m64_fbfd8ff7225c / AGENT.md
SenseNova-U1.5-8B-MoT (8-step) text-to-image
This service turns a text prompt into a 2048x2048 (aspect key '1:1') PNG. It runs SenseTime's SenseNova-U1.5-8B-MoT (Apache-2.0), a native unified multimodal model on a Qwen3-8B backbone that generates in pixel space (no VAE), with the official 8-step distillation LoRA merged: 8 steps, CFG 1, with gx10's sensenova_u1_server.py and the vendor's own inference package. Generation settings are
fixed by this capsule (size "1:1", num_inference_steps 8, guidance_scale 1.0, timestep_shift 3.0, cfg_norm "none", think_mode false). You send a prompt and, optionally, a seed.
Base URL: {{BASE_URL}}
Routes
POST /v1/images/generations
OpenAI-style image generation. Returns one image as base64 PNG.
curl -s {{BASE_URL}}/v1/images/generations {{AUTH}} \
-H "Content-Type: application/json" \
-d '{"prompt": "a lighthouse at sunset", "seed": 42}' \
| python3 -c 'import sys,json,base64; d=json.load(sys.stdin); open("out.png","wb").write(base64.b64decode(d["data"][0]["b64_json"])); print(d["params_used"])'
Request fields:
prompt(string, required): what to draw.seed(integer >= 0, optional): the random seed. The same prompt and seed give the same image. Leave it out for a random one.
Also accepted, for stock OpenAI clients: model (any string), n (must be 1),
response_format (must be "b64_json"), user, and size if it is "1:1". Any other
field, or another value for a fixed one, is rejected with 400.
The response is {"created": ..., "data": [{"b64_json": "<PNG, base64>"}], "params_used": {...}}.
The OpenAI Python SDK works too:
from openai import OpenAI
client = OpenAI(base_url="{{BASE_URL}}/v1", api_key="unused")
r = client.images.generate(model="sensenova-u1.5-8b-mot", prompt="a lighthouse at sunset",
response_format="b64_json", extra_body={"seed": 7})
GET /v1/config
The server's settings and what they mean.
GET /status
Where the service is in its start-up: phase is one of building, downloading,
starting, loading, ready, error, with eta_seconds.
Timings and concurrency
- First start: builds a container image and downloads about 36 GB of weights. After that, loading takes about 7.5 minutes (budgeted at 1200 s).
- Each image takes about 18 s (8 steps, 2048x2048) on a DGX Spark (GB10).
- Send requests one at a time, and allow at least 2 minutes per request in your client timeout.
Errors
503withRetry-Afterwhile the service is still starting.400: a field was rejected; the body'serrorsays which and why.422:promptmissing or a field of the wrong type.5xxotherwise: retry once, then checkGET /status.