Instructions to use rockus/Poocha-E4B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use rockus/Poocha-E4B-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="rockus/Poocha-E4B-GGUF",
	filename="kat-E4B-F16.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use rockus/Poocha-E4B-GGUF 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 rockus/Poocha-E4B-GGUF:F16
# Run inference directly in the terminal:
llama cli -hf rockus/Poocha-E4B-GGUF:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf rockus/Poocha-E4B-GGUF:F16
# Run inference directly in the terminal:
llama cli -hf rockus/Poocha-E4B-GGUF:F16

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 rockus/Poocha-E4B-GGUF:F16
# Run inference directly in the terminal:
./llama-cli -hf rockus/Poocha-E4B-GGUF:F16

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 rockus/Poocha-E4B-GGUF:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf rockus/Poocha-E4B-GGUF:F16

Use Docker

docker model run hf.co/rockus/Poocha-E4B-GGUF:F16

LM Studio
Jan

vLLM

How to use rockus/Poocha-E4B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "rockus/Poocha-E4B-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": "rockus/Poocha-E4B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/rockus/Poocha-E4B-GGUF:F16

Ollama
How to use rockus/Poocha-E4B-GGUF with Ollama:
```
ollama run hf.co/rockus/Poocha-E4B-GGUF:F16
```

Unsloth Studio

How to use rockus/Poocha-E4B-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 rockus/Poocha-E4B-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 rockus/Poocha-E4B-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for rockus/Poocha-E4B-GGUF to start chatting

How to use rockus/Poocha-E4B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf rockus/Poocha-E4B-GGUF:F16

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": "rockus/Poocha-E4B-GGUF:F16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use rockus/Poocha-E4B-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf rockus/Poocha-E4B-GGUF:F16

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 rockus/Poocha-E4B-GGUF:F16

Run Hermes

hermes

Atomic Chat new
Docker Model Runner
How to use rockus/Poocha-E4B-GGUF with Docker Model Runner:
```
docker model run hf.co/rockus/Poocha-E4B-GGUF:F16
```

Lemonade

How to use rockus/Poocha-E4B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull rockus/Poocha-E4B-GGUF:F16

Run and chat with the model

lemonade run user.Poocha-E4B-GGUF-F16

List all available models

lemonade list

🐱 Poocha-E4B — GGUF

GGUF quantizations of rockus/Poocha-E4B — a child-friendly, India-localized science & nature tutor. The assistant persona is Poocha (പൂച്ച, "cat"), a clever, curious kitten who teaches children (ages ~8–12) using everyday Indian analogies. Fine-tuned from Gemma 4 E4B. Runs locally with llama.cpp · Ollama · LM Studio · Jan.

📦 Files

File	Quant	Bits	Size	Use case
`kat-E4B-F16.gguf`	F16	16	~16 GB	Full precision — maximum fidelity / base for re-quantizing
`kat-E4B-Q8_0.gguf`	Q8_0	8	8.0 GB	Near-lossless reference quality
`kat-E4B-Q6_K.gguf`	Q6_K	6	6.2 GB	⭐ Recommended — fits a 12 GB GPU with KV headroom; the deployed default

All three fit a single 12 GB GPU (e.g. RTX 4070 SUPER). Q6_K is the deploy default; Q8_0/F16 trade size for fidelity. (IQ / sub-4-bit quants are intentionally not shipped — at this size there's no need to squeeze, and they'd cost inference speed.)

▶️ Run (llama.cpp)

llama-server -m kat-E4B-Q6_K.gguf -ngl 99 --port 8080

🎛️ Recommended sampling

🔬 Factual / Q&A: temperature 0.30, min_p 0.08, top_k 0, top_p 1.0
🚀 Adventure / story: temperature 0.95, min_p 0.05
Add repetition_penalty ≈ 1.15 for cleaner long outputs.

System prompt:

You are Poocha, a clever, curious little kitten who teaches Indian children (ages 8-12) about science. Warm, encouraging, plain-spoken. You may use a gentle purr or meow OCCASIONALLY. Use simple Indian examples.

📊 Quality (Round-3 E4B)

ARC-Challenge-Indic (English): 90.5% science accuracy
Engagement loop ("what should we explore next?") in 92% of answers; 0% dry (persona always present)
train_loss ≈ 0.226, eval_loss ≈ 0.799

📚 Training

Trained on a ~12k-row multi-corpus Poocha set, every row in Poocha's voice: a cleaned first-round set + NCERT 6–9 + Science Journal for Kids + Tushe/Siyavula passages re-narrated in Poocha's voice (not raw text) + interactive adventures + behaviours. See the base model card for the full corpus breakdown, data design, and licenses.

License

Apache-2.0 (inherited from Gemma 4).

Downloads last month: -

GGUF

Model size

8B params

Architecture

gemma4

Hardware compatibility

6-bit

8-bit

16-bit

Model tree for rockus/Poocha-E4B-GGUF

Base model

google/gemma-4-E4B

Finetuned

google/gemma-4-E4B-it

Finetuned

rockus/Poocha-E4B

Quantized

(1)

this model