Instructions to use Jackrong/Qwopus3.6-27B-v2-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Jackrong/Qwopus3.6-27B-v2-GGUF with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="Jackrong/Qwopus3.6-27B-v2-GGUF")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("Jackrong/Qwopus3.6-27B-v2-GGUF", dtype="auto")

llama-cpp-python

How to use Jackrong/Qwopus3.6-27B-v2-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Jackrong/Qwopus3.6-27B-v2-GGUF",
	filename="Qwopus3.6-27B-v2-IQ4_XS.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": [
				{
					"type": "text",
					"text": "Describe this image in one sentence."
				},
				{
					"type": "image_url",
					"image_url": {
						"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
					}
				}
			]
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use Jackrong/Qwopus3.6-27B-v2-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M

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 Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M

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 Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M

Use Docker

docker model run hf.co/Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use Jackrong/Qwopus3.6-27B-v2-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Jackrong/Qwopus3.6-27B-v2-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Jackrong/Qwopus3.6-27B-v2-GGUF",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M

SGLang

How to use Jackrong/Qwopus3.6-27B-v2-GGUF with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Jackrong/Qwopus3.6-27B-v2-GGUF" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Jackrong/Qwopus3.6-27B-v2-GGUF",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Jackrong/Qwopus3.6-27B-v2-GGUF" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Jackrong/Qwopus3.6-27B-v2-GGUF",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Ollama
How to use Jackrong/Qwopus3.6-27B-v2-GGUF with Ollama:
```
ollama run hf.co/Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M
```

Unsloth Studio new

How to use Jackrong/Qwopus3.6-27B-v2-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for Jackrong/Qwopus3.6-27B-v2-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for Jackrong/Qwopus3.6-27B-v2-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Jackrong/Qwopus3.6-27B-v2-GGUF to start chatting

Pi new

How to use Jackrong/Qwopus3.6-27B-v2-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use Jackrong/Qwopus3.6-27B-v2-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M

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 Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use Jackrong/Qwopus3.6-27B-v2-GGUF with Docker Model Runner:
```
docker model run hf.co/Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M
```

Lemonade

How to use Jackrong/Qwopus3.6-27B-v2-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Jackrong/Qwopus3.6-27B-v2-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Qwopus3.6-27B-v2-GGUF-Q4_K_M

List all available models

lemonade list

Significantly Lower Context Tokens (nctx) Compared to preview-v1 Variant?

by neoresin - opened 1 day ago

Discussion

neoresin

1 day ago

I have been eagerly awaiting this model for so long! Thank you and Kyle Hessling so very much for performing this work and creating this model! 🙇 It's truly awesome.

I've been using the preview-v1 model for a fair bit with Opencode and I was able to hit 166400 tokens before compaction pretty reliably which gave a good size context window to work with. With this model, though, same Q4_K_M quantization, I only hit ~104000 tokens before it OOMs on my 4090.

Here's my commandline for both models:
/usr/bin/llama-server --port 7996 --ctx-size 166400 --fit on --cache-type-k q8_0 --cache-type-v q8_0 -fa on --api-key 'XXX' --repeat-penalty 1.0 --temp 0.6 --top-p 0.95 --min-p 0.0 --top-k 20 --presence-penalty 0.0 --image-min-tokens 1024 --chat-template-kwargs '{"preserve_thinking":true}' --reasoning on --jinja --chat-template-file '/dir/Qwen-Fixed-Chat-Templates/chat_template.jinja' --mmproj '/models/Qwopus3.6-27B-v1-preview-mmproj.gguf' --model '/models/Qwopus3.6-27B-v1-preview-Q4_K_M.gguf'
/usr/bin/llama-server --port 8076 --ctx-size 166400 --fit on --cache-type-k q8_0 --cache-type-v q8_0 -fa on --api-key 'XXX' --repeat-penalty 1.0 --temp 0.6 --top-p 0.95 --min-p 0.0 --top-k 20 --presence-penalty 0.0 --image-min-tokens 1024 --chat-template-kwargs '{"preserve_thinking":true}' --reasoning on --jinja --chat-template-file '/dir/Qwen-Fixed-Chat-Templates/chat_template.jinja' --mmproj '/models/Qwopus3.6-27B-v2-mmproj.gguf' --model '/models/Qwopus3.6-27B-v2-Q4_K_M.gguf'

I'm not claiming to know everything when it comes to all this, but it just seems strange that it would drop by so much if it's nearly the same model as the preview-v1. Is there an explanation for this? Thanks in advance for any time spent! Sorry if I missed anything in your Model card that should explain this.

bardd

1 day ago

i am thinking to use this model Q4_K_M with https://github.com/TheTom/llama-cpp-turboquant
hoping to get 196k context, i have 2 X T4 (15 GB VRAM each) do you think i should use this model ?
did you saw any quality difference in the output?
i have never used these models before
i am thinking of testing these
also did you found out anything about the context problem you are facing ?

neoresin

1 day ago

•

edited 1 day ago

I just bailed and went back to the v1-preview, which still seems to work for me. It gets stuck in loops sometimes so I have to babysit it (playing with the --repeat-penalty at the moment [set to 1.1 seems to work, though the evaluation is ever-evolving]), but darn it: it's a good model. 😀 Using normal llama-server, though I'm intrigued by explorations into the turboquant or rotorquant arena. MTP didn't really do it for me, personally (while generation might be higher, the prompt processing tanked for me, so I didn't go that route), not that they're the same.
Have you seen good results with turboquant elsewhere?
I should say that I'm interested in getting this working with the context window that v1-preview has, if that's possible or if anyone has any good ideas why it changed from v1-preview to v2.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment