DeepSeek & Open-Source LLMs

Run and fine-tune open models locally or in the cloud, and choose the right one for the job.

Experte 6 Module ca. 8 Std. Kostenlos
Lernpfad starten
  1. 1 The Open-Weight Model Landscape What open-weight models are, the main families, licences and when to choose them over APIs. 25 Min.
  2. 2 Running Models Locally Ollama and llama.cpp, quantised GGUF models, hardware tips and local chat. 30 Min.
  3. 3 Serving Models in Production vLLM and other servers, OpenAI-compatible endpoints, GPUs in the cloud and scaling. 35 Min.
  4. 4 Choosing and Evaluating Models Benchmarks vs your own evals, comparing open and hosted models, and total cost of ownership. 30 Min.
  5. 5 Fine-Tuning and RAG When to fine-tune, LoRA basics, and retrieval-augmented generation for your own data. 35 Min.
  6. 6 Responsible Use, Safety and Maintenance Safety layers for self-hosted models, data governance, licensing and keeping deployments current. 25 Min.