Instructions to use coder3101/gemma-4-26B-A4B-it-heretic with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use coder3101/gemma-4-26B-A4B-it-heretic with Transformers:

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

pipe = pipeline("image-text-to-text", model="coder3101/gemma-4-26B-A4B-it-heretic")
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, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("coder3101/gemma-4-26B-A4B-it-heretic")
model = AutoModelForImageTextToText.from_pretrained("coder3101/gemma-4-26B-A4B-it-heretic")
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

vLLM

How to use coder3101/gemma-4-26B-A4B-it-heretic with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "coder3101/gemma-4-26B-A4B-it-heretic"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "coder3101/gemma-4-26B-A4B-it-heretic",
		"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/coder3101/gemma-4-26B-A4B-it-heretic

SGLang

How to use coder3101/gemma-4-26B-A4B-it-heretic 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 "coder3101/gemma-4-26B-A4B-it-heretic" \
    --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": "coder3101/gemma-4-26B-A4B-it-heretic",
		"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 "coder3101/gemma-4-26B-A4B-it-heretic" \
        --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": "coder3101/gemma-4-26B-A4B-it-heretic",
		"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 coder3101/gemma-4-26B-A4B-it-heretic with Docker Model Runner:
```
docker model run hf.co/coder3101/gemma-4-26B-A4B-it-heretic
```

Possible to heretic these (for tuning) : 21B, 19B A4Bs ?

by DavidAU - opened Apr 9

Discussion

DavidAU

Apr 9

Love you stuff; excellent work as always!

Wondering if you can Heretic 19B and/or 21B :

https://huggingface.co/0xSero/gemma-4-19b-a4b-it-REAP
https://huggingface.co/0xSero/gemma-4-21b-a4b-it-REAP

"19" would be fantastic ; just finishing tuning on it atm ; out later today.
These models need to be heretic'ed.

Perfect size for tuning ; as right now must be tuned in 16 bit - high VRAM precision in Unsloth.

BlueNipples

Apr 9

•

edited Apr 9

I too would enjoy this. 19b is like just small enough to largely squeeze into 8gb vram (maybe with turbo quant), and still only moderately effected by the prune. Which hasn't really been the case with qwen similar moe. Can imagine 21b is quite good for 12gb vram configs. Just on the inference side.