gliched_robot

33 ms·

1.
by gliched_robot 2y ago
If anyone wants to try this out, here is the hugging-face spaces demo: https://huggingface.co/spaces/google/paligemma
2.
by gliched_robot 2y ago
This is far more superior than SORA, there is no comparison.
3.
by gliched_robot 2y ago
Lmsys devs have all the answers, I am not sure how this has not leaked yet. They must a strong NDAs.
4.
by gliched_robot 2y ago
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.
5.
by gliched_robot 2y ago
GPU server locations, maybe?
6.
by gliched_robot 2y ago
Inference speed is not a great metric given the horizontal scalability of LLMs.
7.
by gliched_robot 2y ago
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.
8.
by gliched_robot 2y ago
Maybe a typo?
9.
by gliched_robot 2y ago
This llama model some made it run on an iphone. https://x.com/1littlecoder/status/1781076849335861637?s=46
10.
by gliched_robot 2y ago
I see what you did here <q> carrying the "torch" <q>. LOL
11.
by gliched_robot 2y ago
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.
12.
by gliched_robot 2y ago
Wild considering, GPT-4 is 1.8T.
13.
by gliched_robot 2y ago
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
14.
by gliched_robot 2y ago
Is this real? Seems suspicious given that its just one model not a family of models like llama and llama-2.
15.
by gliched_robot 3y ago
This is very cool and will change the way we do lora now.