Instructions to use Emerald7664/Palette-RP-4B-2609-v0.1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Emerald7664/Palette-RP-4B-2609-v0.1 with Transformers:

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

pipe = pipeline("text-generation", model="Emerald7664/Palette-RP-4B-2609-v0.1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Emerald7664/Palette-RP-4B-2609-v0.1")
model = AutoModelForCausalLM.from_pretrained("Emerald7664/Palette-RP-4B-2609-v0.1", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use Emerald7664/Palette-RP-4B-2609-v0.1 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 Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Emerald7664/Palette-RP-4B-2609-v0.1: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 Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Emerald7664/Palette-RP-4B-2609-v0.1: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 Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M

Use Docker

docker model run hf.co/Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M

LM Studio
Jan

vLLM

How to use Emerald7664/Palette-RP-4B-2609-v0.1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Emerald7664/Palette-RP-4B-2609-v0.1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Emerald7664/Palette-RP-4B-2609-v0.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M

SGLang

How to use Emerald7664/Palette-RP-4B-2609-v0.1 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 "Emerald7664/Palette-RP-4B-2609-v0.1" \
    --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": "Emerald7664/Palette-RP-4B-2609-v0.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "Emerald7664/Palette-RP-4B-2609-v0.1" \
        --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": "Emerald7664/Palette-RP-4B-2609-v0.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use Emerald7664/Palette-RP-4B-2609-v0.1 with Ollama:
```
ollama run hf.co/Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M
```
Unsloth Desktop

How to use Emerald7664/Palette-RP-4B-2609-v0.1 with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M

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": "Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use Emerald7664/Palette-RP-4B-2609-v0.1 with Docker Model Runner:
```
docker model run hf.co/Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M
```

Lemonade

How to use Emerald7664/Palette-RP-4B-2609-v0.1 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M

Run and chat with the model

lemonade run user.Palette-RP-4B-2609-v0.1-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use Emerald7664/Palette-RP-4B-2609-v0.1 with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Emerald7664/Palette-RP-4B-2609-v0.1: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 Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use Emerald7664/Palette-RP-4B-2609-v0.1 with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M

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 "Emerald7664/Palette-RP-4B-2609-v0.1:Q4_K_M" \
  --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"

Palette: New Series of Small Model Creative Distills With Only The Best Teachers!

Quick Overview: What is Palette-RP-4B-2609-v0.1?

Palette-RP-4B-2609-v0.1 is Qwen 3.5 4B model trained on roleplay data generated by Hy4, why v0.1? Because it's my first try with it, I will be still adjusting training parameters.

It is recommended to keep thinking OFF as all the training data had it turned off, Hy4 scored anomalously high among models with thinking disabled, hence why I choose it to distill, I may give it a try again next month, the data I was able to get is already massive, but I also contemplate going for three times more data next, which will massively hit my wallet.

Still needs a lot of testing in my opinion, I am also experimenting with LFM 2.5 2.6B again, and of course, I will release the next finetune of it with thinking fully removed.

Quants are as always in the repo, safetensors are there too.

Now, on to the advantages this model holds(according to my plan, I dunno if its a success yet lol):

Very strong RP and EPR: The teacher model was very unaligned, which is not surprising, I have not found a single refusal across over 25 thousand assisstant completions.
Vivid prose: The teacher model is a massive 770B A49B MoE, its RP feels incredibly mature.
Improved in-roleplay intelligence: For the reason spoken above, the teacher model is extremely intelligent, which the student also(partially) inherits.

Downloads last month: 515

Safetensors

Model size

4B params

Tensor type

BF16

Model tree for Emerald7664/Palette-RP-4B-2609-v0.1

Base model

Qwen/Qwen3.5-4B-Base

Finetuned

Qwen/Qwen3.5-4B

Finetuned

(840)

this model