v0.8.0 release checklist

[ssube]

Issues:

• 54

• 113

• 122

• 127

• 132

• 138

• 141

• 144

• 145

• 151

• 152

• 155

• 156

• 157

• 158

• 160

• 161

• 166

• 167

• 170

• 172

• 179

• 180

• 181

• 187

• 188

• 193

• 194

• 195

• 221

• 222

• 224

• 226

• 228

• 229

• 233

• 234

Merges:

• 45

• 163

• 169

• 173

• 174

• 175

• 176

• 177

• 178

• 199

• 200

• 204

• 205

• 206

• 208

• 209

• 214

• 215

• 218

• 220

• 223

Testing:

• txt2img
  • 256x256 with SD v1.5
  • [x] ~should fail: neon blobs~ worked, but that's ok
  • [x] has automation
  • 512x512 with SD v1.5
  • DEIS Multi
    • [x] should work
  • DPM Multi
    • [x] should work
    • [x] has automation
  • Euler A
    • [x] should work
  • Heun
    • [x] should work
    • [x] has automation
  • 512x512 with SD v2.1
  • [x] should work
  • [x] has automation
  • 768x768 with SD v2.1
  • [x] should work, given sufficient memory
  • [x] has automation
  • extra models
  • 512x512 with Knollingcase
    • [x] should work
    • [x] has automation
  • 512x512 with OpenJourney
    • [x] should work
    • [x] has automation
  • 256x256 with OpenJourney
    • [x] should work
  • textual inversion
  • 512x512 with SD v1.5 + Autumn (embeddings)
    • [x] should work
  • 512x512 with SD v1.5 + Minecraft (concept)
    • [x] should work
• img2img
  • 256x256 input
  • [x] should work: expanded into valid image
  • [x] has automation
  • 512x512 input
  • [x] should work
  • [x] has automation
  • 1024x768 input
  • [x] should work
• inpaint
  • regular inpaint
  • black mask
    • [x] should keep all pixels, same image
    • [x] has automation
  • white mask
    • [x] should replace all pixels, different image
    • [x] has automation
  • outpaint
  • 0 all sides
    • [x] should work, run 1 tile
    • primarily a client-side test
  • 256 all sides
    • [x] should work, run 8 tiles
    • [x] has automation
  • 512 top and bottom, 0 left and right
    • [x] should work, run 3 tiles
    • [x] has automation
  • 512 left and right, 0 top and bottom
    • [x] should work, run 3 tiles
    • [x] has automation
• upscale
  • Real ESRGAN
  • x4 with CodeFormer
    • [x] should work
  • x4 with GFPGAN
    • [x] should work
  • x4 without face correction
    • [x] should work
    • [x] has automation
  • x2 without face correction
    • [x] should work
    • [x] has automation
  • x2 model and x4 scale
    • [x] should sort of work: ignores scale and uses x2
  • x4 model and x2 scale
    • [x] should fail: tiles
  • v3 model and x4 scale
    • [x] should work
  • Stable Diffusion
  • using x2 scale
    • [x] should fail: tiles
  • using x4 scale
    • [x] should work
  • with CodeFormer
    • [x] should work
  • with GFPGAN
    • [x] should work
  • without face correction
    • [x] should work
• blend
  • two 512x512 inputs
  • [x] should work
  • two 1024x1024 inputs
  • [x] should work
  • two different size inputs
  • 256x256 and 512x512
    • [x] should work
  • 512x512 and 1024x1024
    • [x] should work
• interactions
  • generate a new image
  • [x] should request and then load an image from the server
  • delete a pending image
  • [x] should remove a single image and leave the rest
  • delete a finished image
  • [x] should remove a single image and leave the rest
  • copy an image to img2img
  • [x] should switch to the img2img tab
  • [x] should populate the image source
  • [x] the generate button should be enabled
  • copy an image to inpaint
  • [x] should switch to the inpaint tab
  • [x] should populate the image source
  • [x] the generate button should be enabled
  • copy an image to upscale
  • [x] should switch to the upscale tab
  • [x] should populate the image source
  • [x] the generate button should be enabled
  • copy two images to blend
  • [x] should switch to the blend tab
  • [x] should populate the image sources
  • [x] the generate button should be enabled once both sources have been populated
  • state should persist on refresh
  • [x] loading images
  • [x] switching tabs
  • [x] images sources do not
• schedulers
  • [x] DDIM
  • [x] DDPM
  • [x] DEIS Multi
  • [x] DPM Multi
  • [x] DPM Single
  • [x] Euler A
  • [x] Euler
  • [x] Heun
  • [ ] iPNDM
  • [x] KDPM2 A
  • [x] KDPM2
  • [ ] Karras Ve
  • [x] LMS
  • [x] PNDM
  • [x] UniPC Multi

Release:

• [x] fix any late arrival bugs
• [x] check on/add UniPCMultistepScheduler
• [x] document ONNX_WEB_EXTRA_ARGS
• [x] more bugs
• [x] update release tests to include Textual Inversions
• [x] check and fix lint
• [x] update package versions and stage files
• [x] run commit-and-tag-version --sign --git-tag-fallback --commit-all --release-as=minor to make VERSION
• [x] make sure packages and images have been built
• [x] update GH pages bundle and default version
• [x] post release on GH
• [x] make follow up tickets
• [x] close milestone and checklist

[ssube]

There are a few bugs around model conversion that still need to be fixed, as well as LoRA token expansion.

[ssube]

All of the bugs that have been reported so far have been fixed, and the release test script passed on my Win10 + AMD + DirectML and Ubuntu + Nvidia + CUDA systems. Unfortunately, my Ubuntu + AMD + ROCm test machine is not cooperating, so I need to fix that or will have to release without specifically testing ROCm, which I don't like very much.

The new ONNX_WEB_EXTRA_ARGS variable still needs to be documented and I need to check if the UniPCMultistepScheduler can be added yet.

[ssube]

• the Karras Ve scheduler is still broken (TypeError: KarrasVeScheduler.step() missing 1 required positional argument: 'sample_hat')
• iPNDM is still producing chips (LMS and PNDM may as well without enough steps)
• the Generate button on the Blend tab is enabled even without the second source