Instructions to use daeunj/gpt-oss-120b-modelopt-nvfp4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use daeunj/gpt-oss-120b-modelopt-nvfp4 with Transformers:

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

pipe = pipeline("text-generation", model="daeunj/gpt-oss-120b-modelopt-nvfp4")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("daeunj/gpt-oss-120b-modelopt-nvfp4")
model = AutoModelForCausalLM.from_pretrained("daeunj/gpt-oss-120b-modelopt-nvfp4", 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

vLLM

How to use daeunj/gpt-oss-120b-modelopt-nvfp4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "daeunj/gpt-oss-120b-modelopt-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": "daeunj/gpt-oss-120b-modelopt-nvfp4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/daeunj/gpt-oss-120b-modelopt-nvfp4

SGLang

How to use daeunj/gpt-oss-120b-modelopt-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 "daeunj/gpt-oss-120b-modelopt-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": "daeunj/gpt-oss-120b-modelopt-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 "daeunj/gpt-oss-120b-modelopt-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": "daeunj/gpt-oss-120b-modelopt-nvfp4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use daeunj/gpt-oss-120b-modelopt-nvfp4 with Docker Model Runner:
```
docker model run hf.co/daeunj/gpt-oss-120b-modelopt-nvfp4
```

GPT-OSS-120B modelopt NVFP4 repack

This repository contains an experimental modelopt W4A16_NVFP4 repack of twhitworth/gpt-oss-120b-fp16.

The checkpoint was produced for RTX PRO 6000 / SM120 fallback experiments in the SpecPrefill reproduction workflow. It is intended to be smoke-tested with TensorRT-LLM before benchmark use.

Expected local generation path:

MODEL_ENV_FILE=config/trtllm_modelopt_nvfp4.env \
  bash scripts/trtllm_start_target.sh

MODEL_ENV_FILE=config/trtllm_modelopt_nvfp4.env \
  bash scripts/trtllm_smoke_test_target.sh

Notes:

Source checkpoint: twhitworth/gpt-oss-120b-fp16
Quantization/repack format: modelopt / W4A16_NVFP4
GPT-OSS expert tensor packing is experimental.
Verify license and redistribution requirements from the source model before making this repository public.

Downloads last month: 22

Safetensors

Model size

66B params

Tensor type

I32

BF16

F16

F8_E4M3

F32

Model tree for daeunj/gpt-oss-120b-modelopt-nvfp4

Base model

openai/gpt-oss-120b

Finetuned

twhitworth/gpt-oss-120b-fp16

Quantized

(1)

this model