ReadyArt
/

Omega-Evolution-27B-v2.2-GGUF

GGUF

nsfw

explicit

Model card Files Files and versions

xet

Community

Instructions to use ReadyArt/Omega-Evolution-27B-v2.2-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

llama-cpp-python

How to use ReadyArt/Omega-Evolution-27B-v2.2-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="ReadyArt/Omega-Evolution-27B-v2.2-GGUF",
	filename="Omega-Evolution-27B-v2.2-Q3_K_M_attn8_ssm8.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use ReadyArt/Omega-Evolution-27B-v2.2-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf ReadyArt/Omega-Evolution-27B-v2.2-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf ReadyArt/Omega-Evolution-27B-v2.2-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 ReadyArt/Omega-Evolution-27B-v2.2-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf ReadyArt/Omega-Evolution-27B-v2.2-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 ReadyArt/Omega-Evolution-27B-v2.2-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf ReadyArt/Omega-Evolution-27B-v2.2-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 ReadyArt/Omega-Evolution-27B-v2.2-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ReadyArt/Omega-Evolution-27B-v2.2-GGUF:Q4_K_M

Use Docker

docker model run hf.co/ReadyArt/Omega-Evolution-27B-v2.2-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use ReadyArt/Omega-Evolution-27B-v2.2-GGUF with Ollama:
```
ollama run hf.co/ReadyArt/Omega-Evolution-27B-v2.2-GGUF:Q4_K_M
```

Unsloth Studio

How to use ReadyArt/Omega-Evolution-27B-v2.2-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 ReadyArt/Omega-Evolution-27B-v2.2-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 ReadyArt/Omega-Evolution-27B-v2.2-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for ReadyArt/Omega-Evolution-27B-v2.2-GGUF to start chatting

How to use ReadyArt/Omega-Evolution-27B-v2.2-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf ReadyArt/Omega-Evolution-27B-v2.2-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": "ReadyArt/Omega-Evolution-27B-v2.2-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use ReadyArt/Omega-Evolution-27B-v2.2-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 ReadyArt/Omega-Evolution-27B-v2.2-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 ReadyArt/Omega-Evolution-27B-v2.2-GGUF:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use ReadyArt/Omega-Evolution-27B-v2.2-GGUF with Docker Model Runner:
```
docker model run hf.co/ReadyArt/Omega-Evolution-27B-v2.2-GGUF:Q4_K_M
```

Lemonade

How to use ReadyArt/Omega-Evolution-27B-v2.2-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ReadyArt/Omega-Evolution-27B-v2.2-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Omega-Evolution-27B-v2.2-GGUF-Q4_K_M

List all available models

lemonade list

Will there be MTP support?

by prjsingularity - opened 15 days ago

Discussion

prjsingularity

15 days ago

Qwen 3.5 supports MTP, which is a very good booster for tps, but it's present in the new checkpoints, maybe not from those these project was made. so i'm curious, do you plan for implementation of MTP?

FrenzyBiscuit

Ready.Art org 14 days ago

Qwen 3.5 supports MTP, which is a very good booster for tps, but it's present in the new checkpoints, maybe not from those these project was made. so i'm curious, do you plan for implementation of MTP?

Can you provide a step by step guide with axolotl?

At the time this model was trained we trained on all layers. MTP should be supported. If it's not, it's because axolotl didn't support it at the time.

FrenzyBiscuit

Ready.Art org 14 days ago

Infact, it does appear the MTP layers are in the model weights.

What issue are you running into?

FrenzyBiscuit

Ready.Art org 14 days ago

Hmm... do we need to recreate the GGUF for MTP support?

@gecfdo thoughts?

llmfan46

14 days ago

Hmm... do we need to recreate the GGUF for MTP support?

@gecfdo thoughts?

Old llama.cpp used to just drop all MTPs, so the old GGUFs should normally not have any MTPs.

FrenzyBiscuit

Ready.Art org 13 days ago

Hmm... do we need to recreate the GGUF for MTP support?

@gecfdo thoughts?

Old llama.cpp used to just drop all MTPs, so the old GGUFs should normally not have any MTPs.

That's the conclusion I've come to after reviewing the GGUF data, yeah.

I'm mulling on how to handle this.

Also, getting ready to make training data for qwen 3.6 27B so... ;)

veldierin

13 days ago

FYI, it is also possible to graft the MTP heads on as long as the finetune hasnt diverged too far from the norm. i successfully tested it out last night.

grab the convert.py and the 27B_MTP.gguf from here:https://huggingface.co/havenoammo/Qwen3.6-27B-MTP-UD-GGUF

then run it like this:
python.exe convert.py .\Qwen3.6-27B-NEO-CODE-HERE-2T-OT-Q5_K_S.gguf .\27B_MTP.gguf .\Qwen3.6-27B-NEO-CODE-HERE-2T-OT-MTP-Q5_K_S.gguf
Reading target: .\Qwen3.6-27B-NEO-CODE-HERE-2T-OT-Q5_K_S.gguf
Reading source: .\27B_MTP.gguf
Target tensors: 851, KVs: 42
Source tensors: 15, KVs: 24

Arch: qwen35
Target block_count: 64
Source block_count: 65, nextn_predict_layers: 1

Extra tensors to transplant: 15

Writing output: .\Qwen3.6-27B-NEO-CODE-HERE-2T-OT-MTP-Q5_K_S.gguf
Copying 866 tensors...
Copied 50/866 tensors
Copied 100/866 tensors
Copied 150/866 tensors
Copied 200/866 tensors
Copied 250/866 tensors
Copied 300/866 tensors
Copied 350/866 tensors
Copied 400/866 tensors
Copied 450/866 tensors
Copied 500/866 tensors
Copied 550/866 tensors
Copied 600/866 tensors
Copied 650/866 tensors
Copied 700/866 tensors
Copied 750/866 tensors
Copied 800/866 tensors
Copied 850/866 tensors
Copied 866/866 tensors

Output: .\Qwen3.6-27B-NEO-CODE-HERE-2T-OT-MTP-Q5_K_S.gguf
Size: 19.44 GB
Tensors: 866

Validating output...
Spot-checking tensor data integrity...
token_embd.weight: OK (28225bb43049ef5c)
blk.64.nextn.eh_proj.weight: OK (f89b048f279f958b)
OK — all checks passed

Done. Output: .\Qwen3.6-27B-NEO-CODE-HERE-2T-OT-MTP-Q5_K_S.gguf

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