fix: restore precision_config, prequant_weights, fused_mod to ModelLoader
Upstream 5dea735 removed these from both ModelLoader and AutoLoader. Restore them in ModelLoader only — AutoLoader uses auto-detect and doesn't need manual precision/prequant configuration. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Q
QuantFunc committed
69f192babcba3ec2f05aceb4797440d70d5fc9f7
Parent: 545d341