Make model card visual, clear, and easier to explore

README.md

@@ -5,6 +5,11 @@ inference: false
 base_model: nota-ai/bk-sdm-tiny-2m
 license: creativeml-openrail-m
 tags:
+  - local-ai
+  - image-generation
+  - diffusers
+  - lora
+  - apple-silicon
   - clover-image
   - text-to-image
   - diffusion
@@ -57,545 +62,122 @@ widget:
       url: examples/prompt-gallery/original/image_81.png
 ---
 
-# Clover Image Tiny 🍀🤗
 
-![Clover Image Tiny mosaic banner](assets/clover-image-tiny-banner.png)
+# 🍀 Clover Image Tiny
+
+### Small model. Room for big ideas.
+
+**Turn a few words into an image. Explore a different style. Make a focused edit.**
+Clover Image Tiny is a compact, open-weight AI image generator for local text-to-image
+workflows on Mac, Windows, and Linux, with Core ML releases for Apple devices.
+Download the model once, then create offline on your own hardware.
 
 <p align="center">
-  <a href="https://huggingface.co/neonforestmist/Clover-Image-Tiny"><img src="https://img.shields.io/badge/Hugging%20Face-Model-FFD21E?logo=huggingface&logoColor=black" alt="Hugging Face model" /></a>
-  <a href="https://github.com/neonforestmist/Clover-Image-Tiny"><img src="https://img.shields.io/badge/GitHub-Source-181717?logo=github&logoColor=white" alt="GitHub source" /></a>
-  <a href="https://huggingface.co/neonforestmist/Clover-Image-Tiny-Inpaint"><img src="https://img.shields.io/badge/Hugging%20Face-Inpainting-FFB000?logo=huggingface&logoColor=black" alt="Clover Image Tiny Inpaint model" /></a>
-  <a href="https://huggingface.co/spaces/neonforestmist/Clover-Image-Tiny-Demo"><img src="https://img.shields.io/badge/Hugging%20Face-Live-FFD21E?logo=huggingface&logoColor=black" alt="Live Hugging Face demo" /></a>
-  <a href="https://huggingface.co/spaces/neonforestmist/Clover-Image-Tiny-Demo"><img src="https://img.shields.io/badge/ZeroGPU-Demo-20BEFF?logo=huggingface&logoColor=white" alt="ZeroGPU demo" /></a>
-  <a href="https://github.com/neonforestmist/Clover-Image-Tiny-iOS"><img src="https://img.shields.io/badge/iPhone-Core%20ML-111111?logo=apple&logoColor=white" alt="iPhone Core ML" /></a>
-  <a href="https://github.com/neonforestmist/clover-image-tiny-lora-trainer"><img src="https://img.shields.io/badge/Visual%20Trainer-LoRA-7C3AED?logo=python&logoColor=white" alt="LoRA visual trainer" /></a>
-  <a href="https://github.com/neonforestmist/Clover-Image-Tiny/actions"><img src="https://img.shields.io/badge/Quality%20checks-passing-22C55E?logo=github&logoColor=white" alt="Quality checks passing" /></a>
-  <a href="LICENSE"><img src="https://img.shields.io/badge/License-CreativeML%20Open%20RAIL--M-D32F2F" alt="CreativeML Open RAIL-M license" /></a>
+  <a href="https://huggingface.co/spaces/neonforestmist/Clover-Image-Tiny-Demo"><img alt="Try the demo" src="https://huggingface.co/neonforestmist/Clover-Image-Tiny/resolve/main/assets/links/try-demo.svg" height="40"></a>
+  <a href="https://huggingface.co/neonforestmist/Clover-Image-Tiny"><img alt="Get the model" src="https://huggingface.co/neonforestmist/Clover-Image-Tiny/resolve/main/assets/links/get-model.svg" height="40"></a>
+  <a href="https://huggingface.co/neonforestmist/Clover-Image-Tiny-Inpaint"><img alt="Inpainting" src="https://huggingface.co/neonforestmist/Clover-Image-Tiny/resolve/main/assets/links/inpainting.svg" height="40"></a>
+  <a href="https://github.com/neonforestmist/Clover-Image-Tiny"><img alt="GitHub source" src="https://huggingface.co/neonforestmist/Clover-Image-Tiny/resolve/main/assets/links/github-source.svg" height="40"></a>
 </p>
 
-Clover Image Tiny 🤗 is a compact 512×512 diffusion model built for
-**low-latency, consumer-grade generation**: local creative tools, edge
-applications, and private on-device workflows such as phone inference through
-Core ML. It starts from BK-SDM-Tiny's proven compressed Stable Diffusion v1.4
-U-Net and gives it a second, fully recorded distillation pass against the full
-frozen SD v1.4 teacher. Clover preserves the tiny 323.4M-parameter denoiser and
-standard Diffusers compatibility while adding its own refreshed weights,
-creator-selected visual character, reproducibility record, adapters,
-inpainting, and local/on-device product stack.
-
-The goal is straightforward: bring useful text-to-image generation closer to
-the consumer device instead of assuming a large cloud model for every request.
-That local-first design is built with even lower end-to-end product latency in
-mind—especially where on-device execution can remove a cloud round trip.
-Results vary with the prompt, seed, scheduler, and number of denoising steps.
-
-For masked image editing, use the dedicated
-[Clover Image Tiny Inpaint model →](https://huggingface.co/neonforestmist/Clover-Image-Tiny-Inpaint).
-It is a separate 9-channel adaptation with its own recommended pipeline and
-Core ML resources.
-
-[**Try Clover Image Tiny & Inpainting in the demo space →**](https://huggingface.co/spaces/neonforestmist/Clover-Image-Tiny-Demo)
-
-[**Native Core ML implementation →**](https://github.com/neonforestmist/Clover-Image-Tiny-iOS)
-
-The demo exposes prompt, negative prompt, seed, guidance, dimensions,
-scheduler, and 4–100 conventional Diffusers inference steps. It creates one
-image per request and keeps the packaged safety checker enabled.
-
-## Contents
-
-1. [Overview](#1-overview)
-2. [Examples](#2-examples)
-3. [Small-model benchmark](#3-small-model-benchmark)
-4. [iPhone and Core ML](#4-iphone-and-core-ml)
-5. [Run locally](#5-run-locally)
-6. [Generation controls](#6-generation-controls)
-7. [Hardware and operating systems](#7-hardware-and-operating-systems)
-8. [Python API](#8-python-api)
-9. [About this release](#9-about-this-release)
-10. [Quality and known behavior](#10-quality-and-known-behavior)
-11. [Safety](#11-safety)
-12. [Training lineage and data](#12-training-lineage-and-data)
-13. [Citation](#13-citation)
-14. [Licenses](#14-licenses)
-15. [Reproducibility and artifact identity](#15-reproducibility-and-artifact-identity)
-
-## 1. Overview
-
-### 1.1 Model profile
-
-| Profile | Value |
-|---|---:|
-| Published PyTorch/Diffusers package | **~1.67 GB** |
-| Denoiser / U-Net | **323,384,964 parameters** · ~646.8 MB |
-| Text encoder | ~246.1 MB |
-| VAE | ~167.3 MB |
-| Safety checker | ~608.0 MB |
-| Native output | 512×512 |
-| Inference range | 4–100 steps |
-| Published formats | PyTorch/Diffusers + Core ML adapters |
-
-The package figure includes the text encoder, VAE, tokenizer/configuration, and
-the enabled upstream safety checker; it is therefore larger than the denoiser
-alone. Component sizes are approximate runtime-footprint context, not a claim
-that every byte is active in every backend.
