- Datasets (images, text, code, etc.)
- Training scripts (optional but recommended)
- QOR tokens for staking (100-2500 QOR depending on tier)
- Quality bond (automatically calculated)
- Upload Resources - Drag & drop your datasets
- Set Duration - Choose 5 minutes to 6 hours
- Preview Rewards - See estimated QOR earnings
- Stake Tokens - Select your tier (Bronze to Platinum)
- Post Task - Get paid while AI trains on your data
| Tier | Stake Required | Max Task Size | Concurrent Tasks | Reward Multiplier | Bond Reduction |
|---|---|---|---|---|---|
| Bronze | 100 QOR | 30 minutes | 1 | 0.8x (36% rewards) | 0% |
| Silver | 500 QOR | 2 hours | 3 | 1.0x (45% rewards) | 20% |
| Gold | 1,000 QOR | 6 hours | 5 | 1.2x (54% rewards) | 40% |
| Platinum | 2,500 QOR | Unlimited | 10 | 1.5x (67.5% rewards) | 60% |
Gold (1,000 QOR) - Best balance of investment vs rewards. Pays for itself in 1-2 successful tasks.
- Total Reward Pool: 90,000 QOR (1800 × 50 QOR per block)
- Your Share: 45% base rate
| Your Tier | Your Earnings | ROI per Task |
|---|---|---|
| Bronze | 32,400 QOR | 324x stake |
| Silver | 40,500 QOR | 81x stake |
| Gold | 48,600 QOR | 49x stake |
| Platinum | 60,750 QOR | 24x stake |
Note: Higher tiers have better absolute earnings despite lower ROI percentages.
✅ Clean Your Data
- Remove corrupted files
- Ensure consistent formatting
- Label data clearly
- Include metadata/descriptions
✅ Optimize File Sizes
- Compress large datasets appropriately
- Use efficient formats (PNG vs BMP, CSV vs Excel)
- Remove unnecessary duplicates
✅ Provide Context
- Describe what the data contains
- Explain intended use case
- Set realistic expectations
- Include any preprocessing steps
✅ Quality Scripts
- Write efficient, well-commented code
- Test scripts before uploading
- Include documentation
- Optimize for GPU training
✅ Set Realistic Timelines
- Small datasets: 5-30 minutes
- Medium datasets: 30 minutes - 2 hours
- Large datasets: 2-6 hours
- Massive datasets: 6+ hours (Platinum only)
✅ Build Reputation
- Start with smaller, high-quality tasks
- Respond to feedback quickly
- Fix issues promptly
- Engage with the community
❌ Don't Upload Corrupted Data
- Broken files waste GPU time
- Results in automatic task failure
- Damages your reputation score
- Leads to bond forfeiture
❌ Don't Use Copyrighted Material
- Violates platform terms
- Can result in legal issues
- Leads to account suspension
- Causes 100% stake slashing
❌ Don't Post Duplicate Tasks
- Same data uploaded multiple times
- Wastes network resources
- Reduces rewards for everyone
- Triggers anti-spam measures
❌ Don't Misrepresent Data
- Claiming fake dataset sizes
- Promising unrealistic results
- Mislabeling data types
- Leads to reputation penalties
❌ Don't Overstake Initially
- Start with Bronze/Silver tier
- Learn the system first
- Build reputation gradually
- Upgrade tier after success
❌ Don't Post Tasks Too Large for Your Tier
- Bronze: Max 30 minutes
- Silver: Max 2 hours
- Gold: Max 6 hours
- Exceeding limits = automatic rejection
❌ Don't Rush Quality Bonds
- Understand bond requirements
- Ensure sufficient QOR balance
- Calculate total cost upfront
- Don't risk more than you can lose
Good Example:
- Dataset: 10,000 cat photos (clean, labeled)
- Script: Fine-tune ResNet for cat breed classification
- Duration: 2 hours
- Expected output: 95%+ accuracy model
Bad Example:
- Dataset: Random internet images (unlabeled, mixed quality)
- Script: Generic training script
- Duration: 6 hours
- Expected output: "Some kind of image classifier"
Good Example:
- Dataset: 50,000 customer reviews (preprocessed, sentiment labeled)
- Script: Fine-tune BERT for sentiment analysis
- Duration: 1 hour
- Expected output: Production-ready sentiment model
Bad Example:
- Dataset: Raw web scraping data (unprocessed, duplicates)
- Script: Basic text classifier
- Duration: 4 hours
- Expected output: "Text analysis thing"
Good Example:
- Dataset: 100,000 Python functions (documented, tested)
- Script: Fine-tune CodeBERT for code completion
- Duration: 3 hours
- Expected output: VS Code extension ready model
Bad Example:
- Dataset: Random GitHub files (uncommented, mixed languages)
- Script: Generic code model
- Duration: 5 hours
- Expected output: "Code understanding model"
- Bond Amount: 10% of expected total rewards
- Purpose: Ensures task quality and completion
- Reduction: Higher tiers get lower bonds
- Return: Full bond returned on successful completion
1 Hour Task (90,000 QOR total rewards):
- Bronze: 9,000 QOR bond (10% of 90,000)
- Silver: 7,200 QOR bond (20% reduction)
- Gold: 5,400 QOR bond (40% reduction)
- Platinum: 3,600 QOR bond (60% reduction)
- Task Failure: Training doesn't complete successfully
- Poor Quality: Data doesn't meet minimum standards
- Fraud: Fake or misleading information
- Copyright: Using protected material without permission
- Starting Score: 100 points
- Score Range: 0-1000+ points
- Impact: Affects rewards, queue priority, bond requirements
- Successful Task: +10 points
- High Quality Data: +15 points
- Innovative Scripts: +20 points
- Community Feedback: +5 points
- Model Reuse: +3 points per reuse
- Failed Task: -20 points
- Poor Quality: -15 points
- Deadline Miss: -10 points
- Copyright Issue: -100 points
- Fraud Attempt: -500 points + ban
| Score Range | Benefits |
|---|---|
| 0-99 | Reduced rewards, higher bonds |
| 100-199 | Standard treatment |
| 200-299 | 10% reward bonus |
| 300-499 | 20% reward bonus, priority queue |
| 500+ | 30% reward bonus, beta features |
Images: PNG, JPG, TIFF, BMP, WebP Text: TXT, CSV, JSON, XML, MD Code: PY, JS, CPP, JAVA, GO, RS Models: PTH, ONNX, H5, PB Archives: ZIP, TAR, 7Z (auto-extracted)
- Single File: Max 5GB
- Total Dataset: Max 100GB per task
- Archive Files: Max 50GB compressed
- Languages: Python (preferred), R, Julia
- Dependencies: Include requirements.txt
- GPU Support: CUDA-compatible code recommended
- Memory: Optimize for 16GB+ GPU memory
Q: My task failed during training A: Check your data quality, script errors, and file formats. Failed tasks forfeit quality bonds.
