Small language models on phones are getting good

Ren Suzuki · 13 likes

Prompt sensitivity is still the biggest pain in production

Tola Ibrahim · 12 likes

Haha my thoughts exactly

Budi Santoso · 0 likes

Depends but yeah mostly

Lucas Silva · 6 likes

Benchmarks look great but real world usage is where it falls apart for me

Ethan Brooks · 0 likes

Good take, hadnt considered that

Tanvir Rahman · 2 likes

Open weights or it didnt happen

Olena Shevchenko · 0 likes

lol this made me laugh

Arif Hasan · 1 likes

The cost per token matters more than the leaderboard imo

Amelia Hughes · 14 likes

Good take, hadnt considered that

Mehdi Benali · 2 likes

Fair point

Jasmine Cruz · 0 likes

Running a quantized version locally and its surprisingly usable

Mango Rider · 0 likes

Good take, hadnt considered that

Kabir Khan · 0 likes

Tried it on my own dataset, results were mixed

Siti Lestari · 3 likes

Fair point

Camila Rocha · 1 likes

I read that twice and still chuckled

Jonas Schneider · 0 likes

Context window claims are always optimistic, test it yourself

Mango Noodle · 0 likes

Fair point

Jisoo Park · 0 likes

True, learned it the hard way