f0bac67edb
Two new local image-prep operators through the already-installed mflux 0.18 (no new heavy install, ungated weights): - seedvr2_upscale: mflux-upscale-seedvr2 (ByteDance SeedVR2, MIT) — faithful local upscale, free alternative to cloud fal_upscale - mflux_image_edit: mflux-generate-qwen-edit (Qwen-Image-Edit, Apache-2.0) — local prompt-based image editing, free alternative to fal_image_edit Research verdict on dinov3 (recorded): waiting for Meta approval is the only legitimate path — TRELLIS.2's DiT is trained on DINOv3 ViT-L/16 features so the encoder can't be swapped, and the only ungated mirrors are unauthorized re-uploads. trellis_mac stays blocked on the pending gate. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
21 lines
1.1 KiB
JSON
21 lines
1.1 KiB
JSON
{
|
|
"id": "seedvr2_upscale",
|
|
"name": "Upscale (local, SeedVR2)",
|
|
"category": "image-prep",
|
|
"description": "Image → faithful upscale entirely on this Mac via mflux SeedVR2 (ByteDance, ungated/MIT). Free local alternative to the cloud fal_upscale. Sharpen thin/low-res inputs before mesh-gen or texture work. First run downloads weights.",
|
|
"accepts": ["image"],
|
|
"produces": ["image"],
|
|
"resources": "gpu",
|
|
"entry": "run.py",
|
|
"python": "/Users/m3ultra/Documents/MODELBEAST/venvs/mflux/bin/python",
|
|
"params_schema": {
|
|
"type": "object",
|
|
"properties": {
|
|
"resolution": {"type": "string", "default": "2x", "description": "Scale factor (\"2x\", \"4x\") or shortest-edge pixels (e.g. \"2048\")"},
|
|
"softness": {"type": "number", "default": 0.0, "minimum": 0.0, "maximum": 1.0, "description": "Detail/denoise strength (0 = faithful, 1 = max)"},
|
|
"seed": {"type": "integer", "default": 42, "description": "Random seed"},
|
|
"quantize": {"type": "string", "enum": ["none", "8", "4"], "default": "none", "description": "Quantize to cut memory"}
|
|
}
|
|
}
|
|
}
|