Instructions to use Code4me2/clef-flash-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Code4me2/clef-flash-NVFP4 with Transformers:

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

pipe = pipeline("image-text-to-text", model="Code4me2/clef-flash-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)

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

processor = AutoProcessor.from_pretrained("Code4me2/clef-flash-NVFP4")
model = AutoModelForMultimodalLM.from_pretrained("Code4me2/clef-flash-NVFP4", 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

vLLM

How to use Code4me2/clef-flash-NVFP4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Code4me2/clef-flash-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": "Code4me2/clef-flash-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/Code4me2/clef-flash-NVFP4

SGLang

How to use Code4me2/clef-flash-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 "Code4me2/clef-flash-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": "Code4me2/clef-flash-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 "Code4me2/clef-flash-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": "Code4me2/clef-flash-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 Code4me2/clef-flash-NVFP4 with Docker Model Runner:
```
docker model run hf.co/Code4me2/clef-flash-NVFP4
```

Code4me2 commited on 7 days ago

Commit

a7076d2

verified ·

1 Parent(s): c25bd83

Card: gate history and decimal sizes

Browse files

Files changed (1) hide show

README.md +7 -2

README.md CHANGED Viewed

@@ -19,7 +19,7 @@ at revision `17f0b0ad`.
 - **Quantized:** the 96 text MLP projections (`gate/up/down_proj`).
 - **Kept in BF16:** the gated-deltanet and full-attention layers, `lm_head` (the joint head
   reads its rows), the vision tower and the joint head.
-- **Size:** 12 GB, versus 19 GB for BF16.
 ## Method
@@ -45,7 +45,7 @@ banking77 (64, 77 options) and LEDGAR (64, 100 options). Inputs run from 1.7K to
 | Top-1 agreement with BF16 | | 0.926 [0.887, 0.952] | 0.883 [0.838, 0.917] |
 | Agreement where BF16 margin ≥ 0.5 (n=156) | | **1.000** | 0.994 |
 | Mean KL(BF16 ‖ quant) | | 0.020 | 0.048 |
-| Size | 19 GB | 12 GB | 8.6 GB |
 Release gates: the accuracy delta's 95% lower bound must be ≥ −2 pt, and agreement on
 confident decisions must be ≥ 0.98. This model passes both. The variant that also
@@ -54,6 +54,11 @@ tokens (−6.1 pt).
 Every disagreement with BF16 is on a question where BF16's own margin is below 0.5.
 Per-band, per-cohort and per-margin breakdowns are in [`metrics/`](metrics/).
 ## Usage

 - **Quantized:** the 96 text MLP projections (`gate/up/down_proj`).
 - **Kept in BF16:** the gated-deltanet and full-attention layers, `lm_head` (the joint head
   reads its rows), the vision tower and the joint head.
+- **Size:** 12.1 GB, versus 19.1 GB for BF16 (decimal GB).
 ## Method
 | Top-1 agreement with BF16 | | 0.926 [0.887, 0.952] | 0.883 [0.838, 0.917] |
 | Agreement where BF16 margin ≥ 0.5 (n=156) | | **1.000** | 0.994 |
 | Mean KL(BF16 ‖ quant) | | 0.020 | 0.048 |
+| Size (GB) | 19.1 | 12.1 | 9.1 |
 Release gates: the accuracy delta's 95% lower bound must be ≥ −2 pt, and agreement on
 confident decisions must be ≥ 0.98. This model passes both. The variant that also
 Every disagreement with BF16 is on a question where BF16's own margin is below 0.5.
+These gates were set after the first full report. Under the originally proposed gates
+(overall agreement ≥ 0.95, every length band ≥ 0.90) this model fails at 0.926 overall
+and 0.878 on the 8K–32K band. Those gates were dropped because band slices of n=26–53
+are too small to gate on.
 Per-band, per-cohort and per-margin breakdowns are in [`metrics/`](metrics/).
 ## Usage