Pinokio
FeedApps
Install PinokioSign in
Author@cinemazverev
Sign in

@cinemazverev

1 Following0 Followers
PostsReplies
More
ActivityCheckpointsFollowed appsGitHub creations
Posts · 6
@cinemazverev·2mo ago

I2V and T2V loras storage

The left App Menu in the Pinokio interface "T2V Loras (save lora files here)" directs to correct folder, but I2V directs to system C-disk Documents folder. Found that pinokio.js file has this lines: }, { icon: "fa-regular fa-folder-open", text: "T2V Loras (save lora files here)", href: "app/loras", fs: true }, { icon: "fa-regular fa-folder-open", text: "I2V Loras (save lora files here)", href: "app/loras i2v", fs: true }, { But "app/loras i2v" folder does not excist. Shall i create it, or both T2V and I2V can use the same "app/loras" folder and subfolders (like it works in wan2gp)? And what ab

Read full post
AboutMaestroBlizaine/MaestroView app
#question
0
@cinemazverev·2mo ago

Sam 3: numpy dependencies conflict. Versions are incompatible

Every new update i face the numpy version conflict, and every time i need to downgrade numpy Requirement already satisfied: timm =1.0.17 in .\app\services\sam\env\Lib\site-packages (from -r app/services/sam/requirements.txt (line 6)) (1.0.28) Collecting numpy =2 (from -r app/services/sam/requirements.txt (line 7)) Using cached numpy-2.5.1-cp312-cp312-win amd64.whl.metadata (6.6 kB) So, it strats to update from requirements.txt which is =2 - but this versions are incompatible: Using cached numpy-2.2.6-cp312-cp312-win amd64.whl (12.6 MB) Installing collected packages: numpy Attempting uninstall:

Read full post
AboutMaestroBlizaine/MaestroView app
#bug
0
@cinemazverev·2mo ago

Maestro installs in Pinokio, but Open Web UI is blank and Python environment setup is incomplete

NVIDIA GeForce RTX 3080 Ti Laptop GPU (16GB VRAM), 32GB RAM, Intel 12th Gen Core i9 Pinokio 8.0.8. Just installed and got "Installation complete" with this LOG, E:\ AI APPs\pinokio\api\Maestro.git\ui conda hook & conda deactivate & conda deactivate & conda deactivate & conda activate base && npm run build ui@0.0.0 build tsc -b && vite build vite v7.3.2 building client environment for production... ✓ 1798 modules transformed. dist/index.html 1.49 kB │ gzip: 0.77 kB dist/assets/index-CcKRZe0V.css 70.06 kB │ gzip: 11.91 kB dist/assets/index-B17Rab9m.js 769.64 kB │ gzip: 199.30 kB (!) Some chunks

Read full post
AboutMaestroBlizaine/MaestroView app
#bug
1
@cinemazverev·2mo ago

Pinokio Update Breaks VRAM Validation for 4B DiT Models & Flash Attention (Python 3.12 Environment Mismatch)

NVIDIA GeForce RTX 3080 Ti Laptop GPU (16GB VRAM), 32GB RAM, Intel 12th Gen Core i9 A recent Pinokio update migrated the application's underlying environment to Python 3.12.8, breaking the memory validation logic and rendering the Flash Attention initialization toggle completely inactive/grayed out in the Gradio UI. Consequently, the service initialization block now flags and hard-blocks flagship 4B models (acestep-v15-xl-sft and acestep-v15-xl-turbo-bf16) on 16GB VRAM GPUs (such as the RTX 3080 Ti Laptop GPU). It throws a false error stating the model is too heavy, forcing a switch to 2B vari

Read full post
AboutACE-Step 1.5cocktailpeanut/ace-step.pinokioView app
#bug
9
@cinemazverev·3mo ago

Time_Out

found that time-out limit is only 300 sec by default — too short for any large file, especially with slow inet connection. you can fix it by editing app.py file in your reclip folder (..\pinokio\api\reclip-pinokio.git\app) with notepad. search for this line: result = subprocess.run(cmd, capture output=True, text=True, timeout=300) and change to any value you need, for ex timeout=1800 will be 30 min, timeout=3600 will be 1 hour etc. after editing, restart reclip to take effect

Read full post
AboutReClipkrynsky/reclip-pinokioView app
#tip
2
@cinemazverev·3mo ago

ACE-Step 1.5 XL - SFT (4B DiT)

Hello, trying to use xl-sft base model (installed from HF official repo) with initialized offload to CPU and Dit offload to CPU. Extra parameters such as "Use ADG (Angle Domain Guidance)" appeared. UI shows up to 200 steps, but during generation real number of steps differs: 9 instead of 50, 28 inst of 100, and 30 inst of 200. so, real max number of steps is only 30. After risky Offload DiT to CPU - OFF (and deactivated LM model) number of steps grew up, but fell down again: | Set | Real | % | Offload to CPU | Offload DiT to CPU | GPU | | ---------------- | -------------- | ----- | -----------

Read full post
AboutACE-Step 1.5cocktailpeanut/ace-step.pinokioView app
#question
3
You’re all caught up.
Pinokio
PrivacyTerms
FeedApps