# DispatchR Training Launchers ## HF Hub Jobs (Recommended) Serverless GPU training — no Spaces needed. ```bash # 1. Install HF CLI curl -LsSf https://hf.co/cli/install.sh | bash # 2. Login hf auth login # 3. Launch training (L4 = $0.80/hr, 24GB) python scripts/launch_hf_job.py --flavor l4x1 --episodes 200 # Or A100 for 14B model ($2.50/hr, 80GB) python scripts/launch_hf_job.py \ --flavor a100-large \ --episodes 200 \ --model unsloth/Qwen3-14B-unsloth-bnb-4bit # 4. Watch logs hf jobs logs # 5. Check status hf jobs list ``` ### Hardware Options | Flavor | GPU | VRAM | $/hr | Best For | |--------|-----|------|------|----------| | `l4x1` | L4 | 24 GB | $0.80 | 4B model, budget runs | | `a100-large` | A100 | 80 GB | $2.50 | 14B model, max performance | ### Python Launcher Args ```bash python scripts/launch_hf_job.py \ --flavor l4x1 \ --episodes 200 \ --batch-size 8 \ --timeout 8h \ --model unsloth/Qwen3-4B-Thinking-2507-bnb-4bit \ --hub-model-id yourname/dispatchr-grpo ```