-
-### 1.2 What Clover adds to BK-SDM-Tiny-2M
-
-Clover is a downstream, materially modified checkpoint of
-[`nota-ai/bk-sdm-tiny-2m`](https://huggingface.co/nota-ai/bk-sdm-tiny-2m),
-not a new or further-compressed architecture. It retains the same block-removed
-Stable Diffusion 1.x U-Net topology and parameter count. The tokenizer, text
-encoder, and VAE are byte-identical to the pinned upstream components; the
-meaningful model-level change is the additionally distilled Clover U-Net
-weights.
-
-| Aspect | BK-SDM-Tiny-2M | Clover Image Tiny |
-|---|---|---|
-| Training lineage | Its model card reports 50,000 distillation-pretraining iterations on 2,256,472 LAION-Aesthetics pairs | Inherits the pinned `aad3e0e8` checkpoint, then adds a separately audited 500-step Clover distillation run |
-| Teacher and objectives | SD v1.4 compression lineage with output and feature KD | Re-distilled against the exact pinned SD v1.4 teacher using equal-weight diffusion-target, teacher-output, and six-point normalized feature losses |
-| Clover calibration data | — | Deterministically selected 1,000-pair PD3M set; license-filtered to CC0/Public Domain Mark items with a checksummed deletion audit and manifest |
-| Trainable scope | Upstream release | U-Net only; teacher, text encoder, and VAE remained frozen |
-| U-Net architecture | Block-removed U-Net, ~323.4M parameters | Same topology and ~323.4M parameters; weights are modified |
-| Product target | Efficient compressed Stable Diffusion baseline | Low-latency consumer products: desktop-local generation, edge deployment, and Core ML phone inference |
-| Runtime envelope | Original compact baseline | Same parameter count and effectively the same measured latency/memory class—no denoiser-size penalty for the Clover refresh |
-| Release additions | Upstream BK-SDM model | Reproducible checkpoint identity, local runner, benchmark evidence, Core ML resources, style LoRAs, inpainting companion, and iOS integration |
-
-That makes Clover a **quality-refreshed, product-oriented evolution in the same
-compact runtime envelope**, rather than merely a rename or a larger model. Its
-advantage is the complete path from a small conventional checkpoint to
-consumer-facing use: reproducible PyTorch inference, Core ML conversion,
-native iPhone integration, style adapters, inpainting, and a hosted demo. The
-exact checkpoint was selected after paired visual evaluation because its output
-character was preferred for Clover's gallery and product direction.
-
-The small shared-runtime benchmark below reinforces the efficiency story:
-Clover averaged `1.024 s` versus `1.027 s` for BK-SDM-Tiny-2M on the same A10G
-recipe, while peak allocation differed by only about 3 MB. Clover scored higher
-on the CLIP prompt proxy for 6 of the 16 fixed prompts. BK-SDM-Tiny-2M retained
-the higher overall mean (`0.3246` versus `0.3195`), so this small benchmark is
-best read as evidence that Clover preserves the base model's efficiency and
-competitive small-model behavior while delivering a distinct checkpoint and a
-broader consumer-ready ecosystem—not as a universal ranking of image quality.
-
-## 2. Examples
-
-### 2.1 Prompt gallery
-
-These nine examples are paired with the exact prompts used for generation. The
-portable prompt list is also available at
-[`examples/prompt-gallery/original/prompts.txt`](examples/prompt-gallery/original/prompts.txt).
-
-<Gallery />
-
-### 2.2 Base model → style LoRA
-
-The same tiny-glass-greenhouse prompt shows the base output alongside the three
-style adapters. These are visual references for the corresponding LoRA cards.
-
-| [Base model](https://huggingface.co/neonforestmist/Clover-Image-Tiny) | [Monet](https://huggingface.co/neonforestmist/clover-image-tiny-monet-lora) | [Pointillism](https://huggingface.co/neonforestmist/clover-image-tiny-pointillism-lora) | [Watercolor Anime](https://huggingface.co/neonforestmist/clover-image-tiny-watercolor-anime-lora) |
-|---|---|---|---|
-| [![Base model greenhouse](examples/normal-to-lora/normal.png)](https://huggingface.co/neonforestmist/Clover-Image-Tiny) | [![Monet greenhouse](examples/normal-to-lora/monet.png)](https://huggingface.co/neonforestmist/clover-image-tiny-monet-lora) | [![Pointillism greenhouse](examples/normal-to-lora/pointillism.png)](https://huggingface.co/neonforestmist/clover-image-tiny-pointillism-lora) | [![Watercolor Anime greenhouse](examples/normal-to-lora/watercolor_anime.png)](https://huggingface.co/neonforestmist/clover-image-tiny-watercolor-anime-lora) |
-
-## 3. Small-model benchmark
-
-Clover is compared with its pinned BK-SDM-Tiny-2M base and two public
-same-family references using 16 prompts, identical seeds, 512×512 output, 30
-DDIM steps, guidance 7.5, and a shared NVIDIA A10G runtime. The measurement is
-an engineering comparison, not a human-preference leaderboard.
-
-| Model | U-Net parameters | Loaded pipeline parameters | Mean latency | Peak CUDA | Mean CLIP cosine |
-|---|---:|---:|---:|---:|---:|
-| [Clover Image Tiny](https://huggingface.co/neonforestmist/Clover-Image-Tiny) | 323.4M | 834.1M | 1.024 s | 2,233 MB | 0.3195 |
-| [BK-SDM-Tiny-2M](https://huggingface.co/nota-ai/bk-sdm-tiny-2m) | 323.4M | 834.1M | 1.027 s | 2,230 MB | 0.3246 |
-| [Segmind Tiny-SD](https://huggingface.co/segmind/tiny-sd) | 323.4M | 530.1M | 1.028 s | 1,649 MB | 0.3345 |
-| [BK-SDM-v2-Tiny](https://huggingface.co/nota-ai/bk-sdm-v2-tiny) | 326.8M | 750.9M | 0.957 s | 2,067 MB | 0.3303 |
-
-The table keeps denoiser size and loaded pipeline size separate. The former is
-the most useful apples-to-apples model comparison; the latter includes the
-text encoder, VAE, and other loaded components and is runtime context rather
-than a download-size metric.
-
-CLIP cosine is only a prompt-adherence proxy. It is not a human-quality score,
-FID, safety evaluation, or evidence that these models are interchangeable.
-The complete protocol, machine-readable results, and generated examples are in
-[`benchmarks/text-to-image/`](benchmarks/text-to-image/) and the
-[full benchmark report](benchmarks/text-to-image/results/clover-small-model-comparison-20260825/REPORT.md).
-
-![Four-prompt small-model comparison](benchmarks/text-to-image/results/clover-small-model-comparison-20260825/contact-sheet.png)
-
-## 4. iPhone and Core ML
-
-The companion [Clover Image Tiny iOS project](https://github.com/neonforestmist/Clover-Image-Tiny-iOS)
-follows Apple platform conventions and exposes prompt, negative prompt, steps,
-guidance, seed, image count, scheduler, random generator, and compute target.
-Its model picker downloads SHA-256-verified, immutable Core ML resources from
-Hugging Face; the large weights are not committed to the Xcode project or
-bundled into the app.
