How to run Google Gemma 2B- and 7B-parameter instruct models locally on the CPU and the GPU on Apple Silicon Macs.

See transcript ›

Here are some of the commands we used during the Creative Machine Learning Live 97.

First, create an Anaconda environment or install in your Python install with pip.

pip install imaginairy

Before running the commands below, I entered an interactive imaginAIry shell.

aimg
🤖🧠> # Commands here

# Upscale an image 4x with Real-ESRGAN.
upscale image.jpg

# Generate an image and animate the diffusion process.
imagine "a sunflower" --gif

# Generate an image and create a GIF comparing it with the original.
imagine "a sunflower" --compare-gif

# Schedule argument values.
edit input.jpg \
    --prompt "a sunflower" \
    --steps 21 \
    --arg-schedule "prompt_strength[6:8:0.5]" \
    --compilation-anim gif

The Google Research team has published a paper for MusicLM, a machine learning model that generates high-fidelity music from text prompts, and it works extremely well. But they won't release it to the public, at least not yet.

You can browse and play through the examples to listen to results obtained by the research team for a wide variety of text-to-music tasks, including audio generation from rich captions, long generation, story mode, text and melody conditioning, painting caption conditioning, 10s audio generation from text, and generation diversity,

I'm particularly surprised by the text and melody conditioning examples, where a text prompt—say, "piano solo," "string quarter," or "tribal drums"—can be combined with a melody prompt—say "bella ciao - humming"—generating accurate results.

Even when they don't release the model, Google Research has publicly released MusicCaps to support future research, "a dataset composed of 5.5k music-text pairs, with rich text descriptions provided by human experts."

How to encode an image dataset to reduce its dimensionality and visualize it in the 2D space.

See transcript ›

How to use TensorFlow inside of a Docker container.

See transcript ›

Yesterday, a colleague of mine shared Toras, the Toronto Annotation Suite, which reminded me of Microsoft's VoTT. But this one has AI-assisted labeling, ala Photoshop smart selection.

Here's a video in which I test if OpenAI's DALL-E can generate usable texture maps from an uploaded image.

This texture comes with one of Apple's project examples and the idea of generating textures with DALL-E came from Adam Watters on Discord.