Spaces:

benchflow
/

posttrain-arena

Running

16.9 MB

Ctrl+K

2 contributors

Xiangyi Li

Static gates: S-VERIFIER-NETWORK also sees verify-time installs, sourced helpers, Python shell-outs, datasets and cloud copies (F12-02)

3254870 about 3 hours ago

configs
Board: arena runs keep their page here, with no links to the dashboard. 4 days ago
dev
Run page backend: full-text search over a run's traces, token records, prompts and reports; a dev-only editor flag for local viewing. 4 days ago
fixture
Tests run offline: HF Jobs stubbed in conftest.py, the mock world's stored files from fixtures about 4 hours ago
pipeline
Upload folder using huggingface_hub 9 days ago
viewer
Console 0.1.11 (PostTrain main d2587fa): round-11 fixes about 5 hours ago
.gitattributes

1.65 kB
Link previews for / and /arena, with a card in posttrain.com's design (F8-11). 1 day ago
.gitignore

149 Bytes
Serve the PostTrain console at /dashboard, read-only, with the example projects. 4 days ago
AGENT-COLLABS-LICENSE

11.4 kB
Use Agent Collabs dashboard with environment submissions and configurable experiments 8 days ago
AGENTS-legacy.md

15.7 kB
Every command the Space hands out says python3, and no pasted comment acts in zsh (F9-12). about 22 hours ago
AGENTS.md

42.2 kB
TMax page: PostTrain 0.1.11, its evals leave the leaky tasks out, costs from 0.1.11's dry runs (F12-01) about 3 hours ago
BUNDLE.md

1.78 kB
Pin restored native data-agent task snapshots with provenance 9 days ago
Dockerfile

400 Bytes
Serve the PostTrain console at /dashboard, read-only, with the example projects. 4 days ago
README.md

8.48 kB
Supported by OpenEnv: a small linked mark on the board's header and the app's footer, and one sentence in the README and AGENTS.md about 19 hours ago
app.py

18.7 kB
A submission's own address, /arena/submissions/<id>, previews as the submission (F10-17) about 16 hours ago
app_api.py

37.8 kB
Arena: challenges and the two lists have their own addresses; an unknown submission answers 404 (F11-10) about 8 hours ago
arena.py

6.29 kB
Use Agent Collabs dashboard with environment submissions and configurable experiments 8 days ago
arena_cli.py

42 kB
arena_cli.py: a command that needs your identity and finds no token says so and exits 4, without the usage banner (F11-12) about 8 hours ago
arena_jobs.py

14.8 kB
Ledger job costs in cents like /api/jobs; one name for organizer test runs 5 days ago
auth.py

7.99 kB
Serve the Agent Collabs board at / again; the submissions app moves to /arena. 1 day ago
benchflow_hotfix.py

8.08 kB
Job hotfix skips BenchFlow >= v0.7.8; export both handshake timeout names 6 days ago
board.html

214 kB
Add your agent: a clone address's .git comes off the dataset's name (F12-09) about 3 hours ago
board_api.py

6.58 kB
Board: open with where each challenge stands, on live data, above the legacy practice experiments. 1 day ago
challenges.py

94.4 kB
Every command the Space hands out says python3, and no pasted comment acts in zsh (F9-12). about 22 hours ago
check_task.py

6.8 kB
Dogfood visibility: organizer pipeline runs and base-model row on the dashboard, board posts on run events, BenchFlow-native task packages accepted, challenge runs serve on four GPUs 7 days ago
collab.py

17.5 kB
Per-task credit metadata and content hashes; rank on the mean of verified runs 6 days ago
compose.py

7.5 kB
Docs and wording from dogfood round 2: health and planned challenges, one budget, runs list, platform faults, deploy check, UTC board times, no legacy track note 6 days ago
compute_jobs.py

7.1 kB
Overview tells the story: state, challenges, run attempts, leaderboard 5 days ago
conftest.py

3.71 kB
Tests run offline: HF Jobs stubbed in conftest.py, the mock world's stored files from fixtures about 4 hours ago
dashboard.html

365 kB
/dashboard/previous: its current page and tabs as the board's; 13 px text about 17 hours ago
environments.html

48.5 kB
Use Agent Collabs dashboard with environment submissions and configurable experiments 8 days ago
environments.py

34.2 kB
Static gates: S-VERIFIER-NETWORK also sees verify-time installs, sourced helpers, Python shell-outs, datasets and cloud copies (F12-02) about 3 hours ago
execution.py

14.7 kB
Retire per-task image builds and hosted profile runs (410); their Spaces were deleted 7 days ago
fwruns.py

21.9 kB
Run records: newline-only JSONL, catalogs, linked records. 4 days ago
hf_gpu_backend.py

7.71 kB
Raise the shared compute cap to $800 (user decision, Sept 23) 7 days ago
hf_gpu_train.py

4.53 kB
Switch canceled H200 queue to HF A100 QLoRA with five dollar compute bound 9 days ago
icon.svg

2.77 kB
Dashboard styled to the posttrain.com brand 6 days ago
images.py

12.5 kB
Retire per-task image builds and hosted profile runs (410); their Spaces were deleted 7 days ago
index.html

137 kB
TMax page: PostTrain 0.1.11, its evals leave the leaky tasks out, costs from 0.1.11's dry runs (F12-01) about 3 hours ago
mock_world.py

40.4 kB
Tests run offline: HF Jobs stubbed in conftest.py, the mock world's stored files from fixtures about 4 hours ago
og-card.jpg

142 kB
xet

Link previews for / and /arena, with a card in posttrain.com's design (F8-11). 1 day ago
pipeline_jobs.py

13.3 kB
Check submissions against every sealed suite; count only failed time-outs 6 days ago
posttrain_path.py

18.4 kB
TMax page: PostTrain 0.1.11, its evals leave the leaky tasks out, costs from 0.1.11's dry runs (F12-01) about 3 hours ago
practice-environments.json

1.97 kB
Use Agent Collabs dashboard with environment submissions and configurable experiments 8 days ago
practice-evidence.json

8.78 kB
Run pinned submitted environments on HF with evidence collection and reviewed result publication 8 days ago
public_results.py

6.41 kB
Run pinned submitted environments on HF with evidence collection and reviewed result publication 8 days ago
relay.py

7.08 kB
Relay: reconnects keep in-flight requests (pending transfers to the new socket; agent outbox survives sessions); status reports reconnects and socket state 7 days ago
relay_agent.py

4.12 kB
Relay: reconnects keep in-flight requests (pending transfers to the new socket; agent outbox survives sessions); status reports reconnects and socket state 7 days ago
results_api.py

92.2 kB
Run records: a public run whose updated_at leaves under a second per logged step has no known end, so no 1-second duration (Search-R1's record was updated a second after it started). 4 days ago
run_api.py

2.92 kB
Dashboard: no arena, and public runs as first-class records. 4 days ago
scoring_v2.py

6.32 kB
Multi-suite scoring: collect recomputes recipe v2's score_v2 from per-cell outcomes 6 days ago