-
-The base catalog stores the shared text encoder, VAE decoder, safety checker,
-and tokenizer once:
-
-- [`neonforestmist/Clover-Image-Tiny-CoreML`](https://huggingface.co/neonforestmist/Clover-Image-Tiny-CoreML)
-
-Each trained style is available both as a standard Diffusers LoRA and as its
-own Core ML picker download:
-
-| Style | Diffusers LoRA | Core ML LoRA model |
-|---|---|---|
-| Monet | [`clover-image-tiny-monet-lora`](https://huggingface.co/neonforestmist/clover-image-tiny-monet-lora) | [`clover-image-tiny-monet-lora-coreml`](https://huggingface.co/neonforestmist/clover-image-tiny-monet-lora-coreml) |
-| Pointillism | [`clover-image-tiny-pointillism-lora`](https://huggingface.co/neonforestmist/clover-image-tiny-pointillism-lora) | [`clover-image-tiny-pointillism-lora-coreml`](https://huggingface.co/neonforestmist/clover-image-tiny-pointillism-lora-coreml) |
-| Watercolor Anime | [`clover-image-tiny-watercolor-anime-lora`](https://huggingface.co/neonforestmist/clover-image-tiny-watercolor-anime-lora) | [`clover-image-tiny-watercolor-anime-lora-coreml`](https://huggingface.co/neonforestmist/clover-image-tiny-watercolor-anime-lora-coreml) |
-
-See [`COREML.md`](COREML.md) for conversion details and
-[`training/README.md`](training/README.md) for the pinned LoRA jobs.
-
-### 4.1 Inpainting track
-
-The 9-channel SD 1.4-class inpainting adaptation is trained and packaged separately:
-[`neonforestmist/Clover-Image-Tiny-Inpaint`](https://huggingface.co/neonforestmist/Clover-Image-Tiny-Inpaint).
-The companion Core ML resource bundle is published at
-[`neonforestmist/Clover-Image-Tiny-Inpaint-CoreML`](https://huggingface.co/neonforestmist/Clover-Image-Tiny-Inpaint-CoreML).
-Its Modal trainer, Core ML export, and native runtime integration are in
-[`inpainting/`](inpainting/), [`modal_inpaint.py`](modal_inpaint.py), and
-[`training/README-INPAINTING.md`](training/README-INPAINTING.md).
-
-The context-aware v2 release distills a pinned SD inpainting teacher with
-diverse free-form and object-like masks. It improved held-out masked MAE by
-19.3% over v1. Recommended settings are DPM-Solver++, 20 steps, CFG 6.0, and a
-96-pixel mask-context crop; the runtime composites through the exact mask so
-unmasked pixels remain unchanged.
-A worked `add blue sunglasses` edit is shown in the
-[inpainting model card](https://huggingface.co/neonforestmist/Clover-Image-Tiny-Inpaint#inpainting-example).
-
-In Diffusers, Clover's published attention-only style LoRAs can also be loaded
-on the inpainting pipeline because their target layers retain the same shapes.
-Adapters that modify the four-channel input convolution are incompatible with
-the nine-channel U-Net. Dynamic LoRA loading is a separate deployment feature:
-the currently shipped stateless Core ML inpainting path does not expose it.
-
-## 5. Run locally
-
-Download once, then generate offline with the bundled runner. Python 3.11 and
-3.12 are supported.
-
-### 5.1 macOS — Apple silicon
-
-~~~bash
-mkdir clover-image-tiny-local
-cd clover-image-tiny-local
-
-python3.12 -m venv .venv
-source .venv/bin/activate
-python -m pip install --upgrade pip
-python -m pip install "huggingface-hub==0.36.2"
-
-hf download "neonforestmist/Clover-Image-Tiny" --local-dir model
-python -m pip install -r model/requirements.txt
-
-python model/examples/generate.py \
-  --model model \
-  --device mps \
-  --local-files-only \
-  --prompt "a tiny glass greenhouse glowing in a moonlit garden, detailed photography" \
-  --negative-prompt "blurry, distorted, low detail" \
-  --steps 50 \
-  --guidance-scale 7.5 \
-  --scheduler pndm \
-  --seed 1337 \
-  --output clover-image-tiny.png
-
-open clover-image-tiny.png
-~~~
-
-Use `python3.11` instead if that is the installed supported Python.
-
-### 5.2 Windows — PowerShell
-
-~~~powershell
-mkdir clover-image-tiny-local
-cd clover-image-tiny-local
-
-py -3.12 -m venv .venv
-.venv\Scripts\Activate.ps1
-python -m pip install --upgrade pip
-python -m pip install "huggingface-hub==0.36.2"
-
-hf download "neonforestmist/Clover-Image-Tiny" --local-dir model
-python -m pip install -r model\requirements.txt
-
-python model\examples\generate.py `
-  --model model `
-  --device auto `
-  --local-files-only `
-  --prompt "a tiny glass greenhouse glowing in a moonlit garden, detailed photography" `
-  --negative-prompt "blurry, distorted, low detail" `
-  --steps 50 `
-  --guidance-scale 7.5 `
-  --scheduler pndm `
-  --seed 1337 `
-  --output clover-image-tiny.png
-
-Invoke-Item .\clover-image-tiny.png
-~~~
-
-Use `py -3.11` if needed. With `--device auto`, the runner selects an
-available NVIDIA CUDA GPU and otherwise uses CPU.
-
-### 5.3 Linux
-
-~~~bash
-mkdir clover-image-tiny-local
-cd clover-image-tiny-local
-
-python3.12 -m venv .venv
-source .venv/bin/activate
-python -m pip install --upgrade pip
-python -m pip install "huggingface-hub==0.36.2"
-
-hf download "neonforestmist/Clover-Image-Tiny" --local-dir model
-python -m pip install -r model/requirements.txt
-
-python model/examples/generate.py \
-  --model model \
-  --device auto \
-  --local-files-only \
-  --prompt "a tiny glass greenhouse glowing in a moonlit garden, detailed photography" \
-  --negative-prompt "blurry, distorted, low detail" \
-  --steps 50 \
-  --seed 1337 \
-  --output clover-image-tiny.png
-~~~
-
-`--device auto` selects CUDA when PyTorch can see an NVIDIA GPU and otherwise
-uses CPU. After the first download, `--local-files-only` prevents network
-access during generation.
-
-## 6. Generation controls
-
-The command above is ready to copy. Change these flags to explore the model:
-
-| Flag | Accepted values | Default | What it controls |
-|---|---|---|---|
-| `--prompt` | Non-empty text | Required | What to generate |
-| `--negative-prompt` | Text, or empty | Empty | Details to discourage; the starter commands and live demo use `blurry, distorted, low detail` |
-| `--steps` | 4–100 | `50` | Diffusion iterations; more steps take longer and do not guarantee a better image |
-| `--guidance-scale` | 0.0–20.0 | `7.5` | How strongly the image follows the prompt |
-| `--scheduler` | `pndm`, `ddim`, `euler`, `euler-a`, `dpmpp-2m` | `pndm` | Sampling method |
-| `--width` | 256–768, divisible by 64 | `512` | Output width |
-| `--height` | 256–768, divisible by 64 | `512` | Output height |
-| `--num-images` | 1–4 | `1` | Images generated in one run |
-| `--seed` | 0–(2⁶³−1) | `1337` | Repeatable starting seed |
-| `--device` | `auto`, `cuda`, `mps`, `cpu` | `auto` | Compute backend |
-| `--local-files-only` | Flag | Off | Require an already-downloaded local model |
-
-The reference configuration is 50-step PNDM, guidance 7.5, 512×512, one
-image, seed 1337, and an empty negative prompt. The live demo pre-fills
-`blurry, distorted, low detail`; the local runner leaves the field empty unless
-you pass the flag.
