Skip to content

feat: reproducible train/eval split for post-training exports (#103) #112

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
dcfocus merged 1 commit into from
Jun 27, 2026
Merged

feat: reproducible train/eval split for post-training exports (#103) #112

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 9 additions & 0 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -273,6 +273,15 @@ print("SFT examples:", manifest["counts"]["examples"])
ctx.export_training("training/pref.jsonl", task="preference", preference_form="paired")
ctx.export_training("training/rollout.jsonl", task="rollout") # GRPO/RLVR groups

# Reproducible, group-disjoint train/eval split (no session leaks across the
# boundary; same seed reproduces the partition). Writes sft.train.jsonl +
# sft.eval.jsonl, each with its own manifest.
ctx.export_training(
"training/sft.jsonl",
task="sft",
split={"eval_fraction": 0.1, "by": "session_id", "seed": 42},
)

# Remote persistence on any object_store backend uses a generic `storage_options`
# dict, matching the conventions used by `lance` and `lance-graph`.
#
Expand Down
Loading
Loading