Instructions to use primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-FP8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-FP8 with Transformers:

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

pipe = pipeline("image-text-to-text", model="primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-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("primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-FP8")
model = AutoModelForMultimodalLM.from_pretrained("primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-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 primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-FP8 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-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": "primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-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/primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-FP8

SGLang

How to use primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-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 "primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-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": "primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-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 "primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-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": "primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-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 primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-FP8 with Docker Model Runner:
```
docker model run hf.co/primitive-ai/Ornith-1.5-35B-A3B-mixed-NVFP4-FP8
```

MergedColumnParallelLinear has no attribute 'data'

by flaviusburca - opened 25 days ago

Discussion

flaviusburca

25 days ago

fails with vllm 0.27.1

jagat-primitive-org

primitive org 25 days ago

Thanks for the report — could you share a bit more so we can pin it down?

the full traceback (the lines above the AttributeError)
your exact vllm serve command, including any --tensor-parallel-size / --speculative-config flags
GPU model, and your transformers version

For context on what we have already ruled out: on vLLM 0.27.1 this checkpoint loads and generates in four configurations on an RTX PRO 6000 (sm_120) — native FP8/NVFP4, the weight-only Marlin fallback (VLLM_TEST_FORCE_FP8_MARLIN=1, which is the path Ampere-class cards take), MTP speculative decoding, and a bare load with no extra flags. We also audited config.json against the tensors on disk using vLLM's own scheme-matching functions (should_ignore_layer / find_matched_target) across 42,833 modules, with no mismatches.

So this looks environment- or flag-specific rather than a defect in the weights, and the traceback would tell us which. If you are running with tensor parallelism we would especially like to know — that is the one dimension we cannot reproduce on a single-GPU host.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment