CatalogWantedMirrorsLedgerOperating StatementDocsStudio statusCard system
alt.freeweights.image / sd-vae-ft-mse-original
From: stabilityai <huggingface.co/stabilityai>Newsgroups: alt.freeweights.imageSubject:

sd-vae-ft-mse-original

Date: 2026-11-14Commit: 629b3adSigned: 2DC28CC7B26A283B ✓

Text to image model. Safetensors weights, 669 MB in all. Published by stabilityai under MIT.

image669MBsd-vae-ft-mse-originalsafetensorspermissive
FactsParamsnot statedSize669 MBFormatssafetensorsTasktext to imageLicenseMITFrom sourcepermissive
$ fw get stabilityai/sd-vae-ft-mse-original
# manifest sha256 9ddf5db0af1d… · 669 MB · 4 files
$ fw verify ./sd-vae-ft-mse-original stabilityai/sd-vae-ft-mse-original
From sourceHosted (signed)stabilityai/sd-vae-ft-mse-originalLast check: 2026-11-15 02:25 UTC
Run at homeGet help

Get help: custom fine-tune or private deploy. We review your setup before we quote.

Works with

Run16Fine-tune8Convert8Scan2
OverviewModel cardFilesUse itLicenseProvenanceModel tree

Overview

Size
669 MB
Weights
335 MB (safetensors)
Parameters
not stated
Formats
safetensors
Task
text to image
License
MIT (permissive)
Files
4, each with a sha256
Publisher
stabilityai
Added
2026-11-14
Section
alt.freeweights.image
Lineage

Original release from the publisher.

Where these bytes come from
Publisher at the pinned commit→You

Sources are tried in order. A file counts only when its size and sha256 match the manifest. Sources are checked when you ask for them, not on a timer.

Sources
BadgeHostLast checkResult
From sourcehuggingface.co2026-11-15 02:25 UTCpassed

Model card

From the publisher at commit 629b3ad. Shown as plain text with links; images open at the publisher.

Improved Autoencoders#

Utilizing#

These weights are intended to be used with the original CompVis Stable Diffusion codebase. If you are looking for the model to use with the 🧨 diffusers library, come here.

Decoder Finetuning#

We publish two kl-f8 autoencoder versions, finetuned from the original kl-f8 autoencoder on a 1:1 ratio of LAION-Aesthetics and LAION-Humans, an unreleased subset containing only SFW images of humans. The intent was to fine-tune on the Stable Diffusion training set (the autoencoder was originally trained on OpenImages) but also enrich the dataset with images of humans to improve the reconstruction of faces. The first, ft-EMA, was resumed from the original checkpoint, trained for 313198 steps and uses EMA weights. It uses the same loss configuration as the original checkpoint (L1 + LPIPS). The second, ft-MSE, was resumed from ft-EMA and uses EMA weights and was trained for another 280k steps using a different loss, with more emphasis on MSE reconstruction (MSE + 0.1 * LPIPS). It produces somewhat ``smoother'' outputs. The batch size for both versions was 192 (16 A100s, batch size 12 per GPU). To keep compatibility with existing models, only the decoder part was finetuned; the checkpoints can be used as a drop-in replacement for the existing autoencoder..

Original kl-f8 VAE vs f8-ft-EMA vs f8-ft-MSE

Evaluation#

COCO 2017 (256x256, val, 5000 images)#

| Model | train steps | rFID | PSNR | SSIM | PSIM | Link | Comments

|----------|---------|------|--------------|---------------|---------------|-----------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------|

| | | | | | | | |

| original | 246803 | 4.99 | 23.4 +/- 3.8 | 0.69 +/- 0.14 | 1.01 +/- 0.28 | https://ommer-lab.com/files/latent-diffusion/kl-f8.zip | as used in SD |

| ft-EMA | 560001 | 4.42 | 23.8 +/- 3.9 | 0.69 +/- 0.13 | 0.96 +/- 0.27 | https://huggingface.co/stabilityai/sd-vae-ft-ema-original/resolve/main/vae-ft-ema-560000-ema-pruned.ckpt | slightly better overall, with EMA |

| ft-MSE | 840001 | 4.70 | 24.5 +/- 3.7 | 0.71 +/- 0.13 | 0.92 +/- 0.27 | https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt | resumed with EMA from ft-EMA, emphasis on MSE (rec. loss = MSE + 0.1 * LPIPS), smoother outputs |

LAION-Aesthetics 5+ (256x256, subset, 10000 images)#

| Model | train steps | rFID | PSNR | SSIM | PSIM | Link | Comments

|----------|-----------|------|--------------|---------------|---------------|-----------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------|

| | | | | | | | |

| original | 246803 | 2.61 | 26.0 +/- 4.4 | 0.81 +/- 0.12 | 0.75 +/- 0.36 | https://ommer-lab.com/files/latent-diffusion/kl-f8.zip | as used in SD |

| ft-EMA | 560001 | 1.77 | 26.7 +/- 4.8 | 0.82 +/- 0.12 | 0.67 +/- 0.34 | https://huggingface.co/stabilityai/sd-vae-ft-ema-original/resolve/main/vae-ft-ema-560000-ema-pruned.ckpt | slightly better overall, with EMA |

| ft-MSE | 840001 | 1.88 | 27.3 +/- 4.7 | 0.83 +/- 0.11 | 0.65 +/- 0.34 | https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt | resumed with EMA from ft-EMA, emphasis on MSE (rec. loss = MSE + 0.1 * LPIPS), smoother outputs |

Visual#

Visualization of reconstructions on 256x256 images from the COCO2017 validation dataset.