-
-For multiple images, the first uses the requested filename and later images use
-numbered names such as `clover-image-tiny-02.png`. Seeds advance from the
-requested seed. A JSON sidecar beside the first PNG records every resolved
-setting, output filename, seed, checksum, and safety result. Existing planned
-outputs are never overwritten.
-
-Run `python model/examples/generate.py --help` for the complete CLI reference.
-
-## 7. Hardware and operating systems
-
-| System | Automatic backend | Precision | Current evidence |
-|---|---|---|---|
-| iPhone (iOS 17+) | Core ML | mixed/compiled | GitHub project and chunked download path linked above |
-| Apple-silicon Mac | MPS | fp16 | Measured locally on an M4 Pro |
-| Windows/Linux with NVIDIA | CUDA | fp16 | Supported code path; performance not measured |
-| CPU-only macOS/Windows/Linux | CPU | fp32 | Supported code path; performance not measured |
-| Windows AMD/DirectML | — | — | No packaged DirectML path |
-
-Keep at least 2 GB free for the model alone and additional room for the Python
-environment and caches; no formal total-install minimum has been measured.
-Larger images and batches need more memory; lower `--width`, `--height`, or
-`--num-images` if necessary.
-
-The measured Mac reference used a 24 GB Apple M4 Pro and completed one 512×512
-image in 18.21 seconds with fp16 MPS. Its process-lifetime maximum RSS was
-631,341,056 bytes. This is a measured point, not a minimum-RAM claim. No Core
-ML package is required for the Python path.
-
-## 8. Python API
-
-~~~python
-import torch
-from diffusers import DiffusionPipeline, PNDMScheduler
-
-model_id = "neonforestmist/Clover-Image-Tiny"
-if torch.cuda.is_available():
-    device = "cuda"
-elif torch.backends.mps.is_available():
-    device = "mps"
-else:
-    device = "cpu"
-
-dtype = torch.float16 if device in {"cuda", "mps"} else torch.float32
-pipe = DiffusionPipeline.from_pretrained(model_id, torch_dtype=dtype)
-pipe.scheduler = PNDMScheduler.from_config(pipe.scheduler.config)
-pipe = pipe.to(device)
-
-generator_device = "cuda" if device == "cuda" else "cpu"
-generator = torch.Generator(device=generator_device).manual_seed(1337)
-image = pipe(
-    prompt="a tiny greenhouse glowing in a moonlit garden",
-    negative_prompt="blurry, distorted, low detail",
-    num_inference_steps=50,
-    guidance_scale=7.5,
-    height=512,
-    width=512,
-    generator=generator,
-).images[0]
-image.save("clover-image-tiny.png")
-~~~
-
-Seeded generation is repeatable within the selected runtime. Different
-devices, dtypes, kernels, and dependency builds can produce different pixels.
-
-## 9. About this release
-
-Clover Image Tiny is a conventional knowledge-distillation checkpoint trained
-for 500 optimizer steps on an exact licensed 1,000-pair calibration set. This
-was a real U-Net optimization run—not a repackaging operation. Its final cursor
-records 4,000 microsteps and 4,000 sample presentations, with finite training
-rows and nonzero gradients throughout.
-
-The model was initialized from
-`nota-ai/bk-sdm-tiny-2m@aad3e0e8ba61b7cb9f64869dc4e586f8ad9d3665`
-and distilled with a frozen
-`CompVis/stable-diffusion-v1-4@133a221b8aa7292a167afc5127cb63fb5005638b`
-teacher. It is a genuinely modified checkpoint, but it was not trained from
-random initialization.
-
-### 9.1 Clover distillation recipe
-
-| Training field | Recorded value |
+## Why create with Clover?
+
+- **Keep your ideas local.** Run generation on your own computer after downloading the weights.
+- **Start small.** The regular model has a 323.4M-parameter denoiser and generates at a native 512 × 512 resolution.
+- **Find your look.** Explore Monet, Pointillism, and Watercolor Anime with optional style adapters.
+- **Make it part of your workflow.** Use a visual demo, Python, a command-line runner, or the native Core ML app.
+
+Clover is a good fit for visual brainstorming, illustration experiments, and developers
+building local creative tools. You control the prompt, seed, style, and generation settings.
+
+## Create from a prompt
+
+Describe the scene you want to explore. The regular Clover model can turn simple prompts
+into landscapes, still lifes, and stylized artwork. These are existing model examples,
+with the original prompts shown below.
+
+| Moonlit greenhouse | Blue flowers | Stained-glass night |
+|:---:|:---:|:---:|
+| ![Clover text-to-image output: a tiny greenhouse in a moonlit garden](https://huggingface.co/neonforestmist/Clover-Image-Tiny/resolve/3f2a698bc3cbad617970a73d623e0732bb5f87f5/examples/prompt-gallery/original/image_74.png) | ![Clover text-to-image output: a bouquet of blue flowers](https://huggingface.co/neonforestmist/Clover-Image-Tiny/resolve/3f2a698bc3cbad617970a73d623e0732bb5f87f5/examples/prompt-gallery/original/image_79.png) | ![Clover text-to-image output: a stained-glass starry night](https://huggingface.co/neonforestmist/Clover-Image-Tiny/resolve/3f2a698bc3cbad617970a73d623e0732bb5f87f5/examples/prompt-gallery/original/image_78.png) |
+| “a tiny glass greenhouse glowing in a moonlit garden” | “A bouquet of blue flowers” | “A stain glass window of a starry night” |
+
+## Change a detail with inpainting
+
+Keep the image you started with and describe an edit to a selected area. Inpainting means
+painting a mask over the part you want to regenerate—white marks the edit, black marks what
+to keep. The published example below uses the prompt **“add blue sunglasses.”**
+
+| Before | After the masked edit |
+|:---:|:---:|
+| ![Original cat artwork before the masked sunglasses edit](https://huggingface.co/neonforestmist/Clover-Image-Tiny-Inpaint/resolve/3e22009fb6e61944f28b0389f775fc47e2c48724/examples/sunglasses-source.png) | ![Clover inpainting example: the cat with blue sunglasses added](https://huggingface.co/neonforestmist/Clover-Image-Tiny-Inpaint/resolve/3e22009fb6e61944f28b0389f775fc47e2c48724/examples/sunglasses-result.png) |
+
+[**Explore Clover Inpaint HQ →**](https://huggingface.co/neonforestmist/Clover-Image-Tiny-Inpaint) · [**Try generation and inpainting →**](https://huggingface.co/spaces/neonforestmist/Clover-Image-Tiny-Demo)
+
+Inpaint HQ pairs the full Stable Diffusion 1.5 inpainting denoiser with Clover's shared
+components. It is a larger, separate model focused on editing quality. For exact preservation,
+composite the result through the original binary mask.
+
+## One prompt, different styles
+
+Style adapters—also called LoRAs—let you change the visual character of your images.
+Here is the same greenhouse prompt with the base model and three published Clover styles.
