Pinokio
FeedApps
Download Pinokio
Sign inCreate accountJoin
Sign inCreate accountJoin
Home

Tags

Browse
#bug#cool-project#guide#pinokio-feedback#pinokio-question#showcase#testers-wanted
PrivacyTerms

#h3

Browse tags

DiscoverPinned

Most posts · Last 30 days

No recent tagged posts yet.

@bizarro·24d

Phosphene 4.16.3 - three fixes from the field

Three fixes, all from things you ran into. How to get it: press Update in Pinokio, then restart Phosphene. No new downloads. Hailuo H3 progress says how long is left. In the middle of a 15 s clip the progress card read "27m 23s / ~18m 21s", which looked like an 18-minute total. It now reads "27m 23s in - ~18m 21s left", and the estimate also counts the model load for each part still to come. Thanks @abandonedmuse. A LoRA with a cut-short download is named, not a crash. An interrupted download keeps the start of the file, so the LoRA looked fine until the render died with "cannot reshape array

Phosphenemrbizarro/phosphene2.1K followers
#fixes#h3#phosphene
2
@bizarro·26d

Phosphene 4.16 - the Music Studio, 70+ fixes, and everything since 4.15

Everything since the 4.15 post, newest first. To get it: press Update in Pinokio, then restart Phosphene. 4.16.2 - two fixes you reported here The outputs list stays on screen. On a 1080p screen a square or wide clip made the preview so tall that the outputs list went off the bottom, with no way to select or delete clips. The preview now scales down so the outputs header and a full row of clips always stay visible. Big screens look the same as before. Thanks @fuschichou. Honest Hailuo H3 times on 36-59 GB Macs. The estimates were measured on 64 GB. On a 48 GB Mac, 10 s and 15 s clips take abou

Phosphenemrbizarro/phosphene2.1K followers
#fixes#h3#music#phosphene
2
@bizarro·1mo ago

Phosphene 4.15 - HD Hailuo H3 clips in five minutes

What ships: a Speed switch on Hailuo H3. Fast runs a 3-step adapter, Best runs the full schedule - and with Upscale & Face Fix after it, a finished HD clip costs five minutes instead of half an hour. Measured on an M4 Max, 5-second clip, character LoRA on: Draft + Fast - 2.8 min render + 2.5 min Upscale & Face Fix = 5.2 minutes, 1280x768 High + Fast - 7.9 min render + 8.6 min Upscale & Face Fix = 16.5 minutes, 2048x1152 That clip below is the first one: five minutes, start to finish, on a laptop. For scale: the infomercial those Bizarro clips come from was shot before Fast existed. 50 takes, 2

Phosphenemrbizarro/phosphene2.1K followers
#editor#h3#phosphene#speed
10
Cekli.com@yoyokits·1mo ago

[Fixed] Maestro AMD Error: MiniMax H3 renders a flat-grey video on my AMD card— and the fix was one dropdown

TL;DR: If MiniMax H3 in Maestro / WanGP "finishes" a generation but the file is a tiny (~50 KB) solid-grey clip on an AMD GPU, open Advanced Settings → H3 Text Encoder and make sure it is GGUF Q4_K_M, not GGUF Q2_K. The 2-bit Q2_K encoder is unreliable for H3 on the ROCm-for-Windows build. Everything else I tried — TDR registry tweaks, reboots, resolution cuts, rolling Maestro back a version — was a dead end. The setup GPU: Radeon RX 7900 XTX (RDNA 3, gfx1100), 24 GB VRAM RAM: 32 GB OS: Windows 11 App: Maestro AMD (a Pinokio wrapper around Blizaine/Maestro, which is built on the WanGP pipeline

Maestro AMDyoyokits/Maestro-AMD134 followers
#amd#h3#troubleshooting
0
You’re all caught up.

Join the community

Follow people and apps for updates. Explore Announcements and All for more.

Sign inCreate account
Discover people →

Get project updates

Popular builders

View all
PrivacyTerms
Pinokio
PrivacyTerms
FeedApps

Radar

View all

Projects getting attention.

Followed·just now
Maestro
An all-in-one, 100% local AI video, image & music studio. Its Director mode turns a single prompt into a full music video or short film — LLM-planned, shot by shot. Built on the WanGP pipeline (Wan 2.1/2.2, LTX-2.3, Qwen, Hunyuan Video, Flux). Requires an NVIDIA GPU (6GB+ VRAM).
Installable
Followed·1 min
ChatterBox
AI-Powered Text-to-Speech with Voice Cloning using Chatterbox TTS and a Gradio interface. Includes Turbo, Multilingual (23+ languages), and Original models. Runs locally; CUDA GPU recommended, CPU supported. Windows, Mac, and Linux.
Installable
Followed·2 min
AI Harness
Chat local com o Codex CLI, memória real em SQLite (com extração automática pela IA) e um atlas neural 3D da sua memória.
Installable
Followed·3 min
Maestro
An all-in-one, 100% local AI creative studio, director, and multi-track editor. Generate with MiniMax H3, LTX-2.5/2.3, Wan, Flux, Qwen, and more; turn an idea or song into a planned production; then finish it on the timeline. Requires an NVIDIA GPU (6GB+ VRAM).
Installable
Followed·3 min
Maestro
An all-in-one, 100% local AI creative studio, director, and multi-track editor. Generate with MiniMax H3, LTX-2.5/2.3, Wan, Flux, Qwen, and more; turn an idea or song into a planned production; then finish it on the timeline. Requires an NVIDIA GPU (6GB+ VRAM).
Installable

Popular builders

View all
  • cocktailpeanut@cocktailpeanut
    Creator of
    • CozyClay
    • mimetic-a0
    • theprototype.app

    creator of pinokio https://x.com/cocktailpeanut

  • Morpheus@morpheus
    Creator of
    • Z-Fusion
    • Freebuff
    • songgeneration-tencent

    404 Error not found Don't hesitate to ask if you have questions. I'm here to help. Tag me if you can.

  • bilawalsidhu@bilawalsidhu
    Creator of
    • see-through-walls
    • God's Eye View
3 people shown