TheTensorTune is a single-file LoRA fine-tuning workbench for language models: the whole pipeline — from dataset to GGUF export — lives on one page. The UI is a node graph: every stage is a block, and blocks are wired together.
- Canvas: drag blocks freely; drag from a block's output port onto the next block's input port to wire them.
- Topbar: add block, Queue, History, service token, Learn and the FA/EN language key.
- Zoom controls: bottom-left corner — zoom in/out, reset and auto layout.
- Status bar: shows the training stack status at the bottom of the page.
Tip: the FA/EN key switches the entire UI between Persian (full RTL) and English; your choice is stored in the browser.