lars/giant
1
0
Fork 0
Code Issues 4 Pull Requests 2 Actions Packages Projects Releases Wiki Activity

Pass --seed through to the train/val event split
CI / Lint (ruff check) (push) Successful in 27s
Details
CI / Format (ruff format) (push) Successful in 28s
Details
CI / Type check (ty) (push) Successful in 39s
Details
CI / Sync project version with tag (push) Has been skipped
Details
CI / Lint (ruff check) (pull_request) Successful in 41s
Details
CI / Format (ruff format) (pull_request) Successful in 39s
Details
CI / Sync project version with tag (pull_request) Has been skipped
Details
CI / Type check (ty) (pull_request) Successful in 37s
Details
CI / Tests (push) Successful in 1m56s
Details
CI / Tests (pull_request) Successful in 1m27s
Details

Browse Source 
make_event_split() defaults to seed=42, and run_train_job was calling
it without forwarding t["seed"] — so the configured --seed affected
model init/EMA/etc. but not which events landed in train vs. val,
which silently always used seed 42 regardless of --seed.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This commit is contained in:
Lars Bogner
2026-07-30 10:29:19 +02:00
parent e7478c36fb
commit 26aa9d3fde
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
giant/pipeline.py
+1 -1
View File
@@ -70,7 +70,7 @@ def run_train_job(
cache.event_index = (unique_ids, counts)
train_events, val_events = make_event_split(
unique_ids, val_fraction=t["val_fraction"]
unique_ids, val_fraction=t["val_fraction"], seed=t["seed"]
)
events_arr = np.array(sorted(train_events))
n_train_steps = setup_cache.n_train_steps_for_split(unique_ids, counts, events_arr)