Instructions to use hipinis/20260718 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use hipinis/20260718 with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf hipinis/20260718:Q8_0
# Run inference directly in the terminal:
llama cli -hf hipinis/20260718:Q8_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf hipinis/20260718:Q8_0
# Run inference directly in the terminal:
llama cli -hf hipinis/20260718:Q8_0

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf hipinis/20260718:Q8_0
# Run inference directly in the terminal:
./llama-cli -hf hipinis/20260718:Q8_0

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf hipinis/20260718:Q8_0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf hipinis/20260718:Q8_0

Use Docker

docker model run hf.co/hipinis/20260718:Q8_0

LM Studio
Jan
Ollama
How to use hipinis/20260718 with Ollama:
```
ollama run hf.co/hipinis/20260718:Q8_0
```
Unsloth Desktop

How to use hipinis/20260718 with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf hipinis/20260718:Q8_0

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "hipinis/20260718:Q8_0"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use hipinis/20260718 with Docker Model Runner:
```
docker model run hf.co/hipinis/20260718:Q8_0
```

Lemonade

How to use hipinis/20260718 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull hipinis/20260718:Q8_0

Run and chat with the model

lemonade run user.20260718-Q8_0

List all available models

lemonade list

Hermes Agent

How to use hipinis/20260718 with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf hipinis/20260718:Q8_0

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default hipinis/20260718:Q8_0

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use hipinis/20260718 with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf hipinis/20260718:Q8_0

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "hipinis/20260718:Q8_0" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

20260718 / custom_nodes /ComfyUI-VideoHelperSuite /testframework /server.py

hipinis

Upload 56 files

3088c65 verified 3 months ago

Raw History Blame Contribute Delete

2.9 kB

import subprocess
import json
import os
import torch
import shutil

import server
import folder_paths

web = server.web

@server.PromptServer.instance.routes.post("/VHS_test")
async def test(request):
    try:
        req_data = await request.json()
        output = req_data['output']['gifs'][0]
        filename = output['filename']
        typ = output['type']
        base_args = ["ffprobe", "-v", "error", '-count_packets', "-show_entries", "stream", "-of", "json"]
        video = folder_paths.get_annotated_filepath(f'{filename} [{typ}]')
        vprobe = json.loads(subprocess.run(base_args + ['-select_streams', 'v:0', video],
                             capture_output=True, check=True).stdout)['streams'][0]
        aprobe = json.loads(subprocess.run(base_args + ['-select_streams', 'a:0', video],
                             capture_output=True, check=True).stdout)['streams']
        probe = {'video': vprobe}
        if len(aprobe) > 0:
            probe['audio'] = aprobe[0]
        errors = []
        compare = None
        for test in req_data['tests']:
            if test['type'] == 'compare':
                compare = test
                continue
            key = test['key']
            expected = test['value']
            actual = probe[test['type']][key]
            if expected != actual:
                #Consider always dumping type?
                errors.append(f'{key}: {expected} != {actual}')
        if len(errors) == 0 and compare is not None:
            if not os.path.exists(compare['filename']):
                os.makedirs(os.path.split(compare['filename'])[0], exist_ok=True)
                shutil.copy(video, compare['filename'])
                print("Missing comparison file has been initialized from output:", os.path.abspath(compare['filename']))
            else:
                #NOTE: This does not include the full memory optimizations of VHS
                #Tests should be small
                #TODO: Figure out way to do opacity comparison. May need to do blending in python
                #(easy, but slower and more memory intensive)
                diff = subprocess.run(['ffmpeg', '-v', 'error', '-i', video, '-i', compare['filename'], '-filter_complex', 'blend=all_mode=grainextract', '-pix_fmt', 'rgb24', '-f', 'rawvideo', '-'], stdout=subprocess.PIPE, check=True).stdout
                diff = torch.frombuffer(diff, dtype=torch.uint8).to(dtype=torch.float32).div_(255)
                #diff = diff.reshape((-1,4))
                d = (diff-0.5).abs().sum()/diff.size(0)
                if d > compare['tolerance']:
                    errors.append(f'Similarity is outside specified tolerance: {d}')
                else:
                    print('d:', d)
        return web.json_response(errors)
    except Exception as e:
        return web.json_response(str(e))

	import subprocess
	import json
	import os
	import torch
	import shutil

	import server
	import folder_paths

	web = server.web

	@server.PromptServer.instance.routes.post("/VHS_test")
	async def test(request):
	try:
	req_data = await request.json()
	output = req_data['output']['gifs'][0]
	filename = output['filename']
	typ = output['type']
	base_args = ["ffprobe", "-v", "error", '-count_packets', "-show_entries", "stream", "-of", "json"]
	video = folder_paths.get_annotated_filepath(f'{filename} [{typ}]')
	vprobe = json.loads(subprocess.run(base_args + ['-select_streams', 'v:0', video],
	capture_output=True, check=True).stdout)['streams'][0]
	aprobe = json.loads(subprocess.run(base_args + ['-select_streams', 'a:0', video],
	capture_output=True, check=True).stdout)['streams']
	probe = {'video': vprobe}
	if len(aprobe) > 0:
	probe['audio'] = aprobe[0]
	errors = []
	compare = None
	for test in req_data['tests']:
	if test['type'] == 'compare':
	compare = test
	continue
	key = test['key']
	expected = test['value']
	actual = probe[test['type']][key]
	if expected != actual:
	#Consider always dumping type?
	errors.append(f'{key}: {expected} != {actual}')
	if len(errors) == 0 and compare is not None:
	if not os.path.exists(compare['filename']):
	os.makedirs(os.path.split(compare['filename'])[0], exist_ok=True)
	shutil.copy(video, compare['filename'])
	print("Missing comparison file has been initialized from output:", os.path.abspath(compare['filename']))
	else:
	#NOTE: This does not include the full memory optimizations of VHS
	#Tests should be small
	#TODO: Figure out way to do opacity comparison. May need to do blending in python
	#(easy, but slower and more memory intensive)
	diff = subprocess.run(['ffmpeg', '-v', 'error', '-i', video, '-i', compare['filename'], '-filter_complex', 'blend=all_mode=grainextract', '-pix_fmt', 'rgb24', '-f', 'rawvideo', '-'], stdout=subprocess.PIPE, check=True).stdout
	diff = torch.frombuffer(diff, dtype=torch.uint8).to(dtype=torch.float32).div_(255)
	#diff = diff.reshape((-1,4))
	d = (diff-0.5).abs().sum()/diff.size(0)
	if d > compare['tolerance']:
	errors.append(f'Similarity is outside specified tolerance: {d}')
	else:
	print('d:', d)
	return web.json_response(errors)
	except Exception as e:
	return web.json_response(str(e))