Enable benchmarked two-slot medium profile
This commit is contained in:
@@ -59,12 +59,12 @@ def render_docs(data: dict) -> str:
|
||||
"",
|
||||
f"Standardprofil: **{data['default_profile']}** · globales Ausgabelimit: **{data['max_output_tokens']} Token**",
|
||||
"",
|
||||
"| Profil | API-Alias | Kontext | Modell | GPU-Verteilung | Vision | MTP |",
|
||||
"|---|---|---:|---|---|---|---:|",
|
||||
"| Profil | API-Alias | Gesamtkontext | Slots | Modell | GPU-Verteilung | Vision | MTP |",
|
||||
"|---|---|---:|---:|---|---|---|---:|",
|
||||
]
|
||||
for item in data["profiles"]:
|
||||
lines.append(
|
||||
f"| {item['id']} | `{item['alias']}` | {item['context']:,} | "
|
||||
f"| {item['id']} | `{item['alias']}` | {item['context']:,} | {item.get('parallel_slots', 1)} | "
|
||||
f"{item.get('model_family', '')} | {item.get('gpu_split', '')} | "
|
||||
f"{'ja' if item.get('vision') else 'nein'} | {item.get('mtp', '')} |"
|
||||
)
|
||||
@@ -92,6 +92,9 @@ def verify_compose(data: dict, compose: pathlib.Path) -> None:
|
||||
env_prefix = item["id"].upper()
|
||||
if f"${{{env_prefix}_CONTEXT:-{item['context']}}}" not in block:
|
||||
raise SystemExit(f"Compose context drift for {item['id']}")
|
||||
slots = int(item.get("parallel_slots", 1))
|
||||
if f"${{{env_prefix}_PARALLEL_SLOTS:-{slots}}}" not in block:
|
||||
raise SystemExit(f"Compose parallel-slot drift for {item['id']}")
|
||||
|
||||
|
||||
def main() -> None:
|
||||
|
||||
Reference in New Issue
Block a user