256x256: ft-EMA (left), ft-MSE (middle), original (right)

Image: image

Image: image

Image: image

Image: image

Image: image

Image: image

Files

4 files · 669 MBWeights 335 MB (safetensors). Sizes and sha256 from the signed manifest.
Hide all 4 files with their sha256
FileSizesha256
.gitattributes1.53 KB713eb0800b91205feeb3d6ca2fb83487f7fb23c5e5b7ffac3fc37578a4841e28
README.md6.36 KB370143cb387f18a88727d83f6dc5bd86218b58a0cb4d89b07c05b40a80970c83
vae-ft-mse-840000-ema-pruned.ckpt335 MBc6a580b13a5bc05a5e16e4dbb80608ff2ec251a162311590c1f34c013d7f3dab
vae-ft-mse-840000-ema-pruned.safetensors335 MB735e4c3a447a3255760d7f86845f09f937809baa529c17370d83e4c3758f3c75
VersionsEach time the pinned upstream commit or file list changed in this repo. Newest first.
Upstream commitDateContents
629b3ad303 pinned now2026-11-15 02:25 UTC4 files · 669 MB

Use it

$ fw get stabilityai/sd-vae-ft-mse-original
# manifest sha256 9ddf5db0af1d… · 669 MB · 4 files
$ fw verify ./sd-vae-ft-mse-original stabilityai/sd-vae-ft-mse-original
fw.pyChecks the signature and every sha256, tries each source in order, resumes.
$ python3 fw.py get https://freeweights.allternit.com/manifests/stabilityai/sd-vae-ft-mse-original.json --dir ./sd-vae-ft-mse-original $ python3 fw.py verify ./sd-vae-ft-mse-original ./stabilityai__sd-vae-ft-mse-original.json
hfThe Hugging Face CLI with one setting. Same repo id.
$ HF_ENDPOINT=https://freeweights.allternit.com hf download stabilityai/sd-vae-ft-mse-original --revision 629b3ad3030ce36e15e70c5db7d91df0d60c627f --local-dir ./sd-vae-ft-mse-original
Packaged for ComfyUI. Most image, video and music models run in ComfyUI . Check the publisher's page for the workflow. Freeweights does not host it.
Run at home

License

Name
MIT License
Id
MIT
Class
Permissive
Redistribution
yes
Listing
Hosted (signed)
Decided
License as declared by the publisher. No person read it. 2026-11-14
Text in
README.md
Text sha256
370143cb387f18a88727d83f6dc5bd86218b58a0cb4d89b07c05b40a80970c83
Terms gate
No extra terms beyond the license.

Read the license text

Licenses are checked against the license table. A custom license passes only after a person reads it.

Provenance

From: stabilityai <huggingface.co/stabilityai>Newsgroups: alt.freeweights.imageSubject: sd-vae-ft-mse-originalDate: 2026-11-14Commit: 629b3adSigned: 2DC28CC7B26A283B ✓
Upstream
stabilityai/sd-vae-ft-mse-original
Pinned commit
629b3ad3030ce36e15e70c5db7d91df0d60c627f
Manifest
sd-vae-ft-mse-original.json · .minisig
Manifest sha256
9ddf5db0af1d7172f38be8a8fd5fdfc62c273e465a073c47ea89b650246205bb
Signature
minisign, Ed25519, key 2DC28CC7B26A283B
Created
2026-11-15 02:25 UTC
Generator
freeweights-manifest/0.1.0
Takedowns
None.

How to verify

$ curl -O https://freeweights.allternit.com/manifests/stabilityai/sd-vae-ft-mse-original.json $ curl -O https://freeweights.allternit.com/manifests/stabilityai/sd-vae-ft-mse-original.json.minisig $ minisign -Vm sd-vae-ft-mse-original.json -P RWQ7KGqyx4zCLXtUFOqvfv8L3g+8fpKJ1FsuxPyOYMF+4KufbUxKq1Oq $ shasum -a 256 <file> # compare with the Files tab

A manifest whose signature fails must be rejected. A file whose size or sha256 does not match must be discarded. Verify a download

Model tree

No related models in this catalog.Nothing here lists this model as its parent, and it names no parent.
Get it
Download
$ HF_ENDPOINT=https://freeweights.allternit.com hf download stabilityai/sd-vae-ft-mse-original
Open in Freeweights Studio

Pull, verify and run it in the Allternit app on Desktop, web or phone.

Signed manifest
minisign, key 2DC28CC7B26A283B. sd-vae-ft-mse-original.json. Provenance. Takedown history: none.
Source
huggingface.co/stabilityai/sd-vae-ft-mse-original

Related models

All of alt.freeweights.image →
From: Comfy-Org <huggingface.co/Comfy-Org>Subject: Boogu-Image
From: Comfy-Org <huggingface.co/Comfy-Org>Subject: Chroma1-HD
From: lodestones <huggingface.co/lodestones>Subject: Chroma1-HD
From: Comfy-Org <huggingface.co/Comfy-Org>Subject: Chroma1-Radiance
From: Comfy-Org <huggingface.co/Comfy-Org>Subject: ERNIE-Image
From: baidu <huggingface.co/baidu>Subject: ERNIE-Image