Support newer glora format.

* add github action to automatically handle stale user support issues

* improve stale message

* remove token part

.ci/windows_base_files/README_VERY_IMPORTANT.txt

@@ -14,7 +14,7 @@ run_cpu.bat
 
 IF YOU GET A RED ERROR IN THE UI MAKE SURE YOU HAVE A MODEL/CHECKPOINT IN: ComfyUI\models\checkpoints
 
-You can download the stable diffusion 1.5 one from: https://huggingface.co/runwayml/stable-diffusion-v1-5/blob/main/v1-5-pruned-emaonly.ckpt
+You can download the stable diffusion 1.5 one from: https://huggingface.co/Comfy-Org/stable-diffusion-v1-5-archive/blob/main/v1-5-pruned-emaonly-fp16.safetensors
 
 
 RECOMMENDED WAY TO UPDATE:

.github/workflows/stale-issues.yml

@@ -0,0 +1,21 @@
+name: 'Close stale issues'
+on:
+  schedule:
+    # Run daily at 430 am PT
+    - cron: '30 11 * * *'
+permissions:
+  issues: write
+
+jobs:
+  stale:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/stale@v9
+        with:
+          stale-issue-message: "This issue is being marked stale because it has not had any activity for 30 days. Reply below within 7 days if your issue still isn't solved, and it will be left open. Otherwise, the issue will be closed automatically."
+          days-before-stale: 30
+          days-before-close: 7
+          stale-issue-label: 'Stale'
+          only-labels: 'User Support'
+          exempt-all-assignees: true
+          exempt-all-milestones: true

README.md

@@ -1,7 +1,7 @@
 <div align="center">
 
 # ComfyUI
-**The most powerful and modular stable diffusion GUI and backend.**
+**The most powerful and modular diffusion model GUI and backend.**
 
 
 [![Website][website-shield]][website-url]
@@ -135,17 +135,17 @@ Put your VAE in: models/vae
 ### AMD GPUs (Linux only)
 AMD users can install rocm and pytorch with pip if you don't have it already installed, this is the command to install the stable version:
 
-```pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.0```
+```pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.1```
 
-This is the command to install the nightly with ROCm 6.0 which might have some performance improvements:
+This is the command to install the nightly with ROCm 6.2 which might have some performance improvements:
 
-```pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm6.1```
+```pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm6.2```
 
 ### NVIDIA
 
 Nvidia users should install stable pytorch using this command:
 
-```pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu121```
+```pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu124```
 
 This is the command to install pytorch nightly instead which might have performance improvements:
 

api_server/routes/internal/internal_routes.py

@@ -2,6 +2,7 @@ from aiohttp import web
 from typing import Optional
 from folder_paths import models_dir, user_directory, output_directory
 from api_server.services.file_service import FileService
+import app.logger
 
 class InternalRoutes:
     '''
@@ -31,6 +32,9 @@ class InternalRoutes:
             except Exception as e:
                 return web.json_response({"error": str(e)}, status=500)
 
+        @self.routes.get('/logs')
+        async def get_logs(request):
+            return web.json_response(app.logger.get_logs())
 
     def get_app(self):
         if self._app is None:

app/frontend_management.py

@@ -8,7 +8,7 @@ import zipfile
 from dataclasses import dataclass
 from functools import cached_property
 from pathlib import Path
-from typing import TypedDict
+from typing import TypedDict, Optional
 
 import requests
 from typing_extensions import NotRequired
@@ -132,12 +132,13 @@ class FrontendManager:
         return match_result.group(1), match_result.group(2), match_result.group(3)
 
     @classmethod
-    def init_frontend_unsafe(cls, version_string: str) -> str:
+    def init_frontend_unsafe(cls, version_string: str, provider: Optional[FrontEndProvider] = None) -> str:
         """
         Initializes the frontend for the specified version.
 
         Args:
             version_string (str): The version string.
+            provider (FrontEndProvider, optional): The provider to use. Defaults to None.
 
         Returns:
             str: The path to the initialized frontend.
@@ -150,7 +151,7 @@ class FrontendManager:
             return cls.DEFAULT_FRONTEND_PATH
 
         repo_owner, repo_name, version = cls.parse_version_string(version_string)
-        provider = FrontEndProvider(repo_owner, repo_name)
+        provider = provider or FrontEndProvider(repo_owner, repo_name)
         release = provider.get_release(version)
 
         semantic_version = release["tag_name"].lstrip("v")
@@ -158,6 +159,7 @@ class FrontendManager:
             Path(cls.CUSTOM_FRONTENDS_ROOT) / provider.folder_name / semantic_version
         )
         if not os.path.exists(web_root):
+            try:
                 os.makedirs(web_root, exist_ok=True)
                 logging.info(
                     "Downloading frontend(%s) version(%s) to (%s)",
@@ -167,6 +169,11 @@ class FrontendManager:
                 )
                 logging.debug(release)
                 download_release_asset_zip(release, destination_path=web_root)
+            finally:
+                # Clean up the directory if it is empty, i.e. the download failed
+                if not os.listdir(web_root):
+                    os.rmdir(web_root)
+
         return web_root
 
     @classmethod

app/logger.py

@@ -0,0 +1,31 @@
+import logging
+from logging.handlers import MemoryHandler
+from collections import deque
+
+logs = None
+formatter = logging.Formatter("%(asctime)s - %(name)s - %(levelname)s - %(message)s")
+
+
+def get_logs():
+    return "\n".join([formatter.format(x) for x in logs])
+
+
+def setup_logger(verbose: bool = False, capacity: int = 300):
+    global logs
+    if logs:
+        return
+
+    # Setup default global logger
+    logger = logging.getLogger()
+    logger.setLevel(logging.DEBUG if verbose else logging.INFO)
+
+    stream_handler = logging.StreamHandler()
+    stream_handler.setFormatter(logging.Formatter("%(message)s"))
+    logger.addHandler(stream_handler)
+
+    # Create a memory handler with a deque as its buffer
+    logs = deque(maxlen=capacity)
+    memory_handler = MemoryHandler(capacity, flushLevel=logging.INFO)
+    memory_handler.buffer = logs
+    memory_handler.setFormatter(formatter)
+    logger.addHandler(memory_handler)

comfy/cli_args.py

@@ -179,10 +179,3 @@ if args.windows_standalone_build:
 
 if args.disable_auto_launch:
     args.auto_launch = False
-
-import logging
-logging_level = logging.INFO
-if args.verbose:
-    logging_level = logging.DEBUG
-
-logging.basicConfig(format="%(message)s", level=logging_level)

comfy/controlnet.py

@@ -34,7 +34,7 @@ import comfy.t2i_adapter.adapter
 import comfy.ldm.cascade.controlnet
 import comfy.cldm.mmdit
 import comfy.ldm.hydit.controlnet
-import comfy.ldm.flux.controlnet_xlabs
+import comfy.ldm.flux.controlnet
 
 
 def broadcast_image_to(tensor, target_batch_size, batched_number):
@@ -148,7 +148,7 @@ class ControlBase:
                         elif self.strength_type == StrengthType.LINEAR_UP:
                             x *= (self.strength ** float(len(control_output) - i))
 
-                    if x.dtype != output_dtype:
+                    if output_dtype is not None and x.dtype != output_dtype:
                         x = x.to(output_dtype)
 
                 out[key].append(x)
@@ -206,7 +206,6 @@ class ControlNet(ControlBase):
         if self.manual_cast_dtype is not None:
             dtype = self.manual_cast_dtype
 
