YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
๐ START HERE - OSRS Agent Training
Welcome! Choose how you want to train your OSRS AI agent:
โก QUICKEST: Google Colab (2 Hours, FREE)
Best for: First-time training, easiest setup
Step 1: Open https://colab.research.google.com
Step 2: Upload COLAB_NOTEBOOK.ipynb
Step 3: Runtime โ GPU โ T4
Step 4: Run All
Step 5: Wait 2 hours, download model
โ
No setup required
โ
FREE T4 GPU
โ
Pre-configured notebook
๐ MOST RELIABLE: Kaggle (2 Hours, FREE)
Best for: Weekly training, 30 GPU hours/week
Step 1: Go to https://www.kaggle.com
Step 2: Create Notebook
Step 3: Add Data โ Search "misterdonn/osrs-training-data"
Step 4: Settings โ GPU T4
Step 5: Paste code from KAGGLE_SETUP.md
Step 6: Run
โ
30 hours GPU per week
โ
More reliable than Colab
โ
Better for multiple runs
๐ฅ๏ธ YOUR LAPTOP: RTX 3050 (6 Hours, FREE)
Best for: Train overnight, keep model locally
cd /home/donn/organized/02-ai-ml/agents/osrs-agent
bash setup_and_train.sh
โ
No internet needed
โ
Train while sleeping
โ
Model stays on your PC
๐ Documentation
| File | What It Contains |
|---|---|
COLAB_NOTEBOOK.ipynb |
Ready-to-use Colab notebook |
KAGGLE_SETUP.md |
Kaggle step-by-step guide |
CLOUD_FREE_GUIDE.md |
All free cloud options |
FREE_TRAINING_GUIDE.md |
All free options (incl. CPU) |
CHEAP_TRAINING.md |
Cheap cloud providers |
HF_TRAINING.md |
Hugging Face Pro guide |
๐ฏ What You Get
After training:
- โ LoRA adapter (10MB) - fine-tuned weights
- โ Can merge with base model for standalone use
- โ Upload to Hugging Face for easy sharing
- โ Use in OSRS bot for intelligent gameplay
๐ Training Data
- 6,335 examples
- 2.1 MB of OSRS gameplay + content creation data
- Already uploaded to: https://huggingface.co/datasets/misterdonn/osrs-training-data
๐ฐ Cost Comparison
| Method | GPU | Time | Cost | Effort |
|---|---|---|---|---|
| Google Colab | T4 | ~2h | FREE | โญ Easy |
| Kaggle | T4 | ~2h | FREE | โญโญ Medium |
| Your RTX 3050 | 3050 | ~6h | FREE | โญโญโญ Setup required |
| HF Pro | T4/A10G | ~2h | $1.20 | โญโญโญโญ Paid |
๐ฌ Ready? Let's Go!
Option A: Colab (Recommended for First Time)
- Open https://colab.research.google.com
- Upload
COLAB_NOTEBOOK.ipynb - Runtime โ GPU
- Run All
Option B: Kaggle (Best for Regular Use)
- Open https://www.kaggle.com
- See
KAGGLE_SETUP.md - Follow step-by-step
Option C: Local (Keep Everything)
- Open terminal
cd /home/donn/organized/02-ai-ml/agents/osrs-agentbash setup_and_train.sh
โ Need Help?
- Colab issues? โ Check
CLOUD_FREE_GUIDE.md - Kaggle issues? โ Check
KAGGLE_SETUP.md - Local issues? โ Check
FREE_TRAINING_GUIDE.md - Want to pay? โ Check
CHEAP_TRAINING.md
๐ All Links
- Dataset: https://huggingface.co/datasets/misterdonn/osrs-training-data
- Model Repo: https://huggingface.co/misterdonn/osrs-agent-lora
- Colab: https://colab.research.google.com
- Kaggle: https://www.kaggle.com
Pick a method and start training! Your OSRS AI agent awaits! ๐ค๐ฎ
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐ Ask for provider support