Instructions to use ornith-ai/Ornith-1.5-9B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ornith-ai/Ornith-1.5-9B with Transformers:

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

pipe = pipeline("text-generation", model="ornith-ai/Ornith-1.5-9B")
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("ornith-ai/Ornith-1.5-9B")
model = AutoModelForMultimodalLM.from_pretrained("ornith-ai/Ornith-1.5-9B", 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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ornith-ai/Ornith-1.5-9B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ornith-ai/Ornith-1.5-9B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ornith-ai/Ornith-1.5-9B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ornith-ai/Ornith-1.5-9B

SGLang

How to use ornith-ai/Ornith-1.5-9B 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 "ornith-ai/Ornith-1.5-9B" \
    --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": "ornith-ai/Ornith-1.5-9B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "ornith-ai/Ornith-1.5-9B" \
        --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": "ornith-ai/Ornith-1.5-9B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ornith-ai/Ornith-1.5-9B with Docker Model Runner:
```
docker model run hf.co/ornith-ai/Ornith-1.5-9B
```

I'm genuinely impressed it is just a better Qwen 9B

#12

by Mk2Oracle - opened 12 days ago

Discussion

Mk2Oracle

12 days ago

I read the posts here, and that hasn't been my experience. And no, I don't need to invalidate anyone else's experience with the model.

I've been using Qwen 9B on my laptop's RTX 5080 for a while, and it is extremely hard for me to even recommend it. It never used tool calls properly, hallucinated, and so on. Using this model, though? IMO, it is in a league of its own. It is just all around better.

Asked it to patch a task that might be complex for these types of models? It investigated, analyzed, and did all the things I would expect at a bare minimum, things that not even Qwen 9B at the base level could do on its own.

It patched a Svelte application and PHP backend properly. I used it for creative writing as well, and I'm not going to say it did well. It quite didn't. I had to go through a lot of turns before we agreed on a tone, format, etc., and then it turned out okay-ish.

And frankly, I've been wanting a model like this. I've used Gemma models, and they all ended in disappointment. Qwen models as well, although Qwen 3.8 is extremely promising.

I used the AtomicChat model here are the flags

things that did really suck:
vision is bad.... or well, mediocre at best.
again, creative writing it needs a lot of polishing
for article writing (non-fiction): if you tell it to generate a op-ed article example then devolves into some fictional story as it goes missing the whole purpose of the first ask. there's some really weird business going on with writing. i had to like do several turns before it understood the assignment.

llama.cpp

AtomicChat--Ornith-1.5-9B-GGUF/Ornith-1.5-9B-AD-Q5_K-Q4_K.gguf -ngl auto -fa on -c 131072 -b 4096 -ub 2048 -ctk q8_0 -ctv q8_0 -n 8192 --perf --reasoning on --reasoning-preserve --reasoning-budget 2048 --jinja --chat-template-kwargs '"'"'{"enable_thinking":true,"preserve_thinking":true}'"'"' --temp 1.0 --top-p 0.95 --top-k 20 --presence-penalty 0.0 --repeat-penalty 1.0 --min-p 0.00 --mmproj /root/.cache/huggingface/hub/models--AtomicChat--Ornith-1.5-9B-GGUF/snapshots/8fc2368e779489ae43607253fc392ec2acae46f3/mmproj-Ornith-1.5-9B-BF16.gguf --image-max-tokens 4096 --image-min-tokens 1024 --host 0.0.0.0 --port 8000 --timeout 7200 --sse-ping-interval 15 --sleep-idle-seconds -1

Ohjaaja

1 day ago

Replying here with same observation. However word of advice. Do not touch lower than Q8 flat default llama.cpp quants. Q4_k_m is absolute waste of time if using it in some other than chat. Go for q8 or for example atleast Quality-36 from here Soulfate24/Ornith-1.5-9B-ASHQ1-Remix-GGUF.

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