40 support tickets in English and Thai: Jev 90–100%, Laya zero-shot 30–80%
A local MLX Laya run against Jev on 40 real support tickets in EN and Thai. Laya was faster but not close on accuracy yet, per the author.
A local MLX Laya run against Jev on 40 real support tickets in EN and Thai. Laya was faster but not close on accuracy yet, per the author.
A Myanmar-language Laya checkpoint trained on Davlan/sib200, with a Gradio demo space from the same author.
aungthuhein-dev · Multilingual · free
aungthuhein-dev
GitHub
Multilingual
free
-
An Arabic Laya checkpoint trained on AmazonScience/massive and facebook/xnli, with a companion laya-ara-rag ranking model from the same author.
Wouze · Multilingual · free
Wouze
GitHub
Multilingual
free
-
Google's LiteRT community conversion of the multilingual checkpoint (tflite), tagged for Android text classification. An English LiteRT build is published alongside it.
litert-community · Multilingual · free, on-device
litert-community
GitHub
Multilingual
free, on-device
-
Synthetic Chinese workplace-chat scenarios with frozen inputs, prompts and labels. Reports classification quality, mis-generated tasks and latency: Jev 253 ms vs Laya 151 ms.
Feng · Multilingual · 151 ms (Laya) vs 253 ms (Jev)
Feng
GitHub
Multilingual
-
151 ms (Laya) vs 253 ms (Jev)
Tested Jev vs Laya (local, MLX) on 40 support tickets, EN + Thai. Jev 90–100%. Laya zero-shot 30–80%. Laya's faster, but not close on accuracy yet.