MakazhanAlpamys/Soup v0.2.0

on GitHub

What's New in v0.2.0

• Merge LoRA — soup merge merges adapter with base model into a standalone model
• Export to GGUF — soup export converts models to GGUF format for Ollama / llama.cpp
• Resume training — soup train --resume auto continues from latest checkpoint
• Weights & Biases — soup train --wandb sends metrics to W&B

Install / Upgrade

pip install --upgrade soup-cli

Full Changelog: v0.1.0...v0.2.0