-        output_dtype = x_noisy.dtype
         if self.cond_hint is None or x_noisy.shape[2] * self.compression_ratio != self.cond_hint.shape[2] or x_noisy.shape[3] * self.compression_ratio != self.cond_hint.shape[3]:
             if self.cond_hint is not None:
                 del self.cond_hint
@@ -236,7 +235,7 @@ class ControlNet(ControlBase):
         x_noisy = self.model_sampling_current.calculate_input(t, x_noisy)
 
         control = self.control_model(x=x_noisy.to(dtype), hint=self.cond_hint, timesteps=timestep.to(dtype), context=context.to(dtype), **extra)
-        return self.control_merge(control, control_prev, output_dtype)
+        return self.control_merge(control, control_prev, output_dtype=None)
 
     def copy(self):
         c = ControlNet(None, global_average_pooling=self.global_average_pooling, load_device=self.load_device, manual_cast_dtype=self.manual_cast_dtype)
@@ -433,12 +432,30 @@ def load_controlnet_hunyuandit(controlnet_data):
 
 def load_controlnet_flux_xlabs(sd):
     model_config, operations, load_device, unet_dtype, manual_cast_dtype, offload_device = controlnet_config(sd)
-    control_model = comfy.ldm.flux.controlnet_xlabs.ControlNetFlux(operations=operations, device=offload_device, dtype=unet_dtype, **model_config.unet_config)
+    control_model = comfy.ldm.flux.controlnet.ControlNetFlux(operations=operations, device=offload_device, dtype=unet_dtype, **model_config.unet_config)
     control_model = controlnet_load_state_dict(control_model, sd)
     extra_conds = ['y', 'guidance']
     control = ControlNet(control_model, load_device=load_device, manual_cast_dtype=manual_cast_dtype, extra_conds=extra_conds)
     return control
 
+def load_controlnet_flux_instantx(sd):
+    new_sd = comfy.model_detection.convert_diffusers_mmdit(sd, "")
+    model_config, operations, load_device, unet_dtype, manual_cast_dtype, offload_device = controlnet_config(new_sd)
+    for k in sd:
+        new_sd[k] = sd[k]
+
+    num_union_modes = 0
+    union_cnet = "controlnet_mode_embedder.weight"
+    if union_cnet in new_sd:
+        num_union_modes = new_sd[union_cnet].shape[0]
+
+    control_model = comfy.ldm.flux.controlnet.ControlNetFlux(latent_input=True, num_union_modes=num_union_modes, operations=operations, device=offload_device, dtype=unet_dtype, **model_config.unet_config)
+    control_model = controlnet_load_state_dict(control_model, new_sd)
+
+    latent_format = comfy.latent_formats.Flux()
+    extra_conds = ['y', 'guidance']
+    control = ControlNet(control_model, compression_ratio=1, latent_format=latent_format, load_device=load_device, manual_cast_dtype=manual_cast_dtype, extra_conds=extra_conds)
+    return control
 
 def load_controlnet(ckpt_path, model=None):
     controlnet_data = comfy.utils.load_torch_file(ckpt_path, safe_load=True)
@@ -504,8 +521,10 @@ def load_controlnet(ckpt_path, model=None):
     elif "controlnet_blocks.0.weight" in controlnet_data: #SD3 diffusers format
         if "double_blocks.0.img_attn.norm.key_norm.scale" in controlnet_data:
             return load_controlnet_flux_xlabs(controlnet_data)
-        else:
+        elif "pos_embed_input.proj.weight" in controlnet_data:
             return load_controlnet_mmdit(controlnet_data)
+        elif "controlnet_x_embedder.weight" in controlnet_data:
+            return load_controlnet_flux_instantx(controlnet_data)
 
     pth_key = 'control_model.zero_convs.0.0.weight'
     pth = False

comfy/float.py

@@ -1,7 +1,18 @@
 import torch
+import math
+
+def calc_mantissa(abs_x, exponent, normal_mask, MANTISSA_BITS, EXPONENT_BIAS, generator=None):
+    mantissa_scaled = torch.where(
+        normal_mask,
+        (abs_x / (2.0 ** (exponent - EXPONENT_BIAS)) - 1.0) * (2**MANTISSA_BITS),
+        (abs_x / (2.0 ** (-EXPONENT_BIAS + 1 - MANTISSA_BITS)))
+    )
+
+    mantissa_scaled += torch.rand(mantissa_scaled.size(), dtype=mantissa_scaled.dtype, layout=mantissa_scaled.layout, device=mantissa_scaled.device, generator=generator)
+    return mantissa_scaled.floor() / (2**MANTISSA_BITS)
 
 #Not 100% sure about this
-def manual_stochastic_round_to_float8(x, dtype):
+def manual_stochastic_round_to_float8(x, dtype, generator=None):
     if dtype == torch.float8_e4m3fn:
         EXPONENT_BITS, MANTISSA_BITS, EXPONENT_BIAS = 4, 3, 7
     elif dtype == torch.float8_e5m2:
@@ -9,44 +20,33 @@ def manual_stochastic_round_to_float8(x, dtype):
     else:
         raise ValueError("Unsupported dtype")
 
+    x = x.half()
     sign = torch.sign(x)
     abs_x = x.abs()
+    sign = torch.where(abs_x == 0, 0, sign)
 
     # Combine exponent calculation and clamping
     exponent = torch.clamp(
-        torch.floor(torch.log2(abs_x)).to(torch.int32) + EXPONENT_BIAS,
+        torch.floor(torch.log2(abs_x)) + EXPONENT_BIAS,
         0, 2**EXPONENT_BITS - 1
     )
 
     # Combine mantissa calculation and rounding
-    # min_normal = 2.0 ** (-EXPONENT_BIAS + 1)
-    # zero_mask = (abs_x == 0)
-    # subnormal_mask = (exponent == 0) & (abs_x != 0)
     normal_mask = ~(exponent == 0)
 
-    mantissa_scaled = torch.where(
+    abs_x[:] = calc_mantissa(abs_x, exponent, normal_mask, MANTISSA_BITS, EXPONENT_BIAS, generator=generator)
+
+    sign *= torch.where(
         normal_mask,
-        (abs_x / (2.0 ** (exponent - EXPONENT_BIAS)) - 1.0) * (2**MANTISSA_BITS),
-        (abs_x / (2.0 ** (-EXPONENT_BIAS + 1 - MANTISSA_BITS)))
-    )
-    mantissa_floor = mantissa_scaled.floor()
-    mantissa = torch.where(
-        torch.rand_like(mantissa_scaled) < (mantissa_scaled - mantissa_floor),
-        (mantissa_floor + 1) / (2**MANTISSA_BITS),
-        mantissa_floor / (2**MANTISSA_BITS)
-    )
-    result = torch.where(
-        normal_mask,
-        sign * (2.0 ** (exponent - EXPONENT_BIAS)) * (1.0 + mantissa),
-        sign * (2.0 ** (-EXPONENT_BIAS + 1)) * mantissa
+        (2.0 ** (exponent - EXPONENT_BIAS)) * (1.0 + abs_x),
+        (2.0 ** (-EXPONENT_BIAS + 1)) * abs_x
     )
 
-    result = torch.where(abs_x == 0, 0, result)
-    return result.to(dtype=dtype)
+    return sign
 
 
 
