About
I love to work with Blender, Python, C++, and I am excited about new AI technology.
AI is about to change our world, and it’s up to us to shape it and give it direction. It’s my profound belief that the future of AI should be open, transparent, and accessible to all.
My current passion lies in the fascinating world of AI, particularly using Llama.cpp, where I’m committed to provide powerful quantized models on HuggingFace accessible to all.
However, I can’t do it alone. I’m well aware that this mission comes with its challenges, and my resources are limited. Your support can make a significant difference. Soon, I’ll be setting up a Patreon or a similar platform to help fuel this endeavor. Please consider becoming a supporter if you’re willing to contribute.
Please note that my focus is exclusively on quantizing truly free models, such as Falcon up to 40b, MPT, and others. This means of course:
No Llamas. No Llama 2. And, sadly, also no Falcon 180b. Llama models are well-supported elsewhere.
Blog
-
theBloke's Impact: Quantized Models and Accessibility
For months, theBloke has been diligently quantizing models, including both new and finely-tuned variant -
The Unveiling of Llama 2: A Paradigm Shift in AI Models
In the ever-evolving realm of artificial intelligence, there are moments that send ripples through the community—moments that redefine the boundaries of what's possible and challenge our understanding of open source principles. The release of Llama 2 in July 2023 was one such moment, an unexpected leap into uncharted territory. -
Ah, and you think the images on my site are somewhat - random ??
You may have noticed that the images often appear to be random or loosely related to the content. This is how and why: -
Adventures with ZFS
I have a ZFS pool that was originally equipped with 4 disks of 3 TB each. After I had replaced these successively with 4 TB disks, I had expected to be able to use the increased capacity - the existing documentation on ZFS suggests exactly that.