darkmaniac7/TokForge-SD15-LCM-Style-Pixelart-b1

SD1.5 LCM Pixel Art, DreamShaper 7 (MNN, batch-1)

DreamShaper 7 with the LCM-LoRA and a pixel art style LoRA fused in, as an MNN bundle that TokForge uses to make pixel art style images on Android phones.

Published by TokForge for on-device inference.

What it is	Image generator (style: pixel art)
Format	MNN diffusion bundle: batch-1 UNet with 8-bit weights, f16 text encoder and VAE decoder
Runs on	Android phones (CPU)
Phone class	8 GB RAM or more (TokForge catalog)
Download	1.21 GB (9 files)
Base model	Lykon/dreamshaper-7, latent-consistency/lcm-lora-sdv1-5, SedatAl/pixel-art-LoRa
License	CreativeML OpenRAIL-M, plus Lykon's Civitai terms for DreamShaper (no selling)

TokForge

• Website: https://tokforge.ai
• Discord: https://discord.gg/Acv3CBtfVm
• Google Play: https://play.google.com/store/apps/details?id=dev.tokforge
• iOS TestFlight: https://testflight.apple.com/join/jnufjzRr

Runs on-device in the TokForge app.

Use it in TokForge

1. Install TokForge from Google Play.
2. On Android, open Models > Recommended > Image and download Pixel Art (SD1.5 LCM Styled). TokForge lists it on phones with 8 GB RAM or more.
3. Choose it as your image model and type a prompt. The image is made on the phone; once the model is downloaded, no connection is needed.

Collections: TokForge Image Models (Android), TokForge SD1.5 LCM Styles (MNN).

Files

File	Size	What it is
alphas.txt	13.3 KB	Scheduler alphas
merges.txt	525 KB	CLIP tokenizer merges
text_encoder.mnn	76.3 MB	CLIP text encoder graph (SD1.5 reference)
text_encoder.mnn.weight	170 MB	CLIP text encoder weights (f16, SD1.5 reference)
unet.mnn	1.04 MB	UNet graph (batch-1 export, dynamic height and width)
unet.mnn.weight	862 MB	UNet weights (8-bit, LCM and pixel art style fused)
vae_decoder.mnn	154 KB	VAE decoder graph (SD1.5 reference)
vae_decoder.mnn.weight	99.0 MB	VAE decoder weights (f16, SD1.5 reference)
vocab.json	862 KB	CLIP tokenizer vocabulary

Only unet.mnn differs from the plain bundle TokForge-SD15-LCM-Style-Pixelart. Every other file, unet.mnn.weight included, is byte-identical (same sha256).

About the model

This is the TokForge SD1.5 LCM bundle (DreamShaper 7 with the LCM-LoRA fused in) with the pixel-art-LoRa art-style LoRA by SedatAl also fused into the UNet at export time, so every image comes out in that style: a retro, low-resolution pixel-art look like 16-bit game sprites. The LCM adapter lets it make a coherent 512x512 image in 4 steps with guidance-free sampling (CFG 1.0).

Architecture	Stable Diffusion 1.5 (CLIP ViT-L/14 text encoder, UNet, VAE decoder)
Base checkpoint	DreamShaper 7 by Lykon
Few-step adapter	LCM-LoRA SD1.5, fused into the UNet
Style	pixel-art-LoRa by SedatAl, fused into the UNet
UNet	8-bit weights (asymmetric), batch 1, dynamic height and width
Text encoder and VAE decoder	SD1.5 reference MNN files, f16
Output	512x512

Render settings

Setting	Value
Steps	4 (TokForge default); up to 8 for extra refinement
CFG scale	1.0 (guidance-free)
Negative prompt	Not used at CFG 1.0
Size	512x512

Batch-1 and plain bundles

This model is published twice. The weights are the same; only the UNet graph differs.

Repo	UNet export	Downloaded by
TokForge-SD15-LCM-Style-Pixelart-b1 (this repo)	batch 1, dynamic height and width (sample=[1,4,-1,-1])	TokForge 3.6.0 beta 24 and later, including the current 1.3.x releases
TokForge-SD15-LCM-Style-Pixelart	fixed batch 2 (conditional and unconditional)	Older TokForge builds, for example 3.6.0 beta 12

LCM sampling is guidance-free (CFG 1.0), so the unconditional UNet pass adds nothing. With the batch-1 UNet, TokForge's engine runs one conditional UNet pass per step instead of a batch of two. Older builds such as TokForge 3.6.0 beta 12 always feed the UNet a batch of two and would crash on the batch-1 export, so they keep downloading the plain repo by its exact name. The two repos are separate so that each app version gets the files it can run.

How it was made

1. In diffusers, load Lykon/dreamshaper-7, then load and fuse latent-consistency/lcm-lora-sdv1-5 and, on top of it, SedatAl/pixel-art-LoRa (one LoRA at a time, fuse_lora in fp32).
2. Export the fused UNet to ONNX, traced at batch 1 with dynamic height and width, so MNN reports sample=[1,4,-1,-1].
3. Convert the UNet to MNN with 8-bit weights (command below).
4. Add the SD1.5 reference MNN text encoder and VAE decoder (f16) and the tokenizer and scheduler files. The LoRAs only change the UNet, so these are reused unchanged and are byte-identical across the TokForge SD1.5 LCM bundles.

MNNConvert -f ONNX --modelFile unet/model.onnx --MNNModel unet.mnn \
  --saveExternalData=1 --weightQuantBits 8 --weightQuantAsymmetric

Limitations

• Built for TokForge's MNN diffusion engine. This is a runtime bundle, not a diffusers checkpoint.
• Needs an engine that can run a batch-1 UNet. Older TokForge builds (for example 3.6.0 beta 12) must use the plain repo.
• Use 4 to 8 steps: below 4 the output is noise, and quality falls off above 8.
• The style is baked into the UNet, so its strength cannot be changed at render time.
• The UNet weights are 8-bit, which may slightly change fine detail compared with the full-precision model.

License

This bundle is shared under the CreativeML OpenRAIL-M license, like its SD1.5 base, including its use-based restrictions. Sources and their licences:

• Base: Lykon/dreamshaper-7: CreativeML-OpenRAIL-M
• LCM adapter: latent-consistency/lcm-lora-sdv1-5: openrail++
• Style LoRA: SedatAl/pixel-art-LoRa: CreativeML-OpenRAIL-M

DreamShaper also carries Lykon's Civitai terms (no selling the model or merges of it); see Credits. Fused with diffusers.fuse_lora() (fp32 ΔW before INT8 quant); CLIP/VAE reused unchanged from the SD1.5 reference (the style LoRA is UNet-only).

Credits

• DreamShaper 7 by Lykon (Civitai, Hugging Face). Used under the CreativeML OpenRAIL-M license and the creator's Civitai terms. Lykon's Civitai terms: no selling the model or merges of it.
• LCM-LoRA SD1.5 by Latent Consistency (Hugging Face). Used under the CreativeML OpenRAIL++-M license.
• pixel-art-LoRa by SedatAl (Hugging Face). Used under the CreativeML OpenRAIL-M license.
• Stable Diffusion v1.5 by Runway and CompVis (Hugging Face). Used under the CreativeML OpenRAIL-M license.
• Converted to MNN with MNN by Alibaba (Apache-2.0).