Instructions to use bynn-intelligence-com/decisionnode-edge with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use bynn-intelligence-com/decisionnode-edge with Transformers:

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

pipe = pipeline("image-text-to-text", model="bynn-intelligence-com/decisionnode-edge")

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

processor = AutoProcessor.from_pretrained("bynn-intelligence-com/decisionnode-edge")
model = AutoModelForMultimodalLM.from_pretrained("bynn-intelligence-com/decisionnode-edge", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use bynn-intelligence-com/decisionnode-edge with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf bynn-intelligence-com/decisionnode-edge:Q8_0
# Run inference directly in the terminal:
llama cli -hf bynn-intelligence-com/decisionnode-edge:Q8_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf bynn-intelligence-com/decisionnode-edge:Q8_0
# Run inference directly in the terminal:
llama cli -hf bynn-intelligence-com/decisionnode-edge:Q8_0

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf bynn-intelligence-com/decisionnode-edge:Q8_0
# Run inference directly in the terminal:
./llama-cli -hf bynn-intelligence-com/decisionnode-edge:Q8_0

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf bynn-intelligence-com/decisionnode-edge:Q8_0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf bynn-intelligence-com/decisionnode-edge:Q8_0

Use Docker

docker model run hf.co/bynn-intelligence-com/decisionnode-edge:Q8_0

Llama new
LM Studio
Jan

vLLM

How to use bynn-intelligence-com/decisionnode-edge with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "bynn-intelligence-com/decisionnode-edge"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "bynn-intelligence-com/decisionnode-edge",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/bynn-intelligence-com/decisionnode-edge:Q8_0

SGLang

How to use bynn-intelligence-com/decisionnode-edge 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 "bynn-intelligence-com/decisionnode-edge" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "bynn-intelligence-com/decisionnode-edge",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "bynn-intelligence-com/decisionnode-edge" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "bynn-intelligence-com/decisionnode-edge",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Ollama
How to use bynn-intelligence-com/decisionnode-edge with Ollama:
```
ollama run hf.co/bynn-intelligence-com/decisionnode-edge:Q8_0
```
Unsloth Desktop
Docker Model Runner
How to use bynn-intelligence-com/decisionnode-edge with Docker Model Runner:
```
docker model run hf.co/bynn-intelligence-com/decisionnode-edge:Q8_0
```

Lemonade

How to use bynn-intelligence-com/decisionnode-edge with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull bynn-intelligence-com/decisionnode-edge:Q8_0

Run and chat with the model

lemonade run user.decisionnode-edge-Q8_0

List all available models

lemonade list

Atomic Chat

bynn-intelligence commited on 5 days ago

Commit

0beea7e

verified ·

1 Parent(s): c55e78b

Model card: point and box questions on the hosted API

Browse files

Files changed (2) hide show

MANIFEST.sha256 +1 -1
README.md +12 -8

MANIFEST.sha256 CHANGED Viewed

@@ -2,7 +2,7 @@
 50cbab8a892c5f2993b8c7351a99182507472def3b1374558308605d99b86b32  ./LICENSE-base
 2014f13a6fe1d9178c84cde47bb087a7a74adbe38650f1823a944077081e86d0  ./LICENSING.md
 e4dcac398e39a098e07f625ed239949296a2000d95dcc275cb0d8083ebb51dc9  ./NOTICE
-a85c3b1edbf52508e03c489415f1484b00b74325d0fbddeb92c3d6f6127ee4d9  ./README.md
 c44489690e63adb3abd8c65ede675fadc5714126e3f866cbd292df824880cb77  ./assets/decisionnode-banner.png
 5febf5c3cb5843034f2251d959c475a4a629dec65bf3ea494b4fed0a3b8aa3aa  ./calibration.json
 28b44a18f60b7e3039c503971b6d194b51c723f0c8a62d56bd6ce7f7f34e9f87  ./config.json

 50cbab8a892c5f2993b8c7351a99182507472def3b1374558308605d99b86b32  ./LICENSE-base
 2014f13a6fe1d9178c84cde47bb087a7a74adbe38650f1823a944077081e86d0  ./LICENSING.md
 e4dcac398e39a098e07f625ed239949296a2000d95dcc275cb0d8083ebb51dc9  ./NOTICE
+c9e9cc16cb86d7ed4f47cd1e66413f60eadd5f1858c921cbf5877c66902f82c4  ./README.md
 c44489690e63adb3abd8c65ede675fadc5714126e3f866cbd292df824880cb77  ./assets/decisionnode-banner.png
 5febf5c3cb5843034f2251d959c475a4a629dec65bf3ea494b4fed0a3b8aa3aa  ./calibration.json
 28b44a18f60b7e3039c503971b6d194b51c723f0c8a62d56bd6ce7f7f34e9f87  ./config.json

