Training the LLaMA Model from Scratch | by Ebad Sayed | Medium