Instructions to use thinkingmachines/Inkling-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use thinkingmachines/Inkling-NVFP4 with Transformers:

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

pipe = pipeline("image-text-to-text", model="thinkingmachines/Inkling-NVFP4")
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)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("thinkingmachines/Inkling-NVFP4")
model = AutoModelForMultimodalLM.from_pretrained("thinkingmachines/Inkling-NVFP4")
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=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use thinkingmachines/Inkling-NVFP4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "thinkingmachines/Inkling-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": "thinkingmachines/Inkling-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/thinkingmachines/Inkling-NVFP4

SGLang

How to use thinkingmachines/Inkling-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 "thinkingmachines/Inkling-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": "thinkingmachines/Inkling-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "thinkingmachines/Inkling-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": "thinkingmachines/Inkling-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use thinkingmachines/Inkling-NVFP4 with Docker Model Runner:
```
docker model run hf.co/thinkingmachines/Inkling-NVFP4
```

Inkling-NVFP4

File size: 1,110 Bytes

392af0e

{
  "audio_token": "<|unused_200053|>",
  "audio_bos_token": "<|content_audio_input|>",
  "dmel_max_value": 2.0,
  "dmel_min_value": -7.0,
  "feature_extractor": {
    "audio_token_duration_s": 0.05,
    "feature_extractor_type": "InklingFeatureExtractor",
    "feature_size": 80,
    "hop_length": 800,
    "n_fft": 1600,
    "padding_side": "right",
    "padding_value": 0.0,
    "return_attention_mask": true,
    "sampling_rate": 16000,
    "window_size": 1600,
    "window_size_multiplier": 2.0
  },
  "image_processor": {
    "do_convert_rgb": true,
    "do_normalize": true,
    "do_rescale": true,
    "do_resize": true,
    "image_mean": [
      0.48145466,
      0.4578275,
      0.40821073
    ],
    "image_processor_type": "InklingImageProcessor",
    "image_std": [
      0.26862954,
      0.26130258,
      0.27577711
    ],
    "resample": 3,
    "rescale_factor": 0.00392156862745098,
    "size": {
      "height": 40,
      "width": 40
    }
  },
  "image_token": "<|unused_200054|>",
  "image_bos_token": "<|content_image|>",
  "num_dmel_bins": 16,
  "processor_class": "InklingProcessor"
}