+
+| Clover | Monet | Pointillism | Watercolor Anime |
+|:---:|:---:|:---:|:---:|
+| ![Base Clover greenhouse example](https://huggingface.co/neonforestmist/Clover-Image-Tiny/resolve/3f2a698bc3cbad617970a73d623e0732bb5f87f5/examples/normal-to-lora/normal.png) | ![Monet style greenhouse with painterly brushwork](https://huggingface.co/neonforestmist/Clover-Image-Tiny/resolve/3f2a698bc3cbad617970a73d623e0732bb5f87f5/examples/normal-to-lora/monet.png) | ![Pointillism style greenhouse made from colored dots](https://huggingface.co/neonforestmist/Clover-Image-Tiny/resolve/3f2a698bc3cbad617970a73d623e0732bb5f87f5/examples/normal-to-lora/pointillism.png) | ![Watercolor Anime style greenhouse illustration](https://huggingface.co/neonforestmist/Clover-Image-Tiny/resolve/3f2a698bc3cbad617970a73d623e0732bb5f87f5/examples/normal-to-lora/watercolor_anime.png) |
+| [Base model](https://huggingface.co/neonforestmist/Clover-Image-Tiny) | [Get Monet](https://huggingface.co/neonforestmist/clover-image-tiny-monet-lora) | [Get Pointillism](https://huggingface.co/neonforestmist/clover-image-tiny-pointillism-lora) | [Get Watercolor Anime](https://huggingface.co/neonforestmist/clover-image-tiny-watercolor-anime-lora) |
+
+## Choose your way to create
+
+| Your workflow | Start here |
 |---|---|
-| Student initialization | `nota-ai/bk-sdm-tiny-2m@aad3e0e8ba61b7cb9f64869dc4e586f8ad9d3665` |
-| Frozen teacher | `CompVis/stable-diffusion-v1-4@133a221b8aa7292a167afc5127cb63fb5005638b` |
-| Trainable parameters | Clover U-Net only; text encoder, VAE, and teacher frozen |
-| Resolution | 512×512 |
-| Optimization | 500 AdamW steps · effective batch 8 · learning rate `1e-5` |
-| Precision | bfloat16 autocast with float32 master weights |
-| Objective | `1.0 × diffusion + 1.0 × teacher output + 1.0 × normalized feature KD` |
-| Feature transfer | Six source-audited BK-Tiny ↔ SD v1.4 internal feature mappings |
-| Reproducibility | Seed 1337 · atomic checkpoints every 50 steps · exact resume proven at step 100 |
-| Training hardware | One NVIDIA A100-SXM4-80GB |
-
-Each objective contributed something complementary: the diffusion term retained
-the standard epsilon-prediction task, output KD pulled the compact student
-toward the full teacher's denoising prediction, and feature KD aligned internal
-representations at six explicitly mapped points across the down, attention, and
-up paths. Teacher execution used `no_grad`; no teacher, CLIP text-encoder, or
-VAE gradients were accumulated. This gives Clover a targeted weight refresh
-without increasing its U-Net parameter count or abandoning standard Stable
-Diffusion/Diffusers compatibility.
-
-This repository contains the PyTorch/Diffusers checkpoint. Core ML artifacts,
-style adapters, and the companion iOS project are versioned separately and
-linked above.
-
-## 10. Quality and known behavior
-
-- The included gallery demonstrates recognizable subjects across colorful
-  scenes, products, food, an animal, a landscape, and an interior.
-- Individual results vary by prompt, seed, scheduler, and step count. More
-  steps increase runtime but do not guarantee a better result.
-- Hands, anatomy, exact counts and relationships, and readable text can be
-  difficult.
-- The small-model comparison is an engineering benchmark with a CLIP
-  prompt-adherence proxy, not a controlled human-preference study.
-- Resolution and batch size multiply memory use.
-
-## 11. Safety
-
-The upstream safety checker is packaged and enabled in both the supported
-runner and hosted demo. A flagged output may be returned as a black placeholder;
-the JSON sidecar records `nsfw_content_detected` so the result is not silent.
-The checker is useful but not a complete moderation system and can miss harmful
-content or over-filter benign content.
-
-Applications should add controls appropriate to their audience and review
-outputs before sharing them. Do not use the model for consequential decisions,
-identity claims, medical or legal conclusions, harassment, exploitation,
-illegal activity, or uses prohibited by CreativeML OpenRAIL-M.
-
-## 12. Training lineage and data
-
-- Clover fine-tuning data: exactly 1,000 accepted image-caption pairs from
-  `Spawning/PD3M@2a5eb24a8dccf245acd8e56341761aee06da0bdf`
-- Split: 973 train, 17 validation, and 10 test records
-- Data gate: `CDLA-Permissive-2.0`; accepted items retain CC0-1.0 or Public
-  Domain Mark 1.0 provenance
-- Preprocessing: deterministic center crop and 512×512 JPEG conversion,
-  version `clover-pd3m-center-crop-512-jpeg95-v1`
-- Dataset-manifest SHA-256:
-  `50c1249f1cb0d8d690a9acc451ca10c9432eb5a7f4e26f34acb5462096e72322`
-
-The set was chosen by a deterministic hash ordering from the pinned PD3M
-revision, then validated for license, dimensions, MIME type, source
-organization, payload integrity, and deletion-list status. The resulting shard,
-manifest, rejection log, selection statistics, and preprocessing recipe were
-all checksummed. This is a deliberately small calibration pass layered on top
-of BK-SDM-Tiny-2M's much larger inherited pretraining—not a claim that Clover
-learned general image generation from only 1,000 examples.
-
-The 1,000 records describe the Clover fine-tuning run. The student and teacher
-already contain knowledge from larger upstream corpora. Their pinned model
-cards and weight licenses are disclosed, while complete item-level provenance
-for all foundational pretraining is not available to this project.
-
-See `DATA_PROVENANCE.md` for the portable manifest identity and
-`MODEL_DATA_LICENSES.md` for the complete component ledger.
-
-## 13. Citation
-
-If Clover Image Tiny is useful in your work, please cite the model release:
-
-```bibtex
-@software{lozadaperez2026cloverimagetiny,
-  author = {Lukas Lozada Perez},
-  title = {Clover Image Tiny: Compact Local Text-to-Image Diffusion},
-  year = {2026},
-  url = {https://huggingface.co/neonforestmist/Clover-Image-Tiny}
-}
-```
-
-## 14. Licenses
-
-The model weights are a derivative under **CreativeML OpenRAIL-M**. The example
-runner and packaging code are under **Apache-2.0**. Dataset and item-level terms
-remain separate. Read `LICENSE`, `LICENSE-MODEL-CREATIVEML-OPENRAIL-M.txt`,
-`LICENSE-CODE`, and `MODEL_DATA_LICENSES.md` before redistribution or use.
-
-The hero mosaic is user-supplied presentation artwork included by explicit
-request for display in this public model repository. It is not benchmark
-evidence, its panel-generation provenance is not claimed, and this package
-does not grant a downstream reuse license for it.
-
-## 15. Reproducibility and artifact identity
-
-| Field | Value |
+| Try it in your browser | [Hosted demo: generation and inpainting](https://huggingface.co/spaces/neonforestmist/Clover-Image-Tiny-Demo) |
+| Generate locally with a visual interface | [GitHub setup and Gradio app](https://github.com/neonforestmist/Clover-Image-Tiny#run-clover-locally) |
+| Build with Python and Diffusers | [Model weights](https://huggingface.co/neonforestmist/Clover-Image-Tiny) · [Local runner](https://github.com/neonforestmist/Clover-Image-Tiny/blob/main/generate.py) |
+| Create on iPhone or iPad | [Native Clover app](https://github.com/neonforestmist/Clover-Image-Tiny-iOS) · [Core ML resources](https://huggingface.co/neonforestmist/Clover-Image-Tiny-CoreML) |
+| Train a personal style | [Visual LoRA trainer](https://github.com/neonforestmist/clover-image-tiny-lora-trainer) |
+
+The hosted demo runs remotely. Local Python and Core ML workflows run on your hardware
+after setup; model downloads require a network connection.
