Weights reload configuration

[samsja]

Make reload_weights on training from scratch configurable to allow skipping base weight reload.

---

 

---

Note

Introduces configurability for resetting inference weights when starting from scratch.

• Adds orchestrator.reload_weights_on_start (default: True) to OrchestratorConfig with descriptive docs
• Updates orchestrator.py start-up: if no checkpoint, conditionally reload_weights only when the flag is enabled; always skip base reload when model.lora is set
• Documents the new field in CHANGELOG.md

^{Written by Cursor Bugbot for commit ef0da2e. This will update automatically on new commits. Configure here.}

[cursor]

Cursor Agent can help with this pull request. Just @cursor in comments and I'll start working on changes in this branch.
_{Learn more about Cursor Agents}

[cursor]

Cursor Bugbot has reviewed your changes and found 2 potential issues.

^{Bugbot Autofix is OFF. To automatically fix reported issues with Cloud Agents, enable Autofix in the Cursor dashboard.}

[samsja]

@cursoragent can u make it true byt defaut to avoid issue

[cursor]

Done. Default is now True, plus changelog entry.