ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

studio: thread List[str] quantization_method through orchestrator + worker

Browse source 
ExportOrchestrator.export_gguf's parameter type is widened from str to
List[str]. Worker's cmd.get default becomes ["Q4_K_M"] to match.

The orchestrator pickles the list into the mp.Queue cmd dict; the
subprocess worker hands it to ExportBackend.export_gguf unchanged.
Both layers are pure passthrough for this field.
This commit is contained in:
Roland Tannous 2026-04-14 22:46:29 +04:00
commit b9e5a13d4e
2 changed files with 7 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

8
studio/backend/core/export/orchestrator.py
View file

@ -310,12 +310,16 @@ class ExportOrchestrator:
def export_gguf(
self,
save_directory: str,
quantization_method: str = "Q4_K_M",
quantization_method: List[str],
push_to_hub: bool = False,
repo_id: Optional[str] = None,
hf_token: Optional[str] = None,
) -> Tuple[bool, str]:
"""Export model in GGUF format."""
"""
Export model in GGUF format. The caller must supply a normalized
list of lowercase quantization method strings (see
`models.export.normalize_gguf_quantization_method`).
"""
return self._run_export(
"gguf",
{

2
studio/backend/core/export/worker.py
View file

@ -139,7 +139,7 @@ def _handle_export(backend, cmd: dict, resp_queue: Any) -> None:
elif export_type == "gguf":
success, message = backend.export_gguf(
save_directory = cmd.get("save_directory", ""),
quantization_method = cmd.get("quantization_method", "Q4_K_M"),
quantization_method = cmd.get("quantization_method", ["Q4_K_M"]),
push_to_hub = cmd.get("push_to_hub", False),
repo_id = cmd.get("repo_id"),
hf_token = cmd.get("hf_token"),