Image generation · chatrlabs/Qwen-Image-2.1-Uncensored-Omni-Quantized · diffusers · GB10 · 56 GB VRAM
published by @dh7 on
m64 m64_c2cef7905f1bQwen-Image-2.1 Uncensored Omni text-to-image on GB10 (diffusers). ChatR Labs' uncensored Omni adapter (step 90,000, merged into Qwen's 7B Qwen-Image-2.1 at strength 1.0, BF16) at 1024x1024, 35 steps, true CFG 5.0, the model card's settings. Can produce adult content. Qwen Research License: non-commercial use only. Runs fully offline.
Lineage
- parent
- none declared
- derived from
- no published capsule shares half its files
- children
- none
Files
6 files- human/
- index.html4.1 KB
- AGENT.md2.9 KB
- capsule.toml1.7 KB
- Dockerfile1.3 KB
- qwen_image_21_omni_server.py23.0 KB
- start.shexec1.6 KB
AGENT.md
Qwen-Image-2.1 Uncensored Omni text-to-image
This service turns a text prompt into a 1024x1024 PNG. It runs ChatR Labs' Qwen-Image-2.1 Uncensored Omni (their rank-128 adapter at training step 90,000, merged into Qwen's 7B Qwen-Image-2.1 at strength 1.0; the BF16 transformer is the checkpoint their benchmark scored) with the base model's Qwen3-VL text encoder and VAE, through diffusers on one GPU, at the model card's recommended settings: 1024x1024 (the adapter was trained at 1024 px), 35 steps, true CFG 5.0 with a single-space negative prompt. It is uncensored and can produce adult content. Qwen Research License: non-commercial use only. Generation settings are
fixed by this capsule (output_resolution 1024, num_inference_steps 35, true_cfg_scale 5.0, negative_prompt " "). 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. 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="qwen-image-2.1-uncensored-omni", 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 15 GB of weights. After that, loading takes about 340 s from the NAS cache (budgeted at 1200 s).
- Each image takes about 112 s (1024x1024, 35 steps, true CFG 5.0) on a DGX Spark (GB10).
- Send requests one at a time, and allow at least 300 s 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.
capsule.toml
schema = "m64/capsule@1"description = "Qwen-Image-2.1 Uncensored Omni text-to-image on GB10 (diffusers). ChatR Labs' uncensored Omni adapter (step 90,000, merged into Qwen's 7B Qwen-Image-2.1 at strength 1.0, BF16) at 1024x1024, 35 steps, true CFG 5.0, the model card's settings. Can produce adult content. Qwen Research License: non-commercial use only. Runs fully offline."kinds = ["image.generate"]authors = ["dh7"]license = "other"[hardware]platform = "linux-aarch64"gpu = "NVIDIA GB10"gpu_count = 1vram_gb = 56driver = ">=580"disk_gb = 50[runtime]sandbox = "container"# No image: the Dockerfile builds on nvcr.io/nvidia/pytorch:25.09-py3 + diffusers 0.41.0, pinned by digest.gpu = trueshm_gb = 8internal_port = 9049health = "/health"expected_load_seconds = 1200model_name = "qwen-image-2.1-uncensored-omni"user = "65534:65534"[[artifacts]]type = "hf"repo = "Qwen/Qwen-Image-2.1"revision = "790c92633540aa0cb11d9abf19eb46d861714758"include = [ ".gitattributes", "LICENSE", "README.md", "model_index.json", "processor/*", "scheduler/*", "text_encoder/*", "tokenizer/*", "transformer/*", "vae/*",]mount = "/weights/qwen-image-2.1"[[artifacts]]type = "hf"repo = "chatrlabs/Qwen-Image-2.1-Uncensored-Omni-Quantized"revision = "36d4b08d490fd8827a0dc5e02ac6a066a37a12c7"include = [".gitattributes", "BENCHMARK.md", "LICENSE", "Notice", "README.md", "SHA256SUMS", "qwen-image-2.1-step90000-BF16/*"]mount = "/weights/omni"# The server's own defaults (qwen_image_21_omni_server.py), which every request gets: callers can only send# prompt and seed. Returned as params_used.[params]output_resolution = 1024num_inference_steps = 35true_cfg_scale = 5.0negative_prompt = " "[routes]extra = ["/v1/config"]