-def stochastic_rounding(value, dtype):
+def stochastic_rounding(value, dtype, seed=0):
     if dtype == torch.float32:
         return value.to(dtype=torch.float32)
     if dtype == torch.float16:
@@ -54,6 +54,13 @@ def stochastic_rounding(value, dtype):
     if dtype == torch.bfloat16:
         return value.to(dtype=torch.bfloat16)
     if dtype == torch.float8_e4m3fn or dtype == torch.float8_e5m2:
-        return manual_stochastic_round_to_float8(value, dtype)
+        generator = torch.Generator(device=value.device)
+        generator.manual_seed(seed)
+        output = torch.empty_like(value, dtype=dtype)
+        num_slices = max(1, (value.numel() / (4096 * 4096)))
+        slice_size = max(1, round(value.shape[0] / num_slices))
+        for i in range(0, value.shape[0], slice_size):
+            output[i:i+slice_size].copy_(manual_stochastic_round_to_float8(value[i:i+slice_size], dtype, generator=generator))
+        return output
 
     return value.to(dtype=dtype)

comfy/ldm/common_dit.py

@@ -1,4 +1,5 @@
 import torch
+import comfy.ops
 
 def pad_to_patch_size(img, patch_size=(2, 2), padding_mode="circular"):
     if padding_mode == "circular" and torch.jit.is_tracing() or torch.jit.is_scripting():
@@ -6,3 +7,15 @@ def pad_to_patch_size(img, patch_size=(2, 2), padding_mode="circular"):
     pad_h = (patch_size[0] - img.shape[-2] % patch_size[0]) % patch_size[0]
     pad_w = (patch_size[1] - img.shape[-1] % patch_size[1]) % patch_size[1]
     return torch.nn.functional.pad(img, (0, pad_w, 0, pad_h), mode=padding_mode)
+
+try:
+    rms_norm_torch = torch.nn.functional.rms_norm
+except:
+    rms_norm_torch = None
+
+def rms_norm(x, weight, eps=1e-6):
+    if rms_norm_torch is not None:
+        return rms_norm_torch(x, weight.shape, weight=comfy.ops.cast_to(weight, dtype=x.dtype, device=x.device), eps=eps)
+    else:
+        rrms = torch.rsqrt(torch.mean(x**2, dim=-1, keepdim=True) + eps)
+        return (x * rrms) * comfy.ops.cast_to(weight, dtype=x.dtype, device=x.device)

comfy/ldm/flux/controlnet.py

@@ -0,0 +1,151 @@
+#Original code can be found on: https://github.com/XLabs-AI/x-flux/blob/main/src/flux/controlnet.py
+
+import torch
+import math
+from torch import Tensor, nn
+from einops import rearrange, repeat
+
+from .layers import (DoubleStreamBlock, EmbedND, LastLayer,
+                                 MLPEmbedder, SingleStreamBlock,
+                                 timestep_embedding)
+
+from .model import Flux
+import comfy.ldm.common_dit
+
+
+class ControlNetFlux(Flux):
+    def __init__(self, latent_input=False, num_union_modes=0, image_model=None, dtype=None, device=None, operations=None, **kwargs):
+        super().__init__(final_layer=False, dtype=dtype, device=device, operations=operations, **kwargs)
+
+        self.main_model_double = 19
+        self.main_model_single = 38
+        # add ControlNet blocks
+        self.controlnet_blocks = nn.ModuleList([])
+        for _ in range(self.params.depth):
+            controlnet_block = operations.Linear(self.hidden_size, self.hidden_size, dtype=dtype, device=device)
+            self.controlnet_blocks.append(controlnet_block)
+
+        self.controlnet_single_blocks = nn.ModuleList([])
+        for _ in range(self.params.depth_single_blocks):
+            self.controlnet_single_blocks.append(operations.Linear(self.hidden_size, self.hidden_size, dtype=dtype, device=device))
+
+        self.num_union_modes = num_union_modes
+        self.controlnet_mode_embedder = None
+        if self.num_union_modes > 0:
+            self.controlnet_mode_embedder = operations.Embedding(self.num_union_modes, self.hidden_size, dtype=dtype, device=device)
+
+        self.gradient_checkpointing = False
+        self.latent_input = latent_input
+        self.pos_embed_input = operations.Linear(self.in_channels, self.hidden_size, bias=True, dtype=dtype, device=device)
+        if not self.latent_input:
+            self.input_hint_block = nn.Sequential(
+                operations.Conv2d(3, 16, 3, padding=1, dtype=dtype, device=device),
+                nn.SiLU(),
+                operations.Conv2d(16, 16, 3, padding=1, dtype=dtype, device=device),
+                nn.SiLU(),
+                operations.Conv2d(16, 16, 3, padding=1, stride=2, dtype=dtype, device=device),
+                nn.SiLU(),
+                operations.Conv2d(16, 16, 3, padding=1, dtype=dtype, device=device),
+                nn.SiLU(),
+                operations.Conv2d(16, 16, 3, padding=1, stride=2, dtype=dtype, device=device),
+                nn.SiLU(),
+                operations.Conv2d(16, 16, 3, padding=1, dtype=dtype, device=device),
+                nn.SiLU(),
+                operations.Conv2d(16, 16, 3, padding=1, stride=2, dtype=dtype, device=device),
+                nn.SiLU(),
+                operations.Conv2d(16, 16, 3, padding=1, dtype=dtype, device=device)
+            )
+
+    def forward_orig(
+        self,
+        img: Tensor,
+        img_ids: Tensor,
+        controlnet_cond: Tensor,
+        txt: Tensor,
+        txt_ids: Tensor,
+        timesteps: Tensor,
+        y: Tensor,
+        guidance: Tensor = None,
+        control_type: Tensor = None,
+    ) -> Tensor:
+        if img.ndim != 3 or txt.ndim != 3:
+            raise ValueError("Input img and txt tensors must have 3 dimensions.")
+
+        # running on sequences img
+        img = self.img_in(img)
+        if not self.latent_input:
+            controlnet_cond = self.input_hint_block(controlnet_cond)
+            controlnet_cond = rearrange(controlnet_cond, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=2, pw=2)
+
+        controlnet_cond = self.pos_embed_input(controlnet_cond)
+        img = img + controlnet_cond
+        vec = self.time_in(timestep_embedding(timesteps, 256))
+        if self.params.guidance_embed:
+            vec = vec + self.guidance_in(timestep_embedding(guidance, 256))
+        vec = vec + self.vector_in(y)
+        txt = self.txt_in(txt)
+
+        if self.controlnet_mode_embedder is not None and len(control_type) > 0:
+            control_cond = self.controlnet_mode_embedder(torch.tensor(control_type, device=img.device), out_dtype=img.dtype).unsqueeze(0).repeat((txt.shape[0], 1, 1))
+            txt = torch.cat([control_cond, txt], dim=1)
+            txt_ids = torch.cat([txt_ids[:,:1], txt_ids], dim=1)
+
+        ids = torch.cat((txt_ids, img_ids), dim=1)
+        pe = self.pe_embedder(ids)
+
+        controlnet_double = ()
+
+        for i in range(len(self.double_blocks)):
+            img, txt = self.double_blocks[i](img=img, txt=txt, vec=vec, pe=pe)
+            controlnet_double = controlnet_double + (self.controlnet_blocks[i](img),)
+
+        img = torch.cat((txt, img), 1)
+
+        controlnet_single = ()
+
+        for i in range(len(self.single_blocks)):
+            img = self.single_blocks[i](img, vec=vec, pe=pe)
+            controlnet_single = controlnet_single + (self.controlnet_single_blocks[i](img[:, txt.shape[1] :, ...]),)
+
+        repeat = math.ceil(self.main_model_double / len(controlnet_double))
+        if self.latent_input:
+            out_input = ()
+            for x in controlnet_double:
+                    out_input += (x,) * repeat
+        else:
+            out_input = (controlnet_double * repeat)
+
+        out = {"input": out_input[:self.main_model_double]}
+        if len(controlnet_single) > 0:
+            repeat = math.ceil(self.main_model_single / len(controlnet_single))
+            out_output = ()
+            if self.latent_input:
+                for x in controlnet_single:
+                        out_output += (x,) * repeat
+            else:
+                out_output = (controlnet_single * repeat)
+            out["output"] = out_output[:self.main_model_single]
+        return out
+
+    def forward(self, x, timesteps, context, y, guidance=None, hint=None, **kwargs):
+        patch_size = 2
+        if self.latent_input:
+            hint = comfy.ldm.common_dit.pad_to_patch_size(hint, (patch_size, patch_size))
+            hint = rearrange(hint, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=patch_size, pw=patch_size)
+        else:
+            hint = hint * 2.0 - 1.0
+
+        bs, c, h, w = x.shape
+        x = comfy.ldm.common_dit.pad_to_patch_size(x, (patch_size, patch_size))
+
+        img = rearrange(x, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=patch_size, pw=patch_size)
+
+        h_len = ((h + (patch_size // 2)) // patch_size)
+        w_len = ((w + (patch_size // 2)) // patch_size)
+        img_ids = torch.zeros((h_len, w_len, 3), device=x.device, dtype=x.dtype)
+        img_ids[..., 1] = img_ids[..., 1] + torch.linspace(0, h_len - 1, steps=h_len, device=x.device, dtype=x.dtype)[:, None]
+        img_ids[..., 2] = img_ids[..., 2] + torch.linspace(0, w_len - 1, steps=w_len, device=x.device, dtype=x.dtype)[None, :]
+        img_ids = repeat(img_ids, "h w c -> b (h w) c", b=bs)
+
+        txt_ids = torch.zeros((bs, context.shape[1], 3), device=x.device, dtype=x.dtype)
+        return self.forward_orig(img, img_ids, hint, context, txt_ids, timesteps, y, guidance, control_type=kwargs.get("control_type", []))