+
+## Small enough to build around
+
+| Regular Clover model | What it means |
 |---|---|
-| Repository | `neonforestmist/Clover-Image-Tiny` |
-| Release status | **PUBLIC PYTORCH/DIFFUSERS CHECKPOINT RELEASE** |
-| Training experiment | `clover-kd-20260712T050925Z-01KXABNHP0` |
-| Optimizer step | 500 |
-| Checkpoint SHA-256 | `4a5b99ff18478742528a0d31c97dcee939b166a51be858721d40ad5984110893` |
-| Checkpoint-bundle SHA-256 | `384b6515f5f26838aea33ec9a941e06610a20764f0b8637c8b7b0667bfc0d447` |
-| Resolved-config SHA-256 | `80cf9395d1f587dc0c1d440d9f5b55c55c20703187998509bb306d19d463f597` |
-| Denoiser parameters | `323,384,964` |
-| Package bytes | `1676086612` |
-| Package files | `31` |
-| Validated Stage B source-package checksums SHA-256 | `d9a28d5fe6f5b675ee1b9db52e6d0493c8d3d357bb824eac590911acbd5c3ebc` |
-| Builder source commit | `9f5ce495fcb88238ec7fdc33204fa42ec9690c37` |
-
-`checksums.json` covers every file in the immutable validated release package
-at the recorded builder commit. Later model-card-only revisions are additionally
-preserved by the Hugging Face Git history.
+| **323.4M denoiser parameters** | A compact Stable Diffusion 1.4-class image model |
+| **About 1.67 GB of model files** | Includes the text encoder, VAE, and packaged safety checker; allow extra space for dependencies and caches |
+| **512 × 512 native output** | A practical starting point for visual experiments |
+| **Diffusers + separate Core ML exports** | Python integration and a path to on-device Apple apps |
+
+On one NVIDIA A10G benchmark, Clover averaged **1.024 seconds per image** across
+16 prompts at 512 × 512, 30 DDIM steps, and guidance 7.5. That is a specific measured
+GPU result, not an iPhone timing or a speed guarantee. [See the comparison and protocol](https://huggingface.co/neonforestmist/Clover-Image-Tiny/blob/main/benchmarks/text-to-image/results/clover-small-model-comparison-20260825/REPORT.md).
+
+## A few things to know
+
+**Can I use Clover offline?** Yes. Download the model and dependencies first, then use
+`--local-files-only` with the local runner. No hosted generation service is required.
+
+**Does it work on a Mac?** The Python path supports Apple silicon through PyTorch MPS.
+NVIDIA systems use CUDA; CPU inference is also supported, but slower.
+
+**Is normal Clover the same as Inpaint HQ?** They are separate checkpoints. Normal Clover
+creates images from text. Inpaint HQ takes an image, a mask, and a prompt, and uses a larger denoiser.
+
+**What are its limits?** Hands, faces, readable text, precise counts, and complex relationships
+can be unreliable. Examples are selected outputs; results vary with prompts and settings.
+The packaged safety checker in the Python runner and hosted demo is imperfect.
+
+## Model lineage, license, and research details
+
+Clover starts from [BK-SDM-Tiny-2M](https://huggingface.co/nota-ai/bk-sdm-tiny-2m)
+and adds a recorded 500-step distillation pass against Stable Diffusion 1.4.
+It retains the compact architecture; its U-Net weights are refreshed.
+
+Model weights use **CreativeML Open RAIL-M** and the example code uses **Apache-2.0**.
+See the [component license ledger](https://huggingface.co/neonforestmist/Clover-Image-Tiny/blob/main/MODEL_DATA_LICENSES.md) for the full terms and inherited obligations.
+
+[**Technical reference: setup, benchmarks, training, and citation**](https://huggingface.co/neonforestmist/Clover-Image-Tiny/blob/main/docs/MODEL_DETAILS.md) ·
+[Data provenance](https://huggingface.co/neonforestmist/Clover-Image-Tiny/blob/main/DATA_PROVENANCE.md) ·
+[All nine prompt examples](https://huggingface.co/neonforestmist/Clover-Image-Tiny/tree/main/examples/prompt-gallery/original)
+
+Created by **Lukas Lozada Perez**.

docs/MODEL_DETAILS.md

@@ -0,0 +1,357 @@
+# Clover Image Tiny — technical reference
+
+Detailed setup, evaluation, training provenance, and release identity for the regular model.
+[Back to the visual model card](https://huggingface.co/neonforestmist/Clover-Image-Tiny).
+
+## 3. Small-model benchmark
+
+Clover is compared with its pinned BK-SDM-Tiny-2M base and two public
+same-family references using 16 prompts, identical seeds, 512×512 output, 30
+DDIM steps, guidance 7.5, and a shared NVIDIA A10G runtime. The measurement is
+an engineering comparison, not a human-preference leaderboard.
+
+| Model | U-Net parameters | Loaded pipeline parameters | Mean latency | Peak CUDA | Mean CLIP cosine |
+|---|---:|---:|---:|---:|---:|
+| [Clover Image Tiny](https://huggingface.co/neonforestmist/Clover-Image-Tiny) | 323.4M | 834.1M | 1.024 s | 2,233 MB | 0.3195 |
+| [BK-SDM-Tiny-2M](https://huggingface.co/nota-ai/bk-sdm-tiny-2m) | 323.4M | 834.1M | 1.027 s | 2,230 MB | 0.3246 |
+| [Segmind Tiny-SD](https://huggingface.co/segmind/tiny-sd) | 323.4M | 530.1M | 1.028 s | 1,649 MB | 0.3345 |
+| [BK-SDM-v2-Tiny](https://huggingface.co/nota-ai/bk-sdm-v2-tiny) | 326.8M | 750.9M | 0.957 s | 2,067 MB | 0.3303 |
+
+The table keeps denoiser size and loaded pipeline size separate. The former is
+the most useful apples-to-apples model comparison; the latter includes the
+text encoder, VAE, and other loaded components and is runtime context rather
+than a download-size metric.
+
+CLIP cosine is only a prompt-adherence proxy. It is not a human-quality score,
+FID, safety evaluation, or evidence that these models are interchangeable.
+The complete protocol, machine-readable results, and generated examples are in
+[`benchmarks/text-to-image/`](https://huggingface.co/neonforestmist/Clover-Image-Tiny/blob/main/benchmarks/text-to-image/) and the
+[full benchmark report](https://huggingface.co/neonforestmist/Clover-Image-Tiny/blob/main/benchmarks/text-to-image/results/clover-small-model-comparison-20260825/REPORT.md).
+
+![Four-prompt small-model comparison](https://huggingface.co/neonforestmist/Clover-Image-Tiny/blob/main/benchmarks/text-to-image/results/clover-small-model-comparison-20260825/contact-sheet.png)
+
+## 5. Run locally
+
+Download once, then generate offline with the bundled runner. Python 3.11 and
+3.12 are supported.
