AI
GPTs are amazingly useful and practical, and this section will be a log of my explorations and learning.
Demos
Test Drives
Quick explorations of GenAI tools and technologies.
- Dec 2024 - Test Drive: transformers.js
- Nov 2024 - Test Drive: Local Model Servers
- Nov 2024 - Test Drive: Local Logprobs
- Nov 2024 - Test Drive: text to SQL
- Nov 2024 - Test Drive: Hybrid Full-text Search
- Oct 2024 - Test Drive: Transformers
- Oct 2024 - Test Drive: tokenizers
- Oct 2024 - Test Drive: llama-index
- Oct 2024 - Test Drive: Open WebUI
- Oct 2024 - Test Drive: ditto
- Oct 2024 - Test Drive: Open Canvas
- Oct 2024 - Test Drive: Gemma APS
- Oct 2024 - Test Drive: fastdata
- Oct 2024 - Test Drive: Nemotron
- Oct 2024 - Test Drive: F5-TTS
- Oct 2024 - Test Drive: MLX Finetuning
- Oct 2024 - Test Drive: Swarm
- Oct 2024 - Test Drive: Video Scraping
- Oct 2024 - Test Drive: MusicGen via MLX
- Oct 2024 - Test Drive: StabilityAI search and replace
- Oct 2024 - Test Drive: llamafile
- Oct 2024 - Test Drive: e2tts with MLX
- Oct 2024 - Test Drive: OpenAI Evaluations
- Oct 2024 - Test Drive: OpenRouter Chat
- Oct 2024 - Test Drive: crewAI
- Oct 2024 - Test Drive: controlflow
- Oct 2024 - Test Drive: Deepgram TTS
- Oct 2024 - Test Drive: Whisper v3 Turbo
- Oct 2024 - Test Drive: Outlines
- Sep 2024 - Test Drive: ell
- Sep 2024 - Test Drive: Sambanova
- Sep 2024 - Test Drive: Cursor Composer
- Sep 2024 - Test Drive: opik
- Sep 2024 - Test Drive: Langtrace
- Sep 2024 - Test Drive: Qwen2.5 + Openrouter + LiteLLM
- Sep 2024 - Test Drive: logprobs
- Sep 2024 - Test Drive: Instructor
- Sep 2024 - Test Drive: litellm
- Sep 2024 - Test Drive: sqlite-vec
Test Drive Backlog
GenAI tools and technologies that I aspire to explore.
- ControlFlow: Python framework for building agentic AI workflows.
- Crewai: framework for orchestrating role-playing, autonomous AI agents.
- dspy: Programming—not prompting—Language Models
- phoenix-arize:
- https://llm.mlc.ai/docs/get_started/quick_start
- Other stuff: transformers, multimodal stuff, prompting techniques, the anthropic prompt course, evals, structured data, openrouter, prompt caching