comfy/ldm/flux/controlnet_xlabs.py

@@ -1,104 +0,0 @@
-#Original code can be found on: https://github.com/XLabs-AI/x-flux/blob/main/src/flux/controlnet.py
-
-import torch
-from torch import Tensor, nn
-from einops import rearrange, repeat
-
-from .layers import (DoubleStreamBlock, EmbedND, LastLayer,
-                                 MLPEmbedder, SingleStreamBlock,
-                                 timestep_embedding)
-
-from .model import Flux
-import comfy.ldm.common_dit
-
-
-class ControlNetFlux(Flux):
-    def __init__(self, image_model=None, dtype=None, device=None, operations=None, **kwargs):
-        super().__init__(final_layer=False, dtype=dtype, device=device, operations=operations, **kwargs)
-
-        # add ControlNet blocks
-        self.controlnet_blocks = nn.ModuleList([])
-        for _ in range(self.params.depth):
-            controlnet_block = operations.Linear(self.hidden_size, self.hidden_size, dtype=dtype, device=device)
-            # controlnet_block = zero_module(controlnet_block)
-            self.controlnet_blocks.append(controlnet_block)
-        self.pos_embed_input = operations.Linear(self.in_channels, self.hidden_size, bias=True, dtype=dtype, device=device)
-        self.gradient_checkpointing = False
-        self.input_hint_block = nn.Sequential(
-            operations.Conv2d(3, 16, 3, padding=1, dtype=dtype, device=device),
-            nn.SiLU(),
-            operations.Conv2d(16, 16, 3, padding=1, dtype=dtype, device=device),
-            nn.SiLU(),
-            operations.Conv2d(16, 16, 3, padding=1, stride=2, dtype=dtype, device=device),
-            nn.SiLU(),
-            operations.Conv2d(16, 16, 3, padding=1, dtype=dtype, device=device),
-            nn.SiLU(),
-            operations.Conv2d(16, 16, 3, padding=1, stride=2, dtype=dtype, device=device),
-            nn.SiLU(),
-            operations.Conv2d(16, 16, 3, padding=1, dtype=dtype, device=device),
-            nn.SiLU(),
-            operations.Conv2d(16, 16, 3, padding=1, stride=2, dtype=dtype, device=device),
-            nn.SiLU(),
-            operations.Conv2d(16, 16, 3, padding=1, dtype=dtype, device=device)
-        )
-
-    def forward_orig(
-        self,
-        img: Tensor,
-        img_ids: Tensor,
-        controlnet_cond: Tensor,
-        txt: Tensor,
-        txt_ids: Tensor,
-        timesteps: Tensor,
-        y: Tensor,
-        guidance: Tensor = None,
-    ) -> Tensor:
-        if img.ndim != 3 or txt.ndim != 3:
-            raise ValueError("Input img and txt tensors must have 3 dimensions.")
-
-        # running on sequences img
-        img = self.img_in(img)
-        controlnet_cond = self.input_hint_block(controlnet_cond)
-        controlnet_cond = rearrange(controlnet_cond, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=2, pw=2)
-        controlnet_cond = self.pos_embed_input(controlnet_cond)
-        img = img + controlnet_cond
-        vec = self.time_in(timestep_embedding(timesteps, 256))
-        if self.params.guidance_embed:
-            vec = vec + self.guidance_in(timestep_embedding(guidance, 256))
-        vec = vec + self.vector_in(y)
-        txt = self.txt_in(txt)
-
-        ids = torch.cat((txt_ids, img_ids), dim=1)
-        pe = self.pe_embedder(ids)
-
-        block_res_samples = ()
-
-        for block in self.double_blocks:
-            img, txt = block(img=img, txt=txt, vec=vec, pe=pe)
-            block_res_samples = block_res_samples + (img,)
-
-        controlnet_block_res_samples = ()
-        for block_res_sample, controlnet_block in zip(block_res_samples, self.controlnet_blocks):
-            block_res_sample = controlnet_block(block_res_sample)
-            controlnet_block_res_samples = controlnet_block_res_samples + (block_res_sample,)
-
-        return {"input": (controlnet_block_res_samples * 10)[:19]}
-
-    def forward(self, x, timesteps, context, y, guidance=None, hint=None, **kwargs):
-        hint = hint * 2.0 - 1.0
-
-        bs, c, h, w = x.shape
-        patch_size = 2
-        x = comfy.ldm.common_dit.pad_to_patch_size(x, (patch_size, patch_size))
-
-        img = rearrange(x, "b c (h ph) (w pw) -> b (h w) (c ph pw)", ph=patch_size, pw=patch_size)
-
-        h_len = ((h + (patch_size // 2)) // patch_size)
-        w_len = ((w + (patch_size // 2)) // patch_size)
-        img_ids = torch.zeros((h_len, w_len, 3), device=x.device, dtype=x.dtype)
-        img_ids[..., 1] = img_ids[..., 1] + torch.linspace(0, h_len - 1, steps=h_len, device=x.device, dtype=x.dtype)[:, None]
-        img_ids[..., 2] = img_ids[..., 2] + torch.linspace(0, w_len - 1, steps=w_len, device=x.device, dtype=x.dtype)[None, :]
-        img_ids = repeat(img_ids, "h w c -> b (h w) c", b=bs)
-
-        txt_ids = torch.zeros((bs, context.shape[1], 3), device=x.device, dtype=x.dtype)
-        return self.forward_orig(img, img_ids, hint, context, txt_ids, timesteps, y, guidance)

comfy/ldm/flux/layers.py

@@ -6,6 +6,7 @@ from torch import Tensor, nn
 
 from .math import attention, rope
 import comfy.ops
+import comfy.ldm.common_dit
 
 
 class EmbedND(nn.Module):
@@ -63,10 +64,7 @@ class RMSNorm(torch.nn.Module):
         self.scale = nn.Parameter(torch.empty((dim), dtype=dtype, device=device))
 
     def forward(self, x: Tensor):
-        x_dtype = x.dtype
-        x = x.float()
-        rrms = torch.rsqrt(torch.mean(x**2, dim=-1, keepdim=True) + 1e-6)
-        return (x * rrms).to(dtype=x_dtype) * comfy.ops.cast_to(self.scale, dtype=x_dtype, device=x.device)
+        return comfy.ldm.common_dit.rms_norm(x, self.scale, 1e-6)
 
 
 class QKNorm(torch.nn.Module):

comfy/ldm/hydit/models.py

@@ -372,7 +372,7 @@ class HunYuanDiT(nn.Module):
         for layer, block in enumerate(self.blocks):
             if layer > self.depth // 2:
                 if controls is not None:
-                    skip = skips.pop() + controls.pop()
+                    skip = skips.pop() + controls.pop().to(dtype=x.dtype)
                 else:
                     skip = skips.pop()
                 x = block(x, c, text_states, freqs_cis_img, skip)   # (N, L, D)

comfy/ldm/modules/diffusionmodules/mmdit.py

@@ -355,29 +355,9 @@ class RMSNorm(torch.nn.Module):
         else:
             self.register_parameter("weight", None)
 