+
+### 5.1 macOS — Apple silicon
+
+~~~bash
+mkdir clover-image-tiny-local
+cd clover-image-tiny-local
+
+python3.12 -m venv .venv
+source .venv/bin/activate
+python -m pip install --upgrade pip
+python -m pip install "huggingface-hub==0.36.2"
+
+hf download "neonforestmist/Clover-Image-Tiny" --local-dir model
+python -m pip install -r model/requirements.txt
+
+python model/examples/generate.py \
+  --model model \
+  --device mps \
+  --local-files-only \
+  --prompt "a tiny glass greenhouse glowing in a moonlit garden, detailed photography" \
+  --negative-prompt "blurry, distorted, low detail" \
+  --steps 50 \
+  --guidance-scale 7.5 \
+  --scheduler pndm \
+  --seed 1337 \
+  --output clover-image-tiny.png
+
+open clover-image-tiny.png
+~~~
+
+Use `python3.11` instead if that is the installed supported Python.
+
+### 5.2 Windows — PowerShell
+
+~~~powershell
+mkdir clover-image-tiny-local
+cd clover-image-tiny-local
+
+py -3.12 -m venv .venv
+.venv\Scripts\Activate.ps1
+python -m pip install --upgrade pip
+python -m pip install "huggingface-hub==0.36.2"
+
+hf download "neonforestmist/Clover-Image-Tiny" --local-dir model
+python -m pip install -r model\requirements.txt
+
+python model\examples\generate.py `
+  --model model `
+  --device auto `
+  --local-files-only `
+  --prompt "a tiny glass greenhouse glowing in a moonlit garden, detailed photography" `
+  --negative-prompt "blurry, distorted, low detail" `
+  --steps 50 `
+  --guidance-scale 7.5 `
+  --scheduler pndm `
+  --seed 1337 `
+  --output clover-image-tiny.png
+
+Invoke-Item .\clover-image-tiny.png
+~~~
+
+Use `py -3.11` if needed. With `--device auto`, the runner selects an
+available NVIDIA CUDA GPU and otherwise uses CPU.
+
+### 5.3 Linux
+
+~~~bash
+mkdir clover-image-tiny-local
+cd clover-image-tiny-local
+
+python3.12 -m venv .venv
+source .venv/bin/activate
+python -m pip install --upgrade pip
+python -m pip install "huggingface-hub==0.36.2"
+
+hf download "neonforestmist/Clover-Image-Tiny" --local-dir model
+python -m pip install -r model/requirements.txt
+
+python model/examples/generate.py \
+  --model model \
+  --device auto \
+  --local-files-only \
+  --prompt "a tiny glass greenhouse glowing in a moonlit garden, detailed photography" \
+  --negative-prompt "blurry, distorted, low detail" \
+  --steps 50 \
+  --seed 1337 \
+  --output clover-image-tiny.png
+~~~
+
+`--device auto` selects CUDA when PyTorch can see an NVIDIA GPU and otherwise
+uses CPU. After the first download, `--local-files-only` prevents network
+access during generation.
+
+## 6. Generation controls
+
+The command above is ready to copy. Change these flags to explore the model:
+
+| Flag | Accepted values | Default | What it controls |
+|---|---|---|---|
+| `--prompt` | Non-empty text | Required | What to generate |
+| `--negative-prompt` | Text, or empty | Empty | Details to discourage; the starter commands and live demo use `blurry, distorted, low detail` |
+| `--steps` | 4–100 | `50` | Diffusion iterations; more steps take longer and do not guarantee a better image |
+| `--guidance-scale` | 0.0–20.0 | `7.5` | How strongly the image follows the prompt |
+| `--scheduler` | `pndm`, `ddim`, `euler`, `euler-a`, `dpmpp-2m` | `pndm` | Sampling method |
+| `--width` | 256–768, divisible by 64 | `512` | Output width |
+| `--height` | 256–768, divisible by 64 | `512` | Output height |
+| `--num-images` | 1–4 | `1` | Images generated in one run |
+| `--seed` | 0–(2⁶³−1) | `1337` | Repeatable starting seed |
+| `--device` | `auto`, `cuda`, `mps`, `cpu` | `auto` | Compute backend |
+| `--local-files-only` | Flag | Off | Require an already-downloaded local model |
+
+The reference configuration is 50-step PNDM, guidance 7.5, 512×512, one
+image, seed 1337, and an empty negative prompt. The live demo pre-fills
+`blurry, distorted, low detail`; the local runner leaves the field empty unless
+you pass the flag.
+
+For multiple images, the first uses the requested filename and later images use
+numbered names such as `clover-image-tiny-02.png`. Seeds advance from the
+requested seed. A JSON sidecar beside the first PNG records every resolved
+setting, output filename, seed, checksum, and safety result. Existing planned
+outputs are never overwritten.
+
+Run `python model/examples/generate.py --help` for the complete CLI reference.
+
+## 7. Hardware and operating systems
+
+| System | Automatic backend | Precision | Current evidence |
+|---|---|---|---|
+| iPhone (see current app requirements) | Core ML | mixed/compiled | GitHub project and chunked download path linked above |
+| Apple-silicon Mac | MPS | fp16 | Measured locally on an M4 Pro |
+| Windows/Linux with NVIDIA | CUDA | fp16 | Supported code path; performance not measured |
+| CPU-only macOS/Windows/Linux | CPU | fp32 | Supported code path; performance not measured |
+| Windows AMD/DirectML | — | — | No packaged DirectML path |
+
+Keep at least 2 GB free for the model alone and additional room for the Python
+environment and caches; no formal total-install minimum has been measured.
+Larger images and batches need more memory; lower `--width`, `--height`, or
+`--num-images` if necessary.
+
+The measured Mac reference used a 24 GB Apple M4 Pro and completed one 512×512
+image in 18.21 seconds with fp16 MPS. Its process-lifetime maximum RSS was
+631,341,056 bytes. This is a measured point, not a minimum-RAM claim. No Core
+ML package is required for the Python path.
+
+## 8. Python API
+
+~~~python
+import torch
+from diffusers import DiffusionPipeline, PNDMScheduler
+
+model_id = "neonforestmist/Clover-Image-Tiny"
+if torch.cuda.is_available():
+    device = "cuda"
+elif torch.backends.mps.is_available():
+    device = "mps"
+else:
+    device = "cpu"
+
+dtype = torch.float16 if device in {"cuda", "mps"} else torch.float32
+pipe = DiffusionPipeline.from_pretrained(model_id, torch_dtype=dtype)
+pipe.scheduler = PNDMScheduler.from_config(pipe.scheduler.config)
+pipe = pipe.to(device)
+
+generator_device = "cuda" if device == "cuda" else "cpu"
+generator = torch.Generator(device=generator_device).manual_seed(1337)
+image = pipe(
+    prompt="a tiny greenhouse glowing in a moonlit garden",
+    negative_prompt="blurry, distorted, low detail",
+    num_inference_steps=50,
+    guidance_scale=7.5,
+    height=512,
+    width=512,
+    generator=generator,
+).images[0]
+image.save("clover-image-tiny.png")
+~~~
+
+Seeded generation is repeatable within the selected runtime. Different
+devices, dtypes, kernels, and dependency builds can produce different pixels.
+
+## 9. About this release
+
+Clover Image Tiny is a conventional knowledge-distillation checkpoint trained
+for 500 optimizer steps on an exact licensed 1,000-pair calibration set. This
+was a real U-Net optimization run—not a repackaging operation. Its final cursor
+records 4,000 microsteps and 4,000 sample presentations, with finite training
+rows and nonzero gradients throughout.