README.md CHANGED Viewed

@@ -110,7 +110,7 @@ def decide(request, threshold=0.7):
     return answers  # still below the threshold: take the safe action
 ```
-Pass images as base64 `data` with a `media_type` (not a local `path`), so the same request is valid on the device and on the hosted API. Only the questions that are still unsure are sent, and a decision never changes when the same request is sent again, so each question costs at most one call per hosted model.
 ## Model details
@@ -121,7 +121,7 @@ Pass images as base64 `data` with a `media_type` (not a local `path`), so the sa
 | Parameters | 2.21 billion (1.88 billion in the language model, 0.33 billion in the vision encoder) |
 | Base model | [Qwen3.5-2B-Base](https://huggingface.co/Qwen/Qwen3.5-2B-Base) |
 | Input | Text, and up to 16 images per request (JPEG, PNG or WebP) |
-| Output | Typed decisions: `choice`, `score`, `truth` and `number` |
 | Context | Evaluated on inputs of up to 8,192 tokens; the runtime accepts up to 16,384 |
 | Weights | bfloat16 safetensors, plus an 8-bit GGUF build (2.0 GB) |
 | License | AGPL-3.0, or a commercial license from Bynn Intelligence |
@@ -192,12 +192,16 @@ The runtime also runs from the command line (`python -m decisionnode_edge <model
 ## Question types
-| Type | You provide | You receive |
-|---|---|---|
-| `choice` | 1 to 255 options, each with an optional description | the selected option, a confidence and a probability for every option |
-| `score` | 1 to 10 ordered levels, from low to high | the expected level, a confidence and a probability for every level |
-| `truth` | a statement or question, with optional `true` and `false` criteria | the probability that the answer is yes |
-| `number` | `min`, `max` and `step` (up to 4,096 values) | the most probable value, the expected value, a confidence and the probable values |
 Write what the question means into `instructions` and `criteria`; the question's key is never shown to the model. Images are seen by every question in the request.

     return answers  # still below the threshold: take the safe action
 ```
+Pass images as base64 `data` with a `media_type` (not a local `path`), so the same request is valid on the device and on the hosted API. Send point and box questions to the hosted API directly. Only the questions that are still unsure are sent, and a decision never changes when the same request is sent again, so each question costs at most one call per hosted model.
 ## Model details
 | Parameters | 2.21 billion (1.88 billion in the language model, 0.33 billion in the vision encoder) |
 | Base model | [Qwen3.5-2B-Base](https://huggingface.co/Qwen/Qwen3.5-2B-Base) |
 | Input | Text, and up to 16 images per request (JPEG, PNG or WebP) |
+| Output | Typed decisions: `choice`, `score`, `truth` and `number`; `point` and `box` through the hosted API |
 | Context | Evaluated on inputs of up to 8,192 tokens; the runtime accepts up to 16,384 |
 | Weights | bfloat16 safetensors, plus an 8-bit GGUF build (2.0 GB) |
 | License | AGPL-3.0, or a commercial license from Bynn Intelligence |
 ## Question types
+| Type | You provide | You receive | Runs on |
+|---|---|---|---|
+| `choice` | 1 to 255 options, each with an optional description | the selected option, a confidence and a probability for every option | DecisionNode Edge and the hosted API |
+| `score` | 1 to 10 ordered levels, from low to high | the expected level, a confidence and a probability for every level | DecisionNode Edge and the hosted API |
+| `truth` | a statement or question, with optional `true` and `false` criteria | the probability that the answer is yes | DecisionNode Edge and the hosted API |
+| `number` | `min`, `max` and `step` (up to 4,096 values) | the most probable value, the expected value, a confidence and the probable values | DecisionNode Edge and the hosted API |
+| `point` | an image and what to point to | a point (`x`, `y`) in pixels or fractions of the image, the probability that the thing is there, and a confidence | the hosted API: DecisionNode Flash or DecisionNode |
+| `box` | an image and what to box | a bounding box (`x1`, `y1`, `x2`, `y2`), the probability that the thing is there, and a confidence | the hosted API: DecisionNode Flash or DecisionNode |
+Point and box questions are answered today by the hosted models through the [DecisionNode API](https://decisionnode.com/docs/api), with the same request format; send them there directly.
 Write what the question means into `instructions` and `criteria`; the question's key is never shown to the model. Images are seen by every question in the request.