Fizgig: trainer and workbench for image, video and music models
Free and open source; fixes broken LoRAs without retraining.
Verified
Core facts
Read by a script from the official API, not rewritten by a model.
- Repository
- https://github.com/shootthesound/Fizgig
- License
- Apache-2.0
- Languages
- Python
- Maintainer
- shootthesound
- Verified
What it is
Fizgig is a trainer and workbench for image, video and music models, in its own window on your computer, in your browser, or on a rented GPU, and you can watch a run from your phone. It is free and open source. It fine-tunes base models on consumer GPUs, down to 8 GB. It fixes broken LoRAs without retraining.
Main features
Training goes beyond LoRAs: LoKR, or the full base model itself, with Anima on an 8 GB card and the larger models on 16 GB.
LoRA Royale provides every epoch on one seed, an optional likeness score, and MP4/GIF exports.
Studio provides a live slider per block with a side-by-side preview and donor-LoRA blending.
Profiler provides a colour-coded per-block report that Studio reads inline.
Getting started
Installation needs an NVIDIA RTX 30/40/50-series or an AMD Radeon with ROCm, Windows 10/11 or Linux, Python 3.10–3.13, and 32 GB of system RAM recommended.
Clone the repo rather than downloading the ZIP, because the updater pulls with git.
Model files download from the Preferences tab, with one Download models for me button per model.
The Start tab sets your training image folder.
On the Training tab, pick a model and a preset, then click Start Training.
git clone https://github.com/shootthesound/Fizgig.git
cd FizgigWho it is for
Memory plans itself: precision, block swap and previews size to your free VRAM, and if a preview cannot fit, training keeps going. Qwen Image 2.1 trains on the base model and previews on Qwen's own Turbo (8 steps, CFG 1), and LoRAs work on both. MOSS-Image trains on Base and previews on MOSS-Image SFT.
Fizgig loads kohya, PEFT, OneTrainer, AI-Toolkit and LyCORIS LoRAs. Dataset prep includes AI captions with Qwen3-VL or Florence-2, bilingual captions, face crops, and a Look Consistency Filter. Face crops, the Look Consistency Filter and every likeness score can work on anime, furry and drawn characters and fursuits too, not just people.
Runs & outputs keeps every run with its loss graph, notes, previews and checkpoints, and starts any of them again with the same settings. Context LoRA trains on top of a frozen, active LoRA so the two coexist. Adaptive LR is a plateau tracker that raises or lowers the rate within your Min/Max, with rollback on instability.
Related tools
Related tools listed on this site.
书生大模型
The official InternLM language-model project. Its current README highlights InternLM3-8B-Instruct, normal response and deep-thinking modes, with links to model cards across releases.
Dataify
AI data-services platform. Its official site describes datasets for training, evaluation and RAG, webpage/search/video data APIs and network resources, primarily supplying data for models and applications.
OpenBMB
OpenBMB’s open-model community entry. Its public page organizes MiniCPM, AgentCPM, VoxCPM, UltraRAG and other models/projects, with community updates.
Gemma
The Gemma repository describes an open-weight model family based on Gemini research and a JAX library for using and fine-tuning Gemma. It states that it is not an official Google product.
腾讯混元大模型
An official Tencent Hunyuan project entry. This source covers Hunyuan-Large MoE models, pretrained and instruction variants, and inference and fine-tuning guidance.
讯飞星辰MaaS
The public primary description positions iFLYTEK Xingchen MaaS as a developer-facing model fine-tuning service, covering data processing, model optimization, evaluation and deployment/hosting. This review did not run those console workflows.
LLaMA
This LLaMA review is limited to the Llama 2 reference-model description in the original official repository’s current README. The repository is explicitly deprecated and points to successor model/tool projects; it is not a current-model or hosted-service catalog.
FlagEval
Evaluation toolkit and open platform for large models. Its public project covers foundation models, training and fine-tuning algorithms, with language, vision, audio and multimodal evaluation directions.
秒画
SenseTime’s AI image creation product with text-to-image generation, model and LoRA training, image collections, and an inspiration community.
Sources
This page is written from the official sources below. Open them to check the original text.
- GitHub READMEhttps://github.com/shootthesound/FizgigFetched
Verified ·
This page is written by AI from the official sources above and checked automatically. This site has not tested it. Found a mistake? Tell us.