DOWNLOADS: https://huggingface.co/datasets/4eJIoBek/PAIT-Downloads/tree/main
This is my collection of portable AI packages to run it fast without anxious headache in console, sort of "Awesome N" repos, but for portables for win. initially, I made these tools for myself, but maybe someone else will need them. All portables can work offline and tested on gtx 1050 ti 4gb(cuda) and core i3 1005g1 (cpu). OK, heres the list:
!!! IF SOMETHING ISN'T WORKING, MAKE SURE THAT PATH TO TOOL DON'T HAVE SPACES OR NON-ENGLISH SYMBOLS !!!
!!! TO AVOID CUDA OUT OF MEMORY ERRORS, INSTALL NVIDIA DRIVERS 535 OR NEWER !!!
Koboldai (without models) [CPU/CUDA] - link - also in downloads / online demo
Mistral-7B-openorca 4bit koboldcpp webui (can also load mmproj loras to chat with images) [CPU/OpenCL] - in downloads / source / webui / model
Mobile AI Distribution (MAID) (same as koboldcpp, but multiplatform win-linux-andoid) [CPU] - link - also in downloads
LocalGPT Llama2-7b (w/o gui) [CUDA(tokenizing only, chat on cpu)/CPU] - in downloads / source
Facebook NLLB 600m webui [CPU] - in downloads / source / webui / model / online demo
Midi composer app [CUDA][CPU] - link - also in downloads / source / online demo
Multitrack midi music generator (generates short jingles, each instrument generated separately) [CPU] - in downloads / webui
AudioCraft Plus [CUDA/CPU] - in downloads / source / webui / online demo
Coqui XTTS2 webui (voice cloning is more "stable" than bark, but less artistic) [CUDA/CPU] - in downloads / source / webui
Suno ai Bark webui (tts is more chaotic than xtts, but if you have patience, you can roll ideal variant) (with zeroshot voice conversion) [CUDA/CPU] - in downloads / source / webui / online demo
TorToiSe webui (english-only) [CUDA/CPU] - in downloads / source / webui / online demo
RVC singing voice cloning webui [CUDA] - link - also in downloads / source
FreeVC webui [CPU] - in downloads / source / webui
Whispercpp GUI [DirectX/CPU] - link - also in downloads / source / gui / online demo
AudioSR (cli interface) [CUDA/CPU] - in downloads / source
VoiceFixer webui [CPU] - in downloads / source / webui
Dual Path RNN (w/o gui) [CPU] - in downloads / source
AudioSep webui [CUDA/CPU] - in downloads / source / webui
UVR [CPU/CUDA] - link - also in downloads / online demo
Demucs gui [CPU][CUDA] - link - also in downloads / source / gui
DeOldify .NET gui [CPU] - link - also in downloads / source / gui / online demo
DIS (BRIAAI RMBG 1.4 model) webui [CPU] - in downloads / source / webui / model
ZoeDepth-webui [CUDA/CPU][CPU] - in downloads / source / webui
real-ESRGAN-gui [Vulkan] - link - also in downloads / source / gui / online demo
ChaiNNer (supports a LOT of upscaling methods) [CUDA/Vulkan] - link - also in downloads / gui
Automatic1111 sdwebui with StableSR extension [CUDA/CPU] - in downloads / source / webui / extension
Automatic1111 Stable Diffusion base (without models) - link / webui
Automatic1111 deliberate v2 (sd1.5) model [CUDA/CPU][DIRECTX/CPU] - in downloads / source / webui / directx webui / model
Automatic1111 Illuminati Diffusion (sd2.1) model [CUDA/CPU] - in downloads / source / webui / model
Automatic1111 SDXL 1.0 (sdxl) model [CUDA/CPU] - in downloads / source / webui / model / refiner
Fooocus (sdxl) [CUDA] - link - also in downloads / source / webui / model / refiner
ComfyUI (without models) [CUDA/CPU] - link - also in downloads / source / webui
Automatic1111 Instructpix2pix (sd1.5) model (you also can download just model and use in default automatic1111 if you want, webui doesnt downloads any other files while loading this one) [DIRECTX/CPU][CUDA/CPU] - in downloads / source / ip2p source / webui / directx webui / model
Automatic1111 sd-unclip (sd2.1) model (there is an alternative that works without any models - controlnet reference) [CUDA/CPU] - in downloads / source / webui / model
LEDITS webui [CUDA/CPU] - in downloads / source / webui
lama cleaner [CPU] - in downloads / source / webui / online demo
Flowframes [CUDA/Vulkan] - in downloads / source / gui
RealBasicVSR (cli interface) [CUDA/CPU] - in downloads / source
Automatic1111 sdwebui with animatediff extension [CUDA/CPU] - in downloads / source / webui / extension / model / online demo
RobustVideoMatting (w/o gui) [CUDA/CPU] - in downloads / source / online demo
Track-anything webui [CPU] - in downloads / webui / online demo
DeepXTools by Iperov [CUDA] - link - also in downloads
Roop neurogen mod (Refacer model) (lightning fast, has realtime deepfake on webcam function) (the refacer model swaps faces better than simswap, but have only 128px resolution and may have more artifacts when head is on side) [DirectX/CUDA/CPU] - in downloads / source / webui / mod by
Deepinsight Refacer gradio webui (replaces only certain faces, has cool face upscale feature) [CUDA] - in downloads / source / webui / mod by
Simswap (w/o gui) [CUDA/CPU] - in downloads / source
DeepFaceLab (w/o gui) [DirectX][CUDA] - link - also in downloads / source
DeepfaceLive [DirectX][CUDA] - link - also in downloads / source
wav2lip gui [CUDA/CPU] - link - also in downloads / source / gui
TripoSR (outputs is still rough, but better, than shap-e) [CUDA/CPU] - in downloads / source / online demo
Shap-E webui [this model is legacy][CUDA/CPU] -in downloads / source / webui
Point-E webui [this model is legacy][CUDA/CPU] (results are worse than shap-e) - in downloads / source / webui
Dust3r webui (one model that does end-to-end photogrammetry, useful when traditional photogrammetry software like metashape dont determines camera positions, but quality may be bad) [CUDA/CPU] - in downloads / source
NERFStudio (splatfacto, nerfacto) [CUDA/CPU(cpu is extremely slow,but working)] - in downloads / source
You can theoretically run these tools on windows 7, jut download this file and place it along with python.exe
Page on itch.io: https://gz1k.itch.io/ai-portable-tools
Alternative downloads with torrents on Archive.org: https://archive.org/details/@takeonme1?tab=uploads
Page on civitai: https://civitai.com/models/104609