PWM Reference Cloud
Trusted reconstruction, benchmark-aligned validation, and official PWM reports
Community
Students & first-time users
- 3 reference runs / month
- Small datasets only
- Standard methods
- Basic output preview
- Public queue
- No validation reports
- "For research preview only" watermark
Researcher
PhD students & postdocs
- 20 reference runs / month
- Small–medium datasets
- Leaderboard-backed methods
- Downloadable reconstructions
- 5 PWM validation reports / month
- Run history & reproducibility logs
- Priority queue
Pro Lab
Active labs & startups
- 100 reference runs / month
- Large dataset support
- Batch jobs
- 25 PWM validation reports / month
- Method comparison mode
- Benchmark comparison tables
- Calibration & failure-diagnosis summary
- API access
- Exportable PDF / CSV / JSON reports
- Faster queue / priority support
Team
Lab groups & imaging startups
- 500 reference runs / month
- 5–10 seats
- Shared workspace & project folders
- 100 PWM validation reports / month
- Private datasets
- Role-based access
- API + integration support
- Higher compute priority
- Benchmark submission tools
- White-label report option
Enterprise
For hardware vendors, hospitals, and industrial imaging teams
$15,000 – $100,000+ / year — Private deployment, custom SLAs, unlimited reports
Contact SalesWeChat Pay Credit Packs (微信支付充值包)
One-time purchase — same credits as subscription tiers, pay with WeChat
Credits are valid for the specified period. No auto-renewal — buy when you need.
Add-on Credits
Need more runs or reports? Purchase additional credits without upgrading your plan.
Extra Reference Runs
| Small run | $3 |
| Medium run | $10 |
| Large run | $25–$100 |
Extra Validation Reports
| Standard PWM report | $25 |
| Advanced benchmark report | $75 |
| Enterprise signed report | $250+ |
Reference Reconstruction
Trusted runs of PWM's benchmark-standard pipeline
Validation Report
TNA score, method comparison, failure-mode hints, reproducibility ID
Decision Support
Is my system working? How do I compare to reference?
Workflow Convenience
No Docker, CUDA, cluster setup, or repo debugging