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

Libraries

How to use ritiksuman/rudra with Transformers:

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

pipe = pipeline("text-generation", model="ritiksuman/rudra")
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)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("ritiksuman/rudra")
model = AutoModelForMultimodalLM.from_pretrained("ritiksuman/rudra", device_map="auto")
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?"}
        ]
    },
]
inputs = processor.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=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use ritiksuman/rudra 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 ritiksuman/rudra:BF16
# Run inference directly in the terminal:
llama cli -hf ritiksuman/rudra:BF16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf ritiksuman/rudra:BF16
# Run inference directly in the terminal:
llama cli -hf ritiksuman/rudra:BF16

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 ritiksuman/rudra:BF16
# Run inference directly in the terminal:
./llama-cli -hf ritiksuman/rudra:BF16

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 ritiksuman/rudra:BF16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ritiksuman/rudra:BF16

Use Docker

docker model run hf.co/ritiksuman/rudra:BF16

LM Studio
Jan

vLLM

How to use ritiksuman/rudra with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/ritiksuman/rudra:BF16

SGLang

How to use ritiksuman/rudra 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 "ritiksuman/rudra" \
    --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": "ritiksuman/rudra",
		"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 "ritiksuman/rudra" \
        --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": "ritiksuman/rudra",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use ritiksuman/rudra with Ollama:
```
ollama run hf.co/ritiksuman/rudra:BF16
```
Unsloth Desktop

How to use ritiksuman/rudra with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ritiksuman/rudra:BF16

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": "ritiksuman/rudra:BF16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use ritiksuman/rudra with Docker Model Runner:
```
docker model run hf.co/ritiksuman/rudra:BF16
```

Lemonade

How to use ritiksuman/rudra with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ritiksuman/rudra:BF16

Run and chat with the model

lemonade run user.rudra-BF16

List all available models

lemonade list

Hermes Agent

How to use ritiksuman/rudra with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ritiksuman/rudra:BF16

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 ritiksuman/rudra:BF16

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use ritiksuman/rudra with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ritiksuman/rudra:BF16

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 "ritiksuman/rudra:BF16" \
  --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"

🦥 Rudra v1 — 4B Coding Agent, Function-Calling & Reasoning LLM (GGUF + Transformers)

Rudra is a 4B coding-agent / reasoning language model built by Vyomma Intelligence. The brain behind Rudra is Ritik Suman from Vyomma Intelligence.

It is a fine-tune of unsloth/Qwen3.5-4B with QLoRA on Unsloth, trained for coding, agentic tool use, and reasoning, with an anti-hallucination instruction set.

Runs locally on CPU or a small GPU via Ollama / llama.cpp, or with Transformers. Great for a local coding assistant, function-calling experiments, and agent prototypes.

📦 Files in this repo

File	Description
`rudra.Q4_K_M.gguf`	✅ The Rudra model — the best fine-tune (beats base on code). This is the file to use.
`rudra-agent-v1.Q4_K_M.gguf`	Identical copy of `rudra.Q4_K_M.gguf` (kept under the explicit `rudra-agent-v1` name)

Use rudra.Q4_K_M.gguf — it is the Rudra model (the only variant that beats the base model). rudra.Q4_K_M.gguf and rudra-agent-v1.Q4_K_M.gguf are byte-identical (SHA-256 28537c0cd909e339…).

⚡ TL;DR

# 1. Download the GGUF from this repo (rudra.Q4_K_M.gguf)
# 2. Create an Ollama model from it
ollama create rudra -f Modelfile
ollama run rudra "Plan how to fix a failing pytest suite."

🧩 Function calling (JSON <tool_call>)
🛠️ Agentic loop: plan → inspect → edit → run → verify → summarize
🧠 Reasoning with [VERIFY] + CONFIDENCE: markers
🖥️ GGUF Q4_K_M (~2.6 GB) — CPU friendly
🔓 Apache-2.0 (base license)

📊 Honest Gap: Base vs. Fine-tunes

5-example quick tests (HumanEval+ execution + BFCL-style function calling), run on the same harness:

Model	Code pass@1	Function-calling (name-acc)	Verdict
Base `qwen-base`	60% (3/5)	80% (4/5)	Baseline
`rudra` (`rudra.Q4_K_M.gguf`, this repo's main file)	100% (5/5)	80% (4/5)	✅ Beats base on code, ties on FC
`shstra01-mix` (mixed-dataset experimental, no longer on HF)	0% (0/5)	60% (3/5)	❌ Regressed — not recommended
Fresh 2-epoch retrain (`rudra-v1`, LR 2e-4)	20% (1/5)	40% (2/5)	❌ Regressed — not pushed
ORPO variant (`rudra-orpo`)	0% (0/5)	20% (1/5)	❌ Regressed — not pushed

What was rejected and why

Mixed-dataset blend (100 examples from Salesforce/xLAM, Glaive, DeepNLP, NVIDIA Nemotron-Agentic): regressed on both code and function calling, and got stuck in an infinite list_dir thinking loop. Not recommended.
Fresh retrain / ORPO runs: also regressed vs. base under the 6 GB (rank 16, seq 1024) constraint.

