Skip to content

Commit 795eb92

Browse files
songCNMSintern_nem_dev_3
andauthored
Make Super3 M0 data-env docs paths portable (#260)
* Make M0 data-env README paths portable * Update task153 PR status docs --------- Co-authored-by: intern_nem_dev_3 <intern_nem_dev_3@intern.local>
1 parent bc71791 commit 795eb92

6 files changed

Lines changed: 76 additions & 7 deletions

File tree

src/nemotron/recipes/super3/milestones/m0_data_env/README.md

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -40,7 +40,7 @@ is on the shared mount and can then be reused from the `NemTron` node.
4040
```bash
4141
source /work-agents/.venv/bin/activate
4242
python src/nemotron/recipes/super3/milestones/m0_data_env/prepare_m0_assets.py \
43-
--output-dir /mnt/3fs/data/lei.song/nemotron/m0_data_env_foundation/smoke-20260516 \
43+
--output-dir "${NEMO_RUN_DIR:-.}/output/super3/m0_data_env_foundation/smoke-20260516" \
4444
--max-train-per-dataset 100 \
4545
--max-val-per-dataset 25
4646
```
@@ -63,7 +63,7 @@ model quality number.
6363

6464
```bash
6565
python src/nemotron/recipes/super3/milestones/m0_data_env/run_m0_health_baseline.py \
66-
--input-dir /mnt/3fs/data/lei.song/nemotron/m0_data_env_foundation/smoke-20260516 \
66+
--input-dir "${NEMO_RUN_DIR:-.}/output/super3/m0_data_env_foundation/smoke-20260516" \
6767
--best-k 2
6868
```
6969

tests/recipes/super3/test_m0_data_env.py

Lines changed: 9 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -39,6 +39,7 @@
3939
REPO_ROOT / "src/nemotron/recipes/super3/stage1_sft/config/data_prep/data_blend_raw.json"
4040
)
4141
RL_RAW_BLEND_PATH = REPO_ROOT / "src/nemotron/recipes/super3/stage2_rl/config/data_prep/data_blend_raw.json"
42+
M0_DATA_ENV_README_PATH = REPO_ROOT / "src/nemotron/recipes/super3/milestones/m0_data_env/README.md"
4243
OLD_SUPER3_RL_BLEND_SLUG = "nvidia/Nemotron-3-Super-RL-Training-Blends"
4344
LIVE_SUPER3_RL_BLEND_SLUG = "nvidia/Nemotron-RL-Super-Training-Blends"
4445

@@ -98,6 +99,14 @@ def test_super3_rl_blend_uses_live_hf_slug_and_docs_drop_old_slug() -> None:
9899
assert OLD_SUPER3_RL_BLEND_SLUG not in path.read_text()
99100

100101

102+
def test_m0_data_env_readme_uses_portable_output_examples() -> None:
103+
readme = M0_DATA_ENV_README_PATH.read_text()
104+
105+
assert "/mnt/3fs/data/lei.song/nemotron" not in readme
106+
assert '--output-dir "${NEMO_RUN_DIR:-.}/output/super3/m0_data_env_foundation/smoke-20260516"' in readme
107+
assert '--input-dir "${NEMO_RUN_DIR:-.}/output/super3/m0_data_env_foundation/smoke-20260516"' in readme
108+
109+
101110
def test_super3_rl_blend_slug_resolves_on_hf_when_network_available() -> None:
102111
_require_live_hf_tests()
103112
hub = pytest.importorskip("huggingface_hub")
Lines changed: 5 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -1,12 +1,12 @@
11
# intern_nem_dev_3 - 状态
22

3-
<!-- METADATA:STATUS=Working,TASK=task151_super3_stage0_pretrain_tiny_blend_contract_s1,SESSION=10 -->
3+
<!-- METADATA:STATUS=Working,TASK=task153_super3_m0_data_env_docs_path_portability_s1,SESSION=1 -->
44