-    def _norm(self, x):
-        """
-        Apply the RMSNorm normalization to the input tensor.
-        Args:
-            x (torch.Tensor): The input tensor.
-        Returns:
-            torch.Tensor: The normalized tensor.
-        """
-        return x * torch.rsqrt(x.pow(2).mean(-1, keepdim=True) + self.eps)
-
     def forward(self, x):
-        """
-        Forward pass through the RMSNorm layer.
-        Args:
-            x (torch.Tensor): The input tensor.
-        Returns:
-            torch.Tensor: The output tensor after applying RMSNorm.
-        """
-        x = self._norm(x)
-        if self.learnable_scale:
-            return x * self.weight.to(device=x.device, dtype=x.dtype)
-        else:
-            return x
+        return comfy.ldm.common_dit.rms_norm(x, self.weight, self.eps)
+
 
 
 class SwiGLUFeedForward(nn.Module):

comfy/lora.py

@@ -16,6 +16,7 @@
     along with this program.  If not, see <https://www.gnu.org/licenses/>.
 """
 
+from __future__ import annotations
 import comfy.utils
 import comfy.model_management
 import comfy.model_base
@@ -323,6 +324,7 @@ def model_lora_keys_unet(model, key_map={}):
                 to = diffusers_keys[k]
                 key_map["transformer.{}".format(k[:-len(".weight")])] = to #simpletrainer and probably regular diffusers flux lora format
                 key_map["lycoris_{}".format(k[:-len(".weight")].replace(".", "_"))] = to #simpletrainer lycoris
+                key_map["lora_transformer_{}".format(k[:-len(".weight")].replace(".", "_"))] = to #onetrainer
 
     return key_map
 
@@ -347,6 +349,39 @@ def weight_decompose(dora_scale, weight, lora_diff, alpha, strength, intermediat
         weight[:] = weight_calc
     return weight
 
+def pad_tensor_to_shape(tensor: torch.Tensor, new_shape: list[int]) -> torch.Tensor:
+    """
+    Pad a tensor to a new shape with zeros.
+
+    Args:
+        tensor (torch.Tensor): The original tensor to be padded.
+        new_shape (List[int]): The desired shape of the padded tensor.
+
+    Returns:
+        torch.Tensor: A new tensor padded with zeros to the specified shape.
+
+    Note:
+        If the new shape is smaller than the original tensor in any dimension,
+        the original tensor will be truncated in that dimension.
+    """
+    if any([new_shape[i] < tensor.shape[i] for i in range(len(new_shape))]):
+        raise ValueError("The new shape must be larger than the original tensor in all dimensions")
+
+    if len(new_shape) != len(tensor.shape):
+        raise ValueError("The new shape must have the same number of dimensions as the original tensor")
+
+    # Create a new tensor filled with zeros
+    padded_tensor = torch.zeros(new_shape, dtype=tensor.dtype, device=tensor.device)
+
+    # Create slicing tuples for both tensors
+    orig_slices = tuple(slice(0, dim) for dim in tensor.shape)
+    new_slices = tuple(slice(0, dim) for dim in tensor.shape)
+
+    # Copy the original tensor into the new tensor
+    padded_tensor[new_slices] = tensor[orig_slices]
+
+    return padded_tensor
+
 def calculate_weight(patches, weight, key, intermediate_dtype=torch.float32):
     for p in patches:
         strength = p[0]
@@ -375,12 +410,18 @@ def calculate_weight(patches, weight, key, intermediate_dtype=torch.float32):
             v = v[1]
 
         if patch_type == "diff":
-            w1 = v[0]
+            diff: torch.Tensor = v[0]
+            # An extra flag to pad the weight if the diff's shape is larger than the weight
+            do_pad_weight = len(v) > 1 and v[1]['pad_weight']
+            if do_pad_weight and diff.shape != weight.shape:
+                logging.info("Pad weight {} from {} to shape: {}".format(key, weight.shape, diff.shape))
+                weight = pad_tensor_to_shape(weight, diff.shape)
+
             if strength != 0.0:
-                if w1.shape != weight.shape:
-                    logging.warning("WARNING SHAPE MISMATCH {} WEIGHT NOT MERGED {} != {}".format(key, w1.shape, weight.shape))
+                if diff.shape != weight.shape:
+                    logging.warning("WARNING SHAPE MISMATCH {} WEIGHT NOT MERGED {} != {}".format(key, diff.shape, weight.shape))
                 else:
-                    weight += function(strength * comfy.model_management.cast_to_device(w1, weight.device, weight.dtype))
+                    weight += function(strength * comfy.model_management.cast_to_device(diff, weight.device, weight.dtype))
         elif patch_type == "lora": #lora/locon
             mat1 = comfy.model_management.cast_to_device(v[0], weight.device, intermediate_dtype)
             mat2 = comfy.model_management.cast_to_device(v[1], weight.device, intermediate_dtype)
@@ -487,20 +528,40 @@ def calculate_weight(patches, weight, key, intermediate_dtype=torch.float32):
             except Exception as e:
                 logging.error("ERROR {} {} {}".format(patch_type, key, e))
         elif patch_type == "glora":
-            if v[4] is not None:
-                alpha = v[4] / v[0].shape[0]
-            else:
-                alpha = 1.0
-
             dora_scale = v[5]
 
+            old_glora = False
+            if v[3].shape[1] == v[2].shape[0] == v[0].shape[0] == v[1].shape[1]:
+                rank = v[0].shape[0]
+                old_glora = True
+
+            if v[3].shape[0] == v[2].shape[1] == v[0].shape[1] == v[1].shape[0]:
+                if old_glora and v[1].shape[0] == weight.shape[0] and weight.shape[0] == weight.shape[1]:
+                    pass
+                else:
+                    old_glora = False
+                    rank = v[1].shape[0]
+
             a1 = comfy.model_management.cast_to_device(v[0].flatten(start_dim=1), weight.device, intermediate_dtype)
             a2 = comfy.model_management.cast_to_device(v[1].flatten(start_dim=1), weight.device, intermediate_dtype)
             b1 = comfy.model_management.cast_to_device(v[2].flatten(start_dim=1), weight.device, intermediate_dtype)
             b2 = comfy.model_management.cast_to_device(v[3].flatten(start_dim=1), weight.device, intermediate_dtype)
 
