ayushmaninbox/artificial-stupidity-asif

AS-IF — Artificial Stupidity Image (Full)

⚠️ These are not my weights

This repository contains quantized ONNX exports of models trained by other people. I did not train them and I do not own them.

• sd-turbo/ derives from stabilityai/sd-turbo — © Stability AI, governed by the Stability AI licence and Acceptable Use Policy.
• tiny-sd/ derives from segmind/tiny-sd, itself a pruned Stable Diffusion 1.5.

Check the upstream licences before any use, especially commercial. My contribution is only the export, quantization and decoder swap described below. If Stability AI or Segmind object to this redistribution, contact me and I will remove it.

The counterpart to AS-I, which is mine: 13.7M parameters, 14 MB, trained from scratch on a laptop, and it only draws emoji. AS-IF is the other trade — someone else's billion-parameter model, compressed enough to ship.

	AS-I (mine)	AS-IF sd-turbo	AS-IF tiny-sd
Weights by	me	Stability AI	Segmind
Size	14 MB	1216 MB	454 MB
Steps	8	2	4
Time per image	0.19 s	2.2 s	5.7 s
Resolution	64×64	512×512	512×512
Draws	~1250 emoji	anything	anything

Example input and output

python asif_sample.py --prompt "two astronauts playing chess" --steps 2

Prompt	Result
two astronauts playing chess	both astronauts, a real chessboard
a frog running a startup	an excellent frog. No startup.
a red car beside a blue house	red car and blue house — spatial relation held
a cat riding a bicycle	cat and bicycle, not convincingly joined
an apple on a wooden table	correct, and clean
a yellow bird sitting on a tree	correct, and clean

Four of six are what was asked. "A frog running a startup" produced a frog and dropped the abstraction — the honest failure mode of a 2-step distilled model: concrete nouns survive, conceptual framing does not.

What I actually did

1. Exported to ONNX and quantized per component, because they do not tolerate damage equally:

Component	Precision	fp32	after
UNet	int8	3.2 GB	869 MB
Text encoder	int8	1.3 GB	342 MB
VAE decoder	replaced	189 MB	4.9 MB

2. Replaced the VAE decoder with TAESD rather than quantizing it. Quantizing it to int8 causes visible colour banding on flat regions for ~80 MB saved; replacing it saves 193 MB and runs faster, because SD's decoder is roughly a third of generation time at 512px on CPU.

Trap worth recording: TAESD's scaling_factor is 1.0, so it consumes UNet-space latents directly. Dividing by SD's 0.18215 first — the move every SD decode example shows — hands it values 5.5× too large and returns psychedelic noise that reads as a broken model rather than a broken constant.

3. Measured a smaller base. tiny-sd/ is 2.7× smaller but not step-distilled, so it needs classifier-free guidance — two UNet passes per step. Four of its steps is 8 passes against sd-turbo's 2, hence 2.6× slower.

The build that would win does not exist off the shelf. Pruned and step-distilled — 454 MB at 2 passes — is the obvious combination and cannot be assembled: LCM-LoRA is trained against the full SD 1.5 UNet, so its tensors do not fit a pruned one (lora_A wants [64, 1280, 3, 3], the pruned model has [64, 640, 3, 3]). Pruning and step-distillation do not compose after the fact.

Usage

git clone https://github.com/ayushmaninbox/artificial-stupidity
cd artificial-stupidity/as-image-model
pip install -r requirements.txt

python asif_sample.py --prompt "a wizard riding a motorcycle" --tiny-vae

SD-Turbo requires guidance_scale=0.0 — it is distilled without classifier-free guidance, and the usual 7.5 produces washed-out output. Tiny-SD is a normal SD 1.5 model and wants ~7.5.

You can rebuild both from scratch instead of downloading:

python asif_export.py --tiny-vae            # sd-turbo,  ~1216 MB
python asif_export.py --small --tiny-vae    # tiny-sd,   ~454 MB

Related

artificial-stupidity-image	AS-I — 14 MB, from scratch, mine
artificial-stupidity	AS-F — the language model
artificial-stupidity-tiny	AS-0…AS-5, down to 83 KB

License

Upstream terms govern. See Stability AI for sd-turbo/ and segmind/tiny-sd for tiny-sd/. The export scripts in the GitHub repo are MIT.