Ask a question below.
Published answers will appear here.
i also tried to correlate per layer metrics from the public wandb with per layer Jlens statistics but without luck, i'd love to see how we can use this kind of technics to have more information in pre training!
(anonyme feedback if you prefer here https://www.admonymous.co/eliebakouch)
i also tried to correlate per layer metrics from the public wandb with per layer Jlens statistics but without luck, i'd love to see how we can use this kind of technics to have more information in pre training!
(anonyme feedback if you prefer here https://www.admonymous.co/eliebakouch)
Ask a question below.
Published answers will appear here.