+            if v[4] is not None:
+                alpha = v[4] / rank
+            else:
+                alpha = 1.0
+
             try:
-                lora_diff = (torch.mm(b2, b1) + torch.mm(torch.mm(weight.flatten(start_dim=1), a2), a1)).reshape(weight.shape)
+                if old_glora:
+                    lora_diff = (torch.mm(b2, b1) + torch.mm(torch.mm(weight.flatten(start_dim=1).to(dtype=intermediate_dtype), a2), a1)).reshape(weight.shape) #old lycoris glora
+                else:
+                    if weight.dim() > 2:
+                        lora_diff = torch.einsum("o i ..., i j -> o j ...", torch.einsum("o i ..., i j -> o j ...", weight.to(dtype=intermediate_dtype), a1), a2).reshape(weight.shape)
+                    else:
+                        lora_diff = torch.mm(torch.mm(weight.to(dtype=intermediate_dtype), a1), a2).reshape(weight.shape)
+                    lora_diff += torch.mm(b1, b2).reshape(weight.shape)
+
                 if dora_scale is not None:
                     weight = function(weight_decompose(dora_scale, weight, lora_diff, alpha, strength, intermediate_dtype))
                 else:

comfy/model_management.py

@@ -45,6 +45,7 @@ cpu_state = CPUState.GPU
 total_vram = 0
 
 xpu_available = False
+torch_version = ""
 try:
     torch_version = torch.version.__version__
     xpu_available = (int(torch_version[0]) < 2 or (int(torch_version[0]) == 2 and int(torch_version[2]) <= 4)) and torch.xpu.is_available()
@@ -369,12 +370,11 @@ def offloaded_memory(loaded_models, device):
             offloaded_mem += m.model_offloaded_memory()
     return offloaded_mem
 
-def minimum_inference_memory():
-    return (1024 * 1024 * 1024) * 1.2
+WINDOWS = any(platform.win32_ver())
 
-EXTRA_RESERVED_VRAM = 200 * 1024 * 1024
-if any(platform.win32_ver()):
-    EXTRA_RESERVED_VRAM = 500 * 1024 * 1024 #Windows is higher because of the shared vram issue
+EXTRA_RESERVED_VRAM = 400 * 1024 * 1024
+if WINDOWS:
+    EXTRA_RESERVED_VRAM = 600 * 1024 * 1024 #Windows is higher because of the shared vram issue
 
 if args.reserve_vram is not None:
     EXTRA_RESERVED_VRAM = args.reserve_vram * 1024 * 1024 * 1024
@@ -383,6 +383,9 @@ if args.reserve_vram is not None:
 def extra_reserved_memory():
     return EXTRA_RESERVED_VRAM
 
+def minimum_inference_memory():
+    return (1024 * 1024 * 1024) * 0.8 + extra_reserved_memory()
+
 def unload_model_clones(model, unload_weights_only=True, force_unload=True):
     to_unload = []
     for i in range(len(current_loaded_models)):
@@ -405,6 +408,8 @@ def unload_model_clones(model, unload_weights_only=True, force_unload=True):
     if not force_unload:
         if unload_weights_only and unload_weight == False:
             return None
+    else:
+        unload_weight = True
 
     for i in to_unload:
         logging.debug("unload clone {} {}".format(i, unload_weight))
@@ -999,7 +1004,10 @@ def should_use_fp16(device=None, model_params=0, prioritize_performance=True, ma
     nvidia_10_series = ["1080", "1070", "titan x", "p3000", "p3200", "p4000", "p4200", "p5000", "p5200", "p6000", "1060", "1050", "p40", "p100", "p6", "p4"]
     for x in nvidia_10_series:
         if x in props.name.lower():
+            if WINDOWS or manual_cast:
                 return True
+            else:
+                return False #weird linux behavior where fp32 is faster
 
     if manual_cast:
         free_model_memory = maximum_vram_for_weights(device)

comfy/model_patcher.py

@@ -30,6 +30,18 @@ import comfy.model_management
 import comfy.lora
 from comfy.types import UnetWrapperFunction
 
+def string_to_seed(data):
+    crc = 0xFFFFFFFF
+    for byte in data:
+        if isinstance(byte, str):
+            byte = ord(byte)
+        crc ^= byte
+        for _ in range(8):
+            if crc & 1:
+                crc = (crc >> 1) ^ 0xEDB88320
+            else:
+                crc >>= 1
+    return crc ^ 0xFFFFFFFF
 
 def set_model_options_patch_replace(model_options, patch, name, block_name, number, transformer_index=None):
     to = model_options["transformer_options"].copy()
@@ -309,7 +321,7 @@ class ModelPatcher:
         else:
             temp_weight = weight.to(torch.float32, copy=True)
         out_weight = comfy.lora.calculate_weight(self.patches[key], temp_weight, key)
-        out_weight = comfy.float.stochastic_rounding(out_weight, weight.dtype)
+        out_weight = comfy.float.stochastic_rounding(out_weight, weight.dtype, seed=string_to_seed(key))
         if inplace_update:
             comfy.utils.copy_to_param(self.model, key, out_weight)
         else:

comfy/utils.py

@@ -528,6 +528,8 @@ def flux_to_diffusers(mmdit_config, output_prefix=""):
         ("guidance_in.out_layer.weight", "time_text_embed.guidance_embedder.linear_2.weight"),
         ("final_layer.adaLN_modulation.1.bias", "norm_out.linear.bias", swap_scale_shift),
         ("final_layer.adaLN_modulation.1.weight", "norm_out.linear.weight", swap_scale_shift),
+        ("pos_embed_input.bias", "controlnet_x_embedder.bias"),
+        ("pos_embed_input.weight", "controlnet_x_embedder.weight"),
     }
 
     for k in MAP_BASIC:

main.py

@@ -6,6 +6,10 @@ import importlib.util
 import folder_paths
 import time
 from comfy.cli_args import args
+from app.logger import setup_logger
+
+
+setup_logger(verbose=args.verbose)
 
 
 def execute_prestartup_script():

nodes.py

@@ -2129,3 +2129,5 @@ def init_extra_nodes(init_custom_nodes=True):
         else:
             logging.warning("Please do a: pip install -r requirements.txt")
         logging.warning("")
+    
+    return import_failed

notebooks/comfyui_colab.ipynb

@@ -79,7 +79,7 @@
         "#!wget -c https://huggingface.co/comfyanonymous/clip_vision_g/resolve/main/clip_vision_g.safetensors -P ./models/clip_vision/\n",
         "\n",
         "# SD1.5\n",
-        "!wget -c https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned-emaonly.ckpt -P ./models/checkpoints/\n",
+        "!wget -c https://huggingface.co/Comfy-Org/stable-diffusion-v1-5-archive/resolve/main/v1-5-pruned-emaonly-fp16.safetensors -P ./models/checkpoints/\n",
         "\n",
         "# SD2\n",
         "#!wget -c https://huggingface.co/stabilityai/stable-diffusion-2-1-base/resolve/main/v2-1_512-ema-pruned.safetensors -P ./models/checkpoints/\n",

script_examples/basic_api_example.py

@@ -43,7 +43,7 @@ prompt_text = """
     "4": {
         "class_type": "CheckpointLoaderSimple",
         "inputs": {
-            "ckpt_name": "v1-5-pruned-emaonly.ckpt"
+            "ckpt_name": "v1-5-pruned-emaonly.safetensors"
         }
     },
     "5": {

script_examples/websockets_api_example.py

@@ -41,11 +41,10 @@ def get_images(ws, prompt):
             continue #previews are binary data
 
     history = get_history(prompt_id)[prompt_id]
-    for o in history['outputs']:
     for node_id in history['outputs']:
         node_output = history['outputs'][node_id]
-            if 'images' in node_output:
         images_output = []
+        if 'images' in node_output:
             for image in node_output['images']:
                 image_data = get_image(image['filename'], image['subfolder'], image['type'])
                 images_output.append(image_data)
@@ -85,7 +84,7 @@ prompt_text = """
     "4": {
         "class_type": "CheckpointLoaderSimple",
         "inputs": {
-            "ckpt_name": "v1-5-pruned-emaonly.ckpt"
+            "ckpt_name": "v1-5-pruned-emaonly.safetensors"
         }
     },
     "5": {

script_examples/websockets_api_example_ws_images.py

@@ -81,7 +81,7 @@ prompt_text = """
     "4": {
         "class_type": "CheckpointLoaderSimple",
         "inputs": {
-            "ckpt_name": "v1-5-pruned-emaonly.ckpt"
+            "ckpt_name": "v1-5-pruned-emaonly.safetensors"
         }
     },
     "5": {

server.py

@@ -31,7 +31,6 @@ from model_filemanager import download_model, DownloadModelStatus
 from typing import Optional
 from api_server.routes.internal.internal_routes import InternalRoutes
 
