a928900a2e
Fourth round of review follow-ups: - Local candidates are now resolved before the fallback groups are sorted, and multi-quant GGUF folders order by the resolved quant's own size. The backend row size sums every quant in a folder, so a folder holding a 2 GB and an 8 GB quant previously sorted after a 4 GB single-quant model even though its 2 GB quant was the smallest loadable artifact. - The cached/local dedupe keys now carry the model kind. A folder emitting both GGUF and safetensors rows shares one path while holding two different models, so a failing GGUF row no longer shadows its loadable safetensors sibling. Contract tests updated and extended for both. |
||
|---|---|---|
| .. | ||
| backend | ||
| frontend | ||
| src-tauri | ||
| __init__.py | ||
| install_llama_prebuilt.py | ||
| install_node_prebuilt.py | ||
| install_python_stack.py | ||
| install_whisper_prebuilt.py | ||
| LICENSE.AGPL-3.0 | ||
| MCP.md | ||
| node_prebuilt_pins.json | ||
| package-lock.json | ||
| package.json | ||
| prebuilt_core.py | ||
| setup.bat | ||
| setup.ps1 | ||
| setup.sh | ||
| Unsloth_Studio_Colab.ipynb | ||