Q: I can't post large tasks A: Upgrade your staking tier. Bronze is limited to 30 minutes, Silver to 2 hours.
Q: My reputation score dropped A: Review your recent tasks for quality issues. Focus on smaller, high-quality tasks to rebuild.
Q: Bond amount seems too high A: Bonds are 10% of expected rewards. Higher staking tiers get reduced bond requirements.
Q: Task is taking longer than expected A: GPU availability varies. Platinum tier gets priority queue access.
- Discord: #task-poster-help
- Documentation: docs.qoranet.ai
- Support Tickets: support@qoranet.ai
- Community Forum: forum.qoranet.ai
Q: Do I pay anything to post tasks? A: No upfront payment! You stake QOR (refundable) and post a quality bond (returned on success). You actually EARN money for posting good tasks.
Q: Who owns the trained models? A: All models become public goods on IPFS. You retain rights to your original data but trained models are open source.
Q: Can I train private/confidential data? A: Data is encrypted during training. Trainers cannot access raw data, only the training process.
Q: How long until I see results? A: Depends on network congestion and task size. Platinum tier gets priority processing.
Q: How much can I earn? A: Varies by tier and task size. Gold tier earns ~50,000 QOR for 1-hour tasks. At launch, this could be $500-5000 depending on token price.
Q: When do I get paid? A: 50% immediately, 50% vested over 30 days to prevent market dumping.
Q: What if the token price drops? A: You're earning a deflationary asset. With 50% of revenues burned, token should appreciate over time.
Q: Can I lose money? A: Only your quality bond is at risk. Stakes are returned when you stop posting. Choose appropriate tier sizes.
Q: What happens to my data? A: Encrypted, used for training only, then deleted. Original data remains yours.
Q: Can I update tasks after posting? A: No. Plan carefully before posting. Failed tasks forfeit bonds.
Q: Do I need to write code? A: Not required but recommended. Data-only tasks get lower reward multipliers.
Q: What GPUs will train my models? A: Network uses consumer and professional GPUs. Tasks auto-matched to appropriate hardware.
- Background: Needed AI for NPC behavior
- Task: 2-hour training on player interaction data
- Tier: Gold (1,000 QOR stake)
- Result: Earned 48,600 QOR + got perfect NPC AI for free
- ROI: 4,860% return on stake
- Background: Academic research on medical imaging
- Task: 4-hour training on MRI scan dataset
- Tier: Platinum (2,500 QOR stake)
- Result: Earned 121,500 QOR + published paper using model
- ROI: 4,860% return + academic publication
- Background: Wanted AI to generate art in personal style
- Task: 30-minute training on artwork portfolio
- Tier: Silver (500 QOR stake)
- Result: Earned 20,250 QOR + personal AI art generator
- ROI: 4,050% return + custom creative tool
- Start Small: Build reputation with quick, high-quality tasks
- Upgrade Gradually: Move from Bronze → Gold → Platinum
- Batch Similar Tasks: Post related datasets together
- Time Strategically: Post during low-congestion periods
- Community Engagement: Share insights, help others
- Diversify Task Sizes: Mix small and large tasks
- Quality First: Better to post fewer, higher-quality tasks
- Monitor Reputation: Keep score above 200 for bonuses
- Conservative Bonding: Don't risk more than you can afford
- Stay Updated: Follow platform updates and best practices
Platform Links
- Website: https://qoranet.ai
- Documentation: https://docs.qoranet.ai
- Discord: https://discord.gg/qoranet
- Twitter: https://twitter.com/qoranet_ai
Support Channels
- General Help: #general-help
- Technical Issues: #technical-support
- Task Posting: #task-poster-help
- Economics Questions: #tokenomics-help
Emergency Contact
- Critical Issues: emergency@qoranet.ai
- Security Issues: security@qoranet.ai
- Legal Issues: legal@qoranet.ai
Last Updated: Launch Version 1.0 This guide will be updated as the platform evolves