oo_viper_oo

joined 1 year ago
[โ€“] oo_viper_oo@alien.top 1 points 1 year ago

Well, you could calculate distance of all product names or descriptions using some textdistance metric, e.g., rapidfuzz, and then apply clustering on that distance matrix. Resulting clusters would correspond to the products. If you had product images, you could do with some image distance metric.

[โ€“] oo_viper_oo@alien.top 1 points 1 year ago (1 children)

Llama 2 and it's derivatives seem to run just fine on Apple Silicon: https://github.com/ggerganov/llama.cpp