Arc Llama 1.1.1 just dropped.
Battlemage on Vulkan got way faster: +54% prompt processing, +40% generation at 16K on a B60.
Pinned runtime. A starter model that works the first time.
pip install -U arc-llama
https://t.co/02yugteKmB
In other bad news, it's been found that @bestbuy is selling uncensored laptops that could be used for cyber attacks. There are zero restrictions on these machines.
It did not take long for this prediction to come true
"We have to shut down Hugging Face because those models *could* be used to conduct cyber attacks"
Imagine a GPU that doesn't work unless it's connected to the internet
That is where this is going
Quick question, what do you think China can’t buy from other allies that we sell as a consumer product ? What is a GPU, what’s it used for ? Are you aware of what you’re mad about? Are you also aware of the people who were involved in the exposé? What they stand to gain? It’s ok to be mad when it’s something you’re knowledgeable on but “chips” comes off kinda surface level and unaware.
Because your bias for open ai is plain to see. Your audience is naturally captured the same way. T3 code is not a good sample of the average dev population using LLMs. You block all the replies in your comments that disagree or call them a stupid bitch and then proceed to block. I wonder why there’s crickets.
Arc Llama 1.1.1 just dropped.
Battlemage on Vulkan got way faster: +54% prompt processing, +40% generation at 16K on a B60.
Pinned runtime. A starter model that works the first time.
pip install -U arc-llama
https://t.co/02yugteKmB