nadlabs (@naddylabs) / X

nadlabs
@naddylabs
Sep 6
Tried qwopus 3.8, (#intel #arcb70, #llamacpp), I have to say, it's slower than vllm, but MAN is the quality and reduction in YAP impressive, the magic that made this is beyond me, but I've done ~300k output tokens and ~25M input tokens so far in #omp, loved it, only looped once🥹
Kyle Hessling
@KyleHessling1
Sep 4
Good morning, Y'ALL! Qwopus 3.8 27B Flash is out now! I feel like this one is the largest leap of added value we've ever committed in a fine-tune! It's very similar to the base model at xhigh, but is significantly more efficient in its CoT, sometimes by a factor of 10 or
nadlabs
@naddylabs
Aug 26
i finally understand why ppl hate @DrizzleORM and @prisma, but i also understand why ppl prefer to use them over not 🥹 oh my god
nadlabs
@naddylabs
Aug 25
@TablePlus THANK YOU FOR YOUR AMAZING APP, THANK YOU, THANK YOU, THANK YOU, THANK YOU!
nadlabs
@naddylabs
Aug 4
@BramptonTransit one of your drivers is texting at red lights, how do I report this horrendous act, this guys putting at least 30 people on this bus at risk …
1
nadlabs
@naddylabs
Jun 12
I wonder what would happen if ai had a left and right brain at an architectural level hmm