Instructions to use hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8 with Transformers:

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

pipe = pipeline("image-text-to-text", model="hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8")
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("hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8")
model = AutoModelForMultimodalLM.from_pretrained("hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8", 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=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8",
		"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/hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8

SGLang

How to use hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8 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 "hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8" \
    --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": "hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8",
		"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 "hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8" \
        --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": "hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8",
		"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 hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8 with Docker Model Runner:
```
docker model run hf.co/hoborific/Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8
```

Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8 / README.md

hoborific

Upload README.md with huggingface_hub

2e8d4c3 verified about 1 month ago

preview code

Raw

History Blame Contribute Delete

1.63 kB

metadata

base_model: Blazed-Forge/Gemma-4-Giftige-Blume-31B-v1
library_name: transformers
tags:
  - quantized
  - w8a16
  - fp8

Gemma-4-Giftige-Blume-31B-v1-W8A16-FP8

Quantized version of Blazed-Forge/Gemma-4-Giftige-Blume-31B-v1.

Format

Offline-quantized W8A16 FP8 in the compressed-tensors float-quantized format: weights in float8_e4m3fn with per-output-channel symmetric scales, activations kept in bf16/fp16.

How it was quantized

For each linear layer, every output row gets its own scale starting from amax / 448, refined by an MSE clip search over ~9 clip fractions (0.8–1.0× amax) picking the lowest-error scale per row. Weights are then quantized q = e4m3(w / scale) with round-to-nearest and saturation. This per-channel + clipping scheme gives better SNR than vLLM's online per-tensor --quantization fp8 path.

Only 2D linear projection weights are quantized (attention q/k/v/o, MLP gate/up/down). Embeddings, norms, lm_head, routers/experts, and the vision tower stay in bf16 and are listed in the checkpoint's ignore list, so vLLM leaves them untouched.

Supported vLLM platforms

Intel XPU — XPUW8A16FP8LinearKernel (the intended target).
NVIDIA CUDA (SM75+, i.e. Turing and newer) — HummingFP8ScaledMMLinearKernel when the humming package is installed, otherwise MarlinFP8ScaledMMLinearKernel.
Not supported: ROCm, CPU, TPU — vLLM has no W8A16-FP8 kernel for these backends yet, so loading will fail with a "no kernel" error.