I'm trying to get familiar with Cog, so this repo serves as a backup of my learning journey.

• Build a simple GPT inference pipeline.
• Build a streaming GPT pipeline.
• Implement a Stable Diffusion pipeline.
  • Add tracing and model warmup.

Basic inference:

cd nlp/basic
cog predict -i prompt="Hello!" -i max_length=100

Streaming inference:

cd nlp/streaming
cog predict -i prompt="Count form one to twenty: One," -i max_length=100