Only rudra.Q4_K_M.gguf beats the base model, which is why it is the primary file in this repo.

⚠️ Honest interpretation

Results are from small samples (n=5) — treat as directional, not statistically definitive. Function-calling still depends on the custom JSON tool_calls template, which differs from the base model's native format, so the FC gap vs. proprietary models remains. CONFIDENCE: tags are not fully calibrated.

Recommendation: for production use, rudra.Q4_K_M.gguf is the best available option here; for maximum reliability on code tasks the base Qwen3.5-4B remains competitive.

🚀 Quickstart

Ollama

FROM ./rudra.Q4_K_M.gguf
SYSTEM """Built by vyomma intelligence. Brain behind rudra is ritik suman from vyomma intelligence.
You are Rudra, a coding, agentic and reasoning assistant. Never fabricate facts or outputs; if unsure, say so; end with CONFIDENCE: <high|medium|low>."""
PARAMETER temperature 0.4
PARAMETER num_ctx 8192

ollama create rudra -f Modelfile
ollama run rudra "Write a Python LRU cache."

llama.cpp

llama-cli -m rudra.Q4_K_M.gguf -p "Explain what a list comprehension does."

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer
tok = AutoTokenizer.from_pretrained("ritiksuman/rudra")
model = AutoModelForCausalLM.from_pretrained("ritiksuman/rudra", device_map="auto")
msgs = [{"role": "system", "content": "Built by vyomma intelligence. Brain behind rudra is ritik suman from vyomma intelligence."},
        {"role": "user", "content": "Write a Python function to check for primes."}]
ids = tok.apply_chat_template(msgs, add_generation_prompt=True, return_tensors="pt")
print(tok.decode(model.generate(ids.to(model.device), max_new_tokens=256)[0], skip_special_tokens=True))

🧰 Tools (function calling)

Tool	Purpose
`list_dir(path)`	List files/directories
`read_file(path)`	Read a file
`write_file(path, content)`	Create/overwrite a file
`edit_file(path, old, new)`	Replace an exact substring
`run_command(cmd)`	Run a shell command
`search_code(pattern)`	Search the codebase

Calls are emitted as:

<tool_call>
{"name": "read_file", "arguments": {"path": "calc.py"}}
</tool_call>

📊 Larger Evaluation (earlier run, directional)

Measured with the harness in the repo (HumanEval+ with real execution; BFCL-style function calling on a held-out xLAM slice; a private set including unanswerable prompts). Small samples — treat as directional.

Metric	Base Qwen3.5-4B	Rudra
HumanEval+ pass@1 (n=15)	40%	20%
Function-calling name acc (n=30)	96.7%	56.7%
Answerable accuracy (n=6)	6/6	6/6
Decline rate on unanswerable (n=8)	88%	62%
Overconfident (`CONFIDENCE: high` on unanswerable)	0%	50%

These n=15/n=30 numbers reflect an earlier, weaker fine-tune. The current rudra.Q4_K_M.gguf improves code pass@1 to 100% (n=5) over base's 60%.

🏋️ Training

Setting	Value
Base	`unsloth/Qwen3.5-4B`
Method	QLoRA (4-bit), Unsloth
LoRA rank	16 (alpha 16)
Epochs	2
LR	2e-4
Max seq	1024
Gradient accumulation	4
Hardware	NVIDIA RTX 3050 Laptop (6 GB)
Data	338 coding-agent tool trajectories + CodeAlpaca + GSM8K + curated identity/anti-hallucination (`rudra_agent_train.jsonl`)

🔬 Reproduce a stronger run

A ready-to-run Kaggle T4 notebook is included in the source project (~32k examples, rank 32, LR 2e-4, 4096 tokens, better datasets: Magicoder-OSS-Instruct, OpenCodeInstruct, xLAM, Hermes, Orca-Math). This is the recommended path to a model that actually beats base at scale.

❓ FAQ

Is it free to run? Yes — GGUF Q4_K_M runs on CPU; no API needed. Does it need a GPU? No; a small GPU helps. Context length? Trained at 1024; base supports up to 40,960 — raise num_ctx in Ollama on bigger hardware. License? Apache-2.0 (inherited). Which file should I use? rudra.Q4_K_M.gguf — the only variant that beats base.

🙏 Credits

Fine-tune & tooling by Vyomma Intelligence. Brain behind Rudra: Ritik Suman. Base model: Qwen3.5-4B (Alibaba). Training: Unsloth + TRL.

Downloads last month: 517

Safetensors

Model size

5B params

Tensor type

BF16

F32

Model tree for ritiksuman/rudra

Base model

Qwen/Qwen3.5-4B-Base

Finetuned

Qwen/Qwen3.5-4B

Finetuned

unsloth/Qwen3.5-4B

Adapter

(217)

this model

Adapters

2 models