Even Facebook is more open than OpenAI. They've released, under somewhat open licenses, models like Galactica and opt-175b, which is of similar size to GPT-3, though maybe not as good.

Here's a guide to running BLOOM, another 175 billion parameter model, on your local computer, just using the CPU, maybe something similar would also work for Facebooks models. https://towardsdatascience.com/run-bloom-the-largest-open-ac...

With this you can expect it to take around three minutes for each word/token the model outputs.

For the same type on task: gtp-j, gpt-neox for exemple. Here a nice list of model you can use: https://huggingface.co/docs/transformers/index#supported-mod...

Stable diffusion is a bigger deal right now than DALLE. SD is made by Stability AI, a company with a name that doesn’t proclaim to be “open” and yet their model is open to everyone.

Stable diffusion is in the space of generative models.

Sadly there are none.