Good demonstration for WebGPU
#1
by GuestWho - opened
My thoughts on WebGPU were about something else—specifically, the execution and complexities of a small program. Certainly not about downloading 4 GB from the internet. Especially since you can download an LLM once and run it continuously. This will clearly come in handy in the gaming industry.
Moreover, the WebGPU technology itself isn't tied to this specific GGUF model at all... why not just add a field for the model URL—for anyone who actually wants to keep downloading it from the internet? I don't get it. Looking at your other models, it seems that you just like to make landing pages for each model)