mcloh/decision-models-benchmark
1
0
Fork 0
mirror of https://github.com/mcloh/decision-models-benchmark.git synced 2026-11-05 22:04:12 +00:00
Code Issues Packages Projects Releases Wiki Activity

update config and docs

Browse Source
This commit is contained in:
Marcos Lohmann
2026-10-07 14:47:13 -03:00
parent 2af9dae193
commit 344d1582fe
6 changed files with 366 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
infra/ds/gpu_queue.py
View File

@@ -21,6 +21,7 @@ def main():
ap.add_argument("--after")
ap.add_argument("--shape", default="VM.GPU.A10.1")
ap.add_argument("--storage", type=int, default=100)
ap.add_argument("--env", nargs="*", default=[], help="variáveis extras K=V (ex.: DMB_SPLIT=dev DMB_LIMIT=50)")
args = ap.parse_args()
LOG.parent.mkdir(exist_ok=True)
if args.after:
@@ -29,7 +30,8 @@ def main():
while True:
try:
run_id = jobs.run_job(f"{args.mode}-{cand}-gpu", "infra/ds/run_job.sh", args.shape, None, None,
False, {"DMB_CANDIDATE": cand, "DMB_MODE": args.mode, "DMB_DEVICE": "auto"},
False, {"DMB_CANDIDATE": cand, "DMB_MODE": args.mode, "DMB_DEVICE": "auto",
**dict(kv.split("=", 1) for kv in args.env)},
storage_gb=args.storage, region=provision.GPU_REGION)
break
except jobs.oci.exceptions.ServiceError as error: