = function(html, options) {
- def_options = { checkForInput: false, closeOnLeave: true, closeOnLeave_checkModified: true };
+ var def_options = { checkForInput: false, closeOnLeave: true, closeOnLeave_checkModified: true };
options = Object.assign(def_options, options || {});
var dialog = document.createElement("div");
@@ -11993,7 +12006,8 @@ LGraphNode.prototype.executeAction = function(action)
if (root.onClose && typeof root.onClose == "function"){
root.onClose();
}
- root.parentNode.removeChild(root);
+ if(root.parentNode)
+ root.parentNode.removeChild(root);
/* XXX CHECK THIS */
if(this.parentNode){
this.parentNode.removeChild(this);
@@ -12285,7 +12299,7 @@ LGraphNode.prototype.executeAction = function(action)
var ref_window = this.getCanvasWindow();
var that = this;
var graphcanvas = this;
- panel = this.createPanel(node.title || "",{
+ var panel = this.createPanel(node.title || "",{
closable: true
,window: ref_window
,onOpen: function(){
diff --git a/web/scripts/app.js b/web/scripts/app.js
index 64f00c165..dd9c21f89 100644
--- a/web/scripts/app.js
+++ b/web/scripts/app.js
@@ -835,7 +835,7 @@ class ComfyApp {
app.#invokeExtensionsAsync("nodeCreated", this);
},
{
- title: nodeData.name,
+ title: nodeData.display_name || nodeData.name,
comfyClass: nodeData.name,
}
);
diff --git a/web/scripts/defaultGraph.js b/web/scripts/defaultGraph.js
index 967377ad6..9b3cb4a7e 100644
--- a/web/scripts/defaultGraph.js
+++ b/web/scripts/defaultGraph.js
@@ -13,7 +13,7 @@ export const defaultGraph = {
inputs: [{ name: "clip", type: "CLIP", link: 5 }],
outputs: [{ name: "CONDITIONING", type: "CONDITIONING", links: [6], slot_index: 0 }],
properties: {},
- widgets_values: ["bad hands"],
+ widgets_values: ["text, watermark"],
},
{
id: 6,
@@ -26,7 +26,7 @@ export const defaultGraph = {
inputs: [{ name: "clip", type: "CLIP", link: 3 }],
outputs: [{ name: "CONDITIONING", type: "CONDITIONING", links: [4], slot_index: 0 }],
properties: {},
- widgets_values: ["masterpiece best quality girl"],
+ widgets_values: ["beautiful scenery nature glass bottle landscape, , purple galaxy bottle,"],
},
{
id: 5,
@@ -56,7 +56,7 @@ export const defaultGraph = {
],
outputs: [{ name: "LATENT", type: "LATENT", links: [7], slot_index: 0 }],
properties: {},
- widgets_values: [8566257, true, 20, 8, "euler", "normal", 1],
+ widgets_values: [156680208700286, true, 20, 8, "euler", "normal", 1],
},
{
id: 8,
diff --git a/web/scripts/widgets.js b/web/scripts/widgets.js
index 68048e996..5ed8cbda9 100644
--- a/web/scripts/widgets.js
+++ b/web/scripts/widgets.js
@@ -390,7 +390,7 @@ export const ComfyWidgets = {
const fileInput = document.createElement("input");
Object.assign(fileInput, {
type: "file",
- accept: "image/jpeg,image/png",
+ accept: "image/jpeg,image/png,image/webp",
style: "display: none",
onchange: async () => {
if (fileInput.files.length) {