Transformers
Safetensors
decision-model
commerce
system-one
structured-output
calibration
custom-code
qwen3.8
offline
Instructions to use infercrane/Commerce-1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use infercrane/Commerce-1 with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("infercrane/Commerce-1", device_map="auto") - Notebooks
- Google Colab
- Kaggle
|
Download API.md from infercrane/Commerce-1: direct link, hf CLI and curl.
- Browser
- Download file 1.37 kB
-
https://huggingface.co/infercrane/Commerce-1/resolve/main/API.md
- Command line
-
hf download hf://infercrane/Commerce-1/API.md
-
curl -L -o API.md https://huggingface.co/infercrane/Commerce-1/resolve/main/API.md
1.37 kB
Typed local API
The local endpoint is POST /v1/systemone. This example exercises all three
question types:
curl --fail --silent http://127.0.0.1:8000/v1/systemone \
-H 'content-type: application/json' \
--data-binary @- <<'JSON'
{
"model": "infercrane/commerce-1",
"state": {
"cart_value": 149.00,
"customer_tier": "returning",
"destination": "domestic"
},
"questions": {
"route": {
"type": "choice",
"instructions": "Choose the fulfillment route.",
"criteria": {
"standard": "Standard fulfillment",
"review": "Manual review"
}
},
"eligible": {
"type": "noul",
"instructions": "Is the order eligible for automatic processing?"
},
"priority": {
"type": "score",
"instructions": "Score processing priority.",
"criteria": ["low", "medium", "high"]
}
}
}
JSON
Successful responses contain typed answers, raw normalized distributions,
checkpoint and runtime identities, token use, and cost: null. Returned values
depend on the verified local model; this document intentionally includes no
fabricated prediction example.
GET /healthz reports process health. GET /v1/model reports the loaded
content identity. The loopback server serializes model work and may return 429
while busy or 503 when the runtime cannot complete a request.