-
 class BinaryEventTypes:
     PREVIEW_IMAGE = 1
     UNENCODED_PREVIEW_IMAGE = 2
@@ -42,6 +41,21 @@ async def send_socket_catch_exception(function, message):
     except (aiohttp.ClientError, aiohttp.ClientPayloadError, ConnectionResetError) as err:
         logging.warning("send error: {}".format(err))
 
+def get_comfyui_version():
+    comfyui_version = "unknown"
+    repo_path = os.path.dirname(os.path.realpath(__file__))
+    try:
+        import pygit2
+        repo = pygit2.Repository(repo_path)
+        comfyui_version = repo.describe(describe_strategy=pygit2.GIT_DESCRIBE_TAGS)
+    except Exception:
+        try:
+            import subprocess
+            comfyui_version = subprocess.check_output(["git", "describe", "--tags"], cwd=repo_path).decode('utf-8')
+        except Exception as e:
+            logging.warning(f"Failed to get ComfyUI version: {e}")
+    return comfyui_version.strip()
+
 @web.middleware
 async def cache_control(request: web.Request, handler):
     response: web.Response = await handler(request)
@@ -401,16 +415,20 @@ class PromptServer():
             return web.json_response(dt["__metadata__"])
 
         @routes.get("/system_stats")
-        async def get_queue(request):
+        async def system_stats(request):
             device = comfy.model_management.get_torch_device()
             device_name = comfy.model_management.get_torch_device_name(device)
             vram_total, torch_vram_total = comfy.model_management.get_total_memory(device, torch_total_too=True)
             vram_free, torch_vram_free = comfy.model_management.get_free_memory(device, torch_free_too=True)
+
             system_stats = {
                 "system": {
                     "os": os.name,
+                    "comfyui_version": get_comfyui_version(),
                     "python_version": sys.version,
-                    "embedded_python": os.path.split(os.path.split(sys.executable)[0])[1] == "python_embeded"
+                    "pytorch_version": comfy.model_management.torch_version,
+                    "embedded_python": os.path.split(os.path.split(sys.executable)[0])[1] == "python_embeded",
+                    "argv": sys.argv
                 },
                 "devices": [
                     {
@@ -586,7 +604,9 @@ class PromptServer():
         @routes.post("/internal/models/download")
         async def download_handler(request):
             async def report_progress(filename: str, status: DownloadModelStatus):
-                await self.send_json("download_progress", status.to_dict())
+                payload = status.to_dict()
+                payload['download_path'] = filename
+                await self.send_json("download_progress", payload)
 
             data = await request.json()
             url = data.get('url')

tests-unit/app_test/frontend_manager_test.py

@@ -1,6 +1,7 @@
 import argparse
 import pytest
 from requests.exceptions import HTTPError
+from unittest.mock import patch
 
 from app.frontend_management import (
     FrontendManager,
@@ -83,6 +84,35 @@ def test_init_frontend_invalid_provider():
     with pytest.raises(HTTPError):
         FrontendManager.init_frontend_unsafe(version_string)
 
+@pytest.fixture
+def mock_os_functions():
+    with patch('app.frontend_management.os.makedirs') as mock_makedirs, \
+         patch('app.frontend_management.os.listdir') as mock_listdir, \
+         patch('app.frontend_management.os.rmdir') as mock_rmdir:
+        mock_listdir.return_value = []  # Simulate empty directory
+        yield mock_makedirs, mock_listdir, mock_rmdir
+
+@pytest.fixture
+def mock_download():
+    with patch('app.frontend_management.download_release_asset_zip') as mock:
+        mock.side_effect = Exception("Download failed")  # Simulate download failure
+        yield mock
+
+def test_finally_block(mock_os_functions, mock_download, mock_provider):
+    # Arrange
+    mock_makedirs, mock_listdir, mock_rmdir = mock_os_functions
+    version_string = 'test-owner/test-repo@1.0.0'
+
+    # Act & Assert
+    with pytest.raises(Exception):
+        FrontendManager.init_frontend_unsafe(version_string, mock_provider)
+
+    # Assert
+    mock_makedirs.assert_called_once()
+    mock_download.assert_called_once()
+    mock_listdir.assert_called_once()
+    mock_rmdir.assert_called_once()
+
 
 def test_parse_version_string():
     version_string = "owner/repo@1.0.0"

tests/inference/test_execution.py

@@ -95,12 +95,11 @@ class ComfyClient:
                     pass # Probably want to store this off for testing
 
         history = self.get_history(prompt_id)[prompt_id]
-        for o in history['outputs']:
         for node_id in history['outputs']:
             node_output = history['outputs'][node_id]
             result.outputs[node_id] = node_output
-                if 'images' in node_output:
             images_output = []
+            if 'images' in node_output:
                 for image in node_output['images']:
                     image_data = self.get_image(image['filename'], image['subfolder'], image['type'])
                     image_obj = Image.open(BytesIO(image_data))

tests/inference/test_inference.py

@@ -109,11 +109,10 @@ class ComfyClient:
                 continue #previews are binary data
 
         history = self.get_history(prompt_id)[prompt_id]
-        for o in history['outputs']:
         for node_id in history['outputs']:
             node_output = history['outputs'][node_id]
-                if 'images' in node_output:
             images_output = []
+            if 'images' in node_output:
                 for image in node_output['images']:
                     image_data = self.get_image(image['filename'], image['subfolder'], image['type'])
                     images_output.append(image_data)

