Run 35B MoE models in ~2.5 GB of RAM. Weights stream from SSD; a trained router predicts loads a step ahead. Mac, iPhone, Android, Windows supported
Python
fix(fetch): use tuple element for MODEL env var in done hint
ahrazzlebc353dc810bd
diff --git a/scripts/fetch_models.py b/scripts/fetch_models.pyindex 67d1302..998ee12 100755--- a/scripts/fetch_models.py+++ b/scripts/fetch_models.py@@ -56,7 +56,7 @@ def _download(tier: str, target_dir: Path) -> None:print(f"[fetch] {tier} <- {repo} (-> {dest})", flush=True)snapshot_download(repo_id=repo, local_dir=str(dest))print(f"[fetch] done: {dest}\n"- f" export {TIER_REPOS[tier].replace('_REPO','_MODEL')}"+ f" export {TIER_REPOS[tier][0].replace('_REPO','_MODEL')}"f"={dest}")