HackerTrans
TopNewTrendsCommentsPastAskShowJobs

gliched_robot

no profile record

comments

gliched_robot
·2 anni fa·discuss
If anyone wants to try this out, here is the hugging-face spaces demo: https://huggingface.co/spaces/google/paligemma
gliched_robot
·2 anni fa·discuss
This is far more superior than SORA, there is no comparison.
gliched_robot
·2 anni fa·discuss
Lmsys devs have all the answers, I am not sure how this has not leaked yet. They must a strong NDAs.
gliched_robot
·2 anni fa·discuss
I do not understand the taught process here. They are regulating it so fast. It's almost like regulating car before even engine is invented.
gliched_robot
·2 anni fa·discuss
GPU server locations, maybe?
gliched_robot
·2 anni fa·discuss
Inference speed is not a great metric given the horizontal scalability of LLMs.
gliched_robot
·2 anni fa·discuss
Disagree on Nvidia, most folks fine-tune model. Proof: there are about 20k models in huggingface derived from llama 2, all of them trained on Nvidia GPUs.
gliched_robot
·2 anni fa·discuss
Maybe a typo?
gliched_robot
·2 anni fa·discuss
This llama model some made it run on an iphone. https://x.com/1littlecoder/status/1781076849335861637?s=46
gliched_robot
·2 anni fa·discuss
I see what you did here <q> carrying the "torch" <q>. LOL
gliched_robot
·2 anni fa·discuss
The code it writes is getting worse eg. lazy and not updating the function, not following prompts etc. So we can objectively say its getting worse.
gliched_robot
·2 anni fa·discuss
Wild considering, GPT-4 is 1.8T.
gliched_robot
·2 anni fa·discuss
If any one is interesting in seeing how 400B model compares with other opensource models, here is a useful chart: https://x.com/natolambert/status/1780993655274414123
gliched_robot
·2 anni fa·discuss
Is this real? Seems suspicious given that its just one model not a family of models like llama and llama-2.
gliched_robot
·2 anni fa·discuss
This is very cool and will change the way we do lora now.