web/assets/userSelection-CyXKCVy3.js

@@ -0,0 +1,120 @@
+var __defProp = Object.defineProperty;
+var __name = (target, value) => __defProp(target, "name", { value, configurable: true });
+import { j as createSpinner, h as api, $ as $el } from "./index-CI3N807S.js";
+class UserSelectionScreen {
+  static {
+    __name(this, "UserSelectionScreen");
+  }
+  async show(users, user) {
+    const userSelection = document.getElementById("comfy-user-selection");
+    userSelection.style.display = "";
+    return new Promise((resolve) => {
+      const input = userSelection.getElementsByTagName("input")[0];
+      const select = userSelection.getElementsByTagName("select")[0];
+      const inputSection = input.closest("section");
+      const selectSection = select.closest("section");
+      const form = userSelection.getElementsByTagName("form")[0];
+      const error = userSelection.getElementsByClassName("comfy-user-error")[0];
+      const button = userSelection.getElementsByClassName(
+        "comfy-user-button-next"
+      )[0];
+      let inputActive = null;
+      input.addEventListener("focus", () => {
+        inputSection.classList.add("selected");
+        selectSection.classList.remove("selected");
+        inputActive = true;
+      });
+      select.addEventListener("focus", () => {
+        inputSection.classList.remove("selected");
+        selectSection.classList.add("selected");
+        inputActive = false;
+        select.style.color = "";
+      });
+      select.addEventListener("blur", () => {
+        if (!select.value) {
+          select.style.color = "var(--descrip-text)";
+        }
+      });
+      form.addEventListener("submit", async (e) => {
+        e.preventDefault();
+        if (inputActive == null) {
+          error.textContent = "Please enter a username or select an existing user.";
+        } else if (inputActive) {
+          const username = input.value.trim();
+          if (!username) {
+            error.textContent = "Please enter a username.";
+            return;
+          }
+          input.disabled = select.disabled = // @ts-expect-error
+          input.readonly = // @ts-expect-error
+          select.readonly = true;
+          const spinner = createSpinner();
+          button.prepend(spinner);
+          try {
+            const resp = await api.createUser(username);
+            if (resp.status >= 300) {
+              let message = "Error creating user: " + resp.status + " " + resp.statusText;
+              try {
+                const res = await resp.json();
+                if (res.error) {
+                  message = res.error;
+                }
+              } catch (error2) {
+              }
+              throw new Error(message);
+            }
+            resolve({ username, userId: await resp.json(), created: true });
+          } catch (err) {
+            spinner.remove();
+            error.textContent = err.message ?? err.statusText ?? err ?? "An unknown error occurred.";
+            input.disabled = select.disabled = // @ts-expect-error
+            input.readonly = // @ts-expect-error
+            select.readonly = false;
+            return;
+          }
+        } else if (!select.value) {
+          error.textContent = "Please select an existing user.";
+          return;
+        } else {
+          resolve({
+            username: users[select.value],
+            userId: select.value,
+            created: false
+          });
+        }
+      });
+      if (user) {
+        const name = localStorage["Comfy.userName"];
+        if (name) {
+          input.value = name;
+        }
+      }
+      if (input.value) {
+        input.focus();
+      }
+      const userIds = Object.keys(users ?? {});
+      if (userIds.length) {
+        for (const u of userIds) {
+          $el("option", { textContent: users[u], value: u, parent: select });
+        }
+        select.style.color = "var(--descrip-text)";
+        if (select.value) {
+          select.focus();
+        }
+      } else {
+        userSelection.classList.add("no-users");
+        input.focus();
+      }
+    }).then((r) => {
+      userSelection.remove();
+      return r;
+    });
+  }
+}
+window.comfyAPI = window.comfyAPI || {};
+window.comfyAPI.userSelection = window.comfyAPI.userSelection || {};
+window.comfyAPI.userSelection.UserSelectionScreen = UserSelectionScreen;
+export {
+  UserSelectionScreen
+};
+//# sourceMappingURL=userSelection-CyXKCVy3.js.map

web/assets/userSelection-GRU1gtOt.js

@@ -1,142 +0,0 @@
-var __defProp = Object.defineProperty;
-var __name = (target, value) => __defProp(target, "name", { value, configurable: true });
-var __async = (__this, __arguments, generator) => {
-  return new Promise((resolve, reject) => {
-    var fulfilled = (value) => {
-      try {
-        step(generator.next(value));
-      } catch (e) {
-        reject(e);
-      }
-    };
-    var rejected = (value) => {
-      try {
-        step(generator.throw(value));
-      } catch (e) {
-        reject(e);
-      }
-    };
-    var step = (x) => x.done ? resolve(x.value) : Promise.resolve(x.value).then(fulfilled, rejected);
-    step((generator = generator.apply(__this, __arguments)).next());
-  });
-};
-import { j as createSpinner, g as api, $ as $el } from "./index-CaD4RONs.js";
-const _UserSelectionScreen = class _UserSelectionScreen {
-  show(users, user) {
-    return __async(this, null, function* () {
-      const userSelection = document.getElementById("comfy-user-selection");
-      userSelection.style.display = "";
-      return new Promise((resolve) => {
-        const input = userSelection.getElementsByTagName("input")[0];
-        const select = userSelection.getElementsByTagName("select")[0];
-        const inputSection = input.closest("section");
-        const selectSection = select.closest("section");
-        const form = userSelection.getElementsByTagName("form")[0];
-        const error = userSelection.getElementsByClassName("comfy-user-error")[0];
-        const button = userSelection.getElementsByClassName(
-          "comfy-user-button-next"
-        )[0];
-        let inputActive = null;
-        input.addEventListener("focus", () => {
-          inputSection.classList.add("selected");
-          selectSection.classList.remove("selected");
-          inputActive = true;
-        });
-        select.addEventListener("focus", () => {
-          inputSection.classList.remove("selected");
-          selectSection.classList.add("selected");
-          inputActive = false;
-          select.style.color = "";
-        });
-        select.addEventListener("blur", () => {
-          if (!select.value) {
-            select.style.color = "var(--descrip-text)";
-          }
-        });
-        form.addEventListener("submit", (e) => __async(this, null, function* () {
-          var _a, _b, _c;
-          e.preventDefault();
-          if (inputActive == null) {
-            error.textContent = "Please enter a username or select an existing user.";
-          } else if (inputActive) {
-            const username = input.value.trim();
-            if (!username) {
-              error.textContent = "Please enter a username.";
-              return;
-            }
-            input.disabled = select.disabled = // @ts-expect-error
-            input.readonly = // @ts-expect-error
-            select.readonly = true;
-            const spinner = createSpinner();
-            button.prepend(spinner);
-            try {
-              const resp = yield api.createUser(username);
-              if (resp.status >= 300) {
-                let message = "Error creating user: " + resp.status + " " + resp.statusText;
-                try {
-                  const res = yield resp.json();
-                  if (res.error) {
-                    message = res.error;
-                  }
-                } catch (error2) {
-                }
-                throw new Error(message);
-              }
-              resolve({ username, userId: yield resp.json(), created: true });
-            } catch (err) {
-              spinner.remove();
-              error.textContent = (_c = (_b = (_a = err.message) != null ? _a : err.statusText) != null ? _b : err) != null ? _c : "An unknown error occurred.";
-              input.disabled = select.disabled = // @ts-expect-error
-              input.readonly = // @ts-expect-error
-              select.readonly = false;
-              return;
-            }
-          } else if (!select.value) {
-            error.textContent = "Please select an existing user.";
-            return;
-          } else {
-            resolve({
-              username: users[select.value],
-              userId: select.value,
-              created: false
-            });
-          }
-        }));
-        if (user) {
-          const name = localStorage["Comfy.userName"];
-          if (name) {
-            input.value = name;
-          }
-        }
-        if (input.value) {
-          input.focus();
-        }
-        const userIds = Object.keys(users != null ? users : {});
-        if (userIds.length) {
-          for (const u of userIds) {
-            $el("option", { textContent: users[u], value: u, parent: select });
-          }
-          select.style.color = "var(--descrip-text)";
-          if (select.value) {
-            select.focus();
-          }
-        } else {
-          userSelection.classList.add("no-users");
-          input.focus();
-        }
-      }).then((r) => {
-        userSelection.remove();
-        return r;
-      });
-    });
-  }
-};
-__name(_UserSelectionScreen, "UserSelectionScreen");
-let UserSelectionScreen = _UserSelectionScreen;
-window.comfyAPI = window.comfyAPI || {};
-window.comfyAPI.userSelection = window.comfyAPI.userSelection || {};
-window.comfyAPI.userSelection.UserSelectionScreen = UserSelectionScreen;
-export {
-  UserSelectionScreen
-};
-//# sourceMappingURL=userSelection-GRU1gtOt.js.map

web/index.html

@@ -14,8 +14,8 @@
 		</style> -->
 
 
-
-		<link rel="stylesheet" type="text/css" href="user.css" />
+
+		<link rel="stylesheet" type="text/css" href="user.css" />
 		<link rel="stylesheet" type="text/css" href="materialdesignicons.min.css" />
 		<script type="module" crossorigin src="./assets/index-CI3N807S.js"></script>
 		<link rel="stylesheet" crossorigin href="./assets/index-_5czGnTA.css">