ML Model Training

Pipeline

1. Define task and success metrics
2. Collect/prepare training data
3. Choose base model (by size and task fit)
4. Select GPU and estimate cost
5. Configure training (SFT, DPO, or GRPO)
6. Monitor training run
7. Evaluate on test set
8. Deploy model

Model Selection

• Small tasks (classification): Qwen 0.6B-3B
• Medium tasks (generation): Mistral 7B, Llama 8B
• Complex tasks (reasoning): Qwen 27B, Llama 70B