Instructions to use GestaltLabs/Ornstein-3.5-9B-V2-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use GestaltLabs/Ornstein-3.5-9B-V2-NVFP4 with Transformers:

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

pipe = pipeline("text-generation", model="GestaltLabs/Ornstein-3.5-9B-V2-NVFP4")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("GestaltLabs/Ornstein-3.5-9B-V2-NVFP4")
model = AutoModelForMultimodalLM.from_pretrained("GestaltLabs/Ornstein-3.5-9B-V2-NVFP4")
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

vLLM

How to use GestaltLabs/Ornstein-3.5-9B-V2-NVFP4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "GestaltLabs/Ornstein-3.5-9B-V2-NVFP4"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "GestaltLabs/Ornstein-3.5-9B-V2-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/GestaltLabs/Ornstein-3.5-9B-V2-NVFP4

SGLang

How to use GestaltLabs/Ornstein-3.5-9B-V2-NVFP4 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 "GestaltLabs/Ornstein-3.5-9B-V2-NVFP4" \
    --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": "GestaltLabs/Ornstein-3.5-9B-V2-NVFP4",
		"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 "GestaltLabs/Ornstein-3.5-9B-V2-NVFP4" \
        --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": "GestaltLabs/Ornstein-3.5-9B-V2-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use GestaltLabs/Ornstein-3.5-9B-V2-NVFP4 with Docker Model Runner:
```
docker model run hf.co/GestaltLabs/Ornstein-3.5-9B-V2-NVFP4
```

Ornstein 3.5 9B — V2 · NVFP4

NVFP4 4-bit quantization (Blackwell-optimized) with FP8 KV cache, produced with NVIDIA TensorRT Model Optimizer (PTQ) from GestaltLabs/Ornstein-3.5-9B-V2 — the reinforcement-learning post-training (V2) of Ornstein 3.5 9B. Text weights only; pair with the full base model for the vision tower.

Usage

These ModelOpt checkpoints carry an hf_quant_config.json and are intended for ModelOpt-aware runtimes (TensorRT-LLM / vLLM). Load the directory as a standard HF model id with the matching backend.

Support This Work

I'm a PhD student in visual neuroscience at the University of Toronto who also happens to spend way too much time fine-tuning, merging, and quantizing open-weight models on rented H100s and a local DGX Spark. All training compute is self-funded — balancing GPU costs against a student budget. If my uploads have been useful to you, consider buying a PhD student a coffee. It goes a long way toward keeping these experiments running.

Support on Ko-fi