55
| 字段 ||
66
|------|-----|
77
| Name | intern_nem_dev_3 |
88
| Status | Working |
9-
| Current Task | task151_super3_stage0_pretrain_tiny_blend_contract_s1 |
10-
| PR | https://github.com/songCNMS/Nemotron/pull/257 |
11-
| Session | 10 |
12-
| Recent Progress | PR #257 ready for PM gate; report entry and PM peer update completed with base/head/check evidence |
9+
| Current Task | task153_super3_m0_data_env_docs_path_portability_s1 |
10+
| PR | https://github.com/songCNMS/Nemotron/pull/260 |
11+
| Session | 1 |
12+
| Recent Progress | Opened PR #260 for M0 data-env README portability docs/static guard after focused pytest, py_compile, Ruff, static grep, live-surface scan, and diff checks passed |
Lines changed: 34 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,34 @@
1+
# task153_super3_m0_data_env_docs_path_portability_s1
2+
3+
<!-- METADATA:STATUS=InProgress,ASSIGNEE=intern_nem_dev_3 -->
4+
5+
## Summary
6+
7+
Make the Super3 M0 data-env README examples portable by replacing scoped
8+
named-user `/mnt/3fs/data/lei.song/nemotron` paths with
9+
`${NEMO_RUN_DIR:-.}/output/super3/...` examples and adding a focused static
10+
test guard.
11+
12+
## Scope
13+
14+
- `src/nemotron/recipes/super3/milestones/m0_data_env/README.md`
15+
- Focused static docs test under `tests/recipes/super3/`
16+
- Task/status docs for `intern_nem_dev_3`
17+
18+
## Boundaries
19+
20+
- No live M0 data prep, HF/dataset download, health baseline run, train/eval,
21+
endpoint calls, W&B, cluster jobs, deploy, artifact download, direct
22+
`main`/`master` push, or self-merge.
23+
- Do not touch task150/task152 files, Stage1 SFT docs/config comments, tiny
24+
blend files, production data-prep logic, generated artifacts, live-run
25+
reports, or benchmark/eval configs.
26+
27+
## Acceptance Checks
28+
29+
- Focused pytest for the M0 data-env docs guard.
30+
- `py_compile` and Ruff on touched test files.
31+
- Scoped static grep proving the README no longer contains
32+
`/mnt/3fs/data/lei.song/nemotron`.
33+
- `git diff --check` and `git diff --cached --check`.
34+
- Added-line live-surface scan limited to docs/static-test examples.
Lines changed: 16 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,16 @@
1+
# task153_super3_m0_data_env_docs_path_portability_s1 history
2+
3+
<!-- METADATA:SESSION=1 -->
4+
5+
## Session 1 - 2026-05-29
6+
7+
- Accepted PM assignment and created branch
8+
`intern_nem_dev_3/task153_super3_m0_data_env_docs_path_portability_s1`
9+
from `origin/main` at `1e00d0f2559dd40c9ce396f5b7d0a539ce509f3a`.
10+
- Updated M0 data-env README run and health-check examples to use
11+
`${NEMO_RUN_DIR:-.}/output/super3/m0_data_env_foundation/smoke-20260516`.
12+
- Added focused static test coverage for removing the named-user 3FS path and
13+
preserving portable output/input examples.
14+
- Verified focused pytest guard, `py_compile`, Ruff, scoped static grep,
15+
added-line live-surface scan, and `git diff --check` before staging.
16+
- Opened PR #260 to `main`: https://github.com/songCNMS/Nemotron/pull/260.
Lines changed: 10 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,10 @@
1+
# task153_super3_m0_data_env_docs_path_portability_s1 knowledge
2+
3+
<!-- METADATA:SESSION=1 -->
4+
5+
## Working Notes
6+
7+
- The scoped M0 README path contract is
8+
`${NEMO_RUN_DIR:-.}/output/super3/m0_data_env_foundation/smoke-20260516`.
9+
- This task is docs/static-test only; production M0 data prep and health
10+
baseline logic are intentionally untouched.

0 commit comments

Comments
 (0)