flowtyone / comfyui-flowty-lcm

LCM test nodes for comfyui

comfyui-flowty-lcm

This is a comfyui early testing node for LCM, adapted from https://github.com/0xbitches/sd-webui-lcm

It uses the diffusers backend unfortunately and not comfy's model loading mechanism. But the intention here is just to be able to execute lcm inside comfy.

You need to use a special lcm model for this - the only available checkpoint currently is LCM_Dreamshaper_v7.

comfyui-flowty-lcm's Issues

TypeError: slice indices must be integers or None or have an __index__ method

diffuser:0.23.0 or 0.24.0.dev0 , all has errors below:
File "/data/ComfyUI/custom_nodes/ComfyUI-LCM/lcm/lcm_i2i_pipeline.py", line 304, in call
self.scheduler.set_timesteps(strength, num_inference_steps, lcm_origin_steps)
File "/data/miniconda3/envs/env-novelai/lib/python3.10/site-packages/diffusers/schedulers/scheduling_lcm.py", line 388, in set_timesteps
timesteps = lcm_origin_timesteps[::-skipping_step][:num_inference_steps]

Error occurred when executing LCMSampler:

already download LCM_Dreamshaper_v7.safetensors, where should I put it to:
..\ComfyUI\models\checkpoints
..\ComfyUI\models\checkpoints\SimianLuo
..\hub\models--SimianLuo--LCM_Dreamshaper_v7\refs\main
..\hub\models--SimianLuo--LCM_Dreamshaper_v7\refs\main\\SimianLuo
All these path are not work

Error occurred when executing LCMSampler:

[Errno 13] Permission denied: 'I:\sdmodel\Stable-diffusion\cp\hub\models--SimianLuo--LCM_Dreamshaper_v7\refs\main'

File "F:\ComfyUI\ComfyUI\execution.py", line 153, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
File "F:\ComfyUI\ComfyUI\execution.py", line 83, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
File "F:\ComfyUI\ComfyUI\execution.py", line 76, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
File "F:\ComfyUI\ComfyUI\custom_nodes\comfyui-flowty-lcm\nodes.py", line 43, in sample
self.pipe = LatentConsistencyModelPipeline.from_pretrained(
File "F:\ComfyUI\python_embeded\lib\site-packages\diffusers\pipelines\pipeline_utils.py", line 939, in from_pretrained
cached_folder = cls.download(
File "F:\ComfyUI\python_embeded\lib\site-packages\diffusers\pipelines\pipeline_utils.py", line 1599, in download
cached_folder = snapshot_download(
File "F:\ComfyUI\python_embeded\lib\site-packages\huggingface_hub\utils_validators.py", line 118, in _inner_fn
return fn(*args, **kwargs)
File "F:\ComfyUI\python_embeded\lib\site-packages\huggingface_hub_snapshot_download.py", line 169, in snapshot_download
with open(ref_path) as f:

Path to Work

notes.py:
`from .lcm.lcm_scheduler import LCMScheduler
from .lcm.lcm_pipeline import LatentConsistencyModelPipeline
from os import path
import time
import torch
import random
import numpy as np
from comfy.model_management import get_torch_device

MAX_SEED = np.iinfo(np.int32).max

def randomize_seed_fn(seed: int, randomize_seed: bool) -> int:
if randomize_seed:
seed = random.randint(0, MAX_SEED)
return seed

class LCMSampler:

def __init__(self):
    self.scheduler = LCMScheduler.from_pretrained(path.join(path.dirname(__file__), "scheduler_config.json"))
    self.pipe = None

@classmethod
def INPUT_TYPES(s):
    return {"required":
                {
                "seed": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
                "steps": ("INT", {"default": 4, "min": 1, "max": 10000}),
                "cfg": ("FLOAT", {"default": 8.0, "min": 0.0, "max": 100.0, "step":0.5, "round": 0.01}),
                "size": ("INT", {"default": 512, "min": 512, "max": 768}),
                "num_images": ("INT", {"default": 1, "min": 1, "max": 64}),
                 "positive_prompt": ("STRING", {"multiline": True}),
                }
            }

RETURN_TYPES = ("IMAGE",)
FUNCTION = "sample"

CATEGORY = "sampling"

def sample(self, seed, steps, cfg, positive_prompt, size, num_images):
    if self.pipe is None:
        self.pipe = LatentConsistencyModelPipeline.from_pretrained(
            pretrained_model_name_or_path="C:\Matrix\Data\Models\LCM_Dreamshaper_v7",
            local_files_only=True,
            scheduler=self.scheduler`

Change model path to standard path

Tried your node today and the script burped based on the model directory.
Typically, we use ComfyUIs path or StableDiffusions directory..

FileNotFoundError: [Errno 2] No such file or directory: 'foo\.cache\huggingface\hub\models--SimianLuo--LCM_Dreamshaper_v7\refs\main'

Use these paths instead:

ComfyUI: foo:\ComfyUI\models\checkpoints
Auto1111: foo:\stable-diffusion-webui\models\Stable-diffusion

Cant seems to get the nodes after installation

Can not find nodes.
I tried to open the sample.png (not sure if its saved as embedded workflow) but it did not work too.