+
+The model was initialized from
+`nota-ai/bk-sdm-tiny-2m@aad3e0e8ba61b7cb9f64869dc4e586f8ad9d3665`
+and distilled with a frozen
+`CompVis/stable-diffusion-v1-4@133a221b8aa7292a167afc5127cb63fb5005638b`
+teacher. It is a genuinely modified checkpoint, but it was not trained from
+random initialization.
+
+### 9.1 Clover distillation recipe
+
+| Training field | Recorded value |
+|---|---|
+| Student initialization | `nota-ai/bk-sdm-tiny-2m@aad3e0e8ba61b7cb9f64869dc4e586f8ad9d3665` |
+| Frozen teacher | `CompVis/stable-diffusion-v1-4@133a221b8aa7292a167afc5127cb63fb5005638b` |
+| Trainable parameters | Clover U-Net only; text encoder, VAE, and teacher frozen |
+| Resolution | 512×512 |
+| Optimization | 500 AdamW steps · effective batch 8 · learning rate `1e-5` |
+| Precision | bfloat16 autocast with float32 master weights |
+| Objective | `1.0 × diffusion + 1.0 × teacher output + 1.0 × normalized feature KD` |
+| Feature transfer | Six source-audited BK-Tiny ↔ SD v1.4 internal feature mappings |
+| Reproducibility | Seed 1337 · atomic checkpoints every 50 steps · exact resume proven at step 100 |
+| Training hardware | One NVIDIA A100-SXM4-80GB |
+
+Each objective contributed something complementary: the diffusion term retained
+the standard epsilon-prediction task, output KD pulled the compact student
+toward the full teacher's denoising prediction, and feature KD aligned internal
+representations at six explicitly mapped points across the down, attention, and
+up paths. Teacher execution used `no_grad`; no teacher, CLIP text-encoder, or
+VAE gradients were accumulated. This gives Clover a targeted weight refresh
+without increasing its U-Net parameter count or abandoning standard Stable
+Diffusion/Diffusers compatibility.
+
+This repository contains the PyTorch/Diffusers checkpoint. Core ML artifacts,
+style adapters, and the companion iOS project are versioned separately and
+linked above.
+
+## 10. Quality and known behavior
+
+- The included gallery demonstrates recognizable subjects across colorful
+  scenes, products, food, an animal, a landscape, and an interior.
+- Individual results vary by prompt, seed, scheduler, and step count. More
+  steps increase runtime but do not guarantee a better result.
+- Hands, anatomy, exact counts and relationships, and readable text can be
+  difficult.
+- The small-model comparison is an engineering benchmark with a CLIP
+  prompt-adherence proxy, not a controlled human-preference study.
+- Resolution and batch size multiply memory use.
+
+## 11. Safety
+
+The upstream safety checker is packaged and enabled in both the supported
+runner and hosted demo. A flagged output may be returned as a black placeholder;
+the JSON sidecar records `nsfw_content_detected` so the result is not silent.
+The checker is useful but not a complete moderation system and can miss harmful
+content or over-filter benign content.
+
+Applications should add controls appropriate to their audience and review
+outputs before sharing them. Do not use the model for consequential decisions,
+identity claims, medical or legal conclusions, harassment, exploitation,
+illegal activity, or uses prohibited by CreativeML OpenRAIL-M.
+
+## 12. Training lineage and data
+
+- Clover fine-tuning data: exactly 1,000 accepted image-caption pairs from
+  `Spawning/PD3M@2a5eb24a8dccf245acd8e56341761aee06da0bdf`
+- Split: 973 train, 17 validation, and 10 test records
+- Data gate: `CDLA-Permissive-2.0`; accepted items retain CC0-1.0 or Public
+  Domain Mark 1.0 provenance
+- Preprocessing: deterministic center crop and 512×512 JPEG conversion,
+  version `clover-pd3m-center-crop-512-jpeg95-v1`
+- Dataset-manifest SHA-256:
+  `50c1249f1cb0d8d690a9acc451ca10c9432eb5a7f4e26f34acb5462096e72322`
+
+The set was chosen by a deterministic hash ordering from the pinned PD3M
+revision, then validated for license, dimensions, MIME type, source
+organization, payload integrity, and deletion-list status. The resulting shard,
+manifest, rejection log, selection statistics, and preprocessing recipe were
+all checksummed. This is a deliberately small calibration pass layered on top
+of BK-SDM-Tiny-2M's much larger inherited pretraining—not a claim that Clover
+learned general image generation from only 1,000 examples.
+
+The 1,000 records describe the Clover fine-tuning run. The student and teacher
+already contain knowledge from larger upstream corpora. Their pinned model
+cards and weight licenses are disclosed, while complete item-level provenance
+for all foundational pretraining is not available to this project.
+
+See `DATA_PROVENANCE.md` for the portable manifest identity and
+`MODEL_DATA_LICENSES.md` for the complete component ledger.
+
+## 13. Citation
+
+If Clover Image Tiny is useful in your work, please cite the model release:
+
+```bibtex
+@software{lozadaperez2026cloverimagetiny,
+  author = {Lukas Lozada Perez},
+  title = {Clover Image Tiny: Compact Local Text-to-Image Diffusion},
+  year = {2026},
+  url = {https://huggingface.co/neonforestmist/Clover-Image-Tiny}
+}
+```
+
+## 14. Licenses
+
+The model weights are a derivative under **CreativeML OpenRAIL-M**. The example
+runner and packaging code are under **Apache-2.0**. Dataset and item-level terms
+remain separate. Read `LICENSE`, `LICENSE-MODEL-CREATIVEML-OPENRAIL-M.txt`,
+`LICENSE-CODE`, and `MODEL_DATA_LICENSES.md` before redistribution or use.
+
+The legacy hero mosaic is user-supplied presentation artwork included by explicit
+request for display in this public model repository. It is not benchmark
+evidence, its panel-generation provenance is not claimed, and this package
+does not grant a downstream reuse license for it.
+
+## 15. Reproducibility and artifact identity
+
+| Field | Value |
+|---|---|
+| Repository | `neonforestmist/Clover-Image-Tiny` |
+| Release status | **PUBLIC PYTORCH/DIFFUSERS CHECKPOINT RELEASE** |
+| Training experiment | `clover-kd-20260712T050925Z-01KXABNHP0` |
+| Optimizer step | 500 |
+| Checkpoint SHA-256 | `4a5b99ff18478742528a0d31c97dcee939b166a51be858721d40ad5984110893` |
+| Checkpoint-bundle SHA-256 | `384b6515f5f26838aea33ec9a941e06610a20764f0b8637c8b7b0667bfc0d447` |
+| Resolved-config SHA-256 | `80cf9395d1f587dc0c1d440d9f5b55c55c20703187998509bb306d19d463f597` |
+| Denoiser parameters | `323,384,964` |
+| Package bytes | `1676086612` |
+| Package files | `31` |
+| Validated Stage B source-package checksums SHA-256 | `d9a28d5fe6f5b675ee1b9db52e6d0493c8d3d357bb824eac590911acbd5c3ebc` |
+| Builder source commit | `9f5ce495fcb88238ec7fdc33204fa42ec9690c37` |
+
+`checksums.json` covers every file in the immutable validated release package
+at the recorded builder commit. Later model-card-only revisions are additionally
+preserved by the Hugging Face Git history.