Add long-context attention profile closure
This commit is contained in:
@@ -276,6 +276,7 @@ def main() -> None:
|
||||
loads = []
|
||||
config_knobs = BASE.knobs(config, paths, args.output_root / "cache")
|
||||
config_knobs["enable_prefix_caching"] = args.prefix_caching
|
||||
config_knobs["prediction_max_tokens_per_request"] = 40960
|
||||
for trace in traces:
|
||||
run_dir = args.output_root / "runs" / config.name / trace["label"]
|
||||
result_path = run_dir / "result.json"
|
||||
|
||||
Reference in New Issue
Block a user