Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a single forward pass, in 100+ languages, with a router that picks the right checkpoint per request.
Python calibrationclassificationdecision-modelhuggingfacejevmodernbertmultilingualnlppythonpytorchroutingtyped-decisionszero-shot
fix(compose): stop the example override from contradicting the page's model default
Aashish5ae77abd7e37
diff --git a/compose.example.yml b/compose.example.ymlindex 6d0cb477..760215fe 100644--- a/compose.example.yml+++ b/compose.example.yml@@ -6,14 +6,16 @@ services:laya:environment:LAYA_DEVICE: "${LAYA_DEVICE:-cpu}"- LAYA_MODEL: "${LAYA_MODEL:-english}"+ LAYA_MODEL: "${LAYA_MODEL:-auto}"LAYA_REQUEST_FILE: /inputs/request.jsonOMP_NUM_THREADS: "${OMP_NUM_THREADS:-4}"HF_HUB_OFFLINE: "${HF_HUB_OFFLINE:-0}"- # Local checkpoint: set LAYA_MODEL=auto above, then enable both this- # setting and the checkpoint mount below. The directory must be writable- # by UID 10001 for tokenizer compatibility updates.+ # Local checkpoint: leave LAYA_MODEL at the auto default above. The entry+ # script reads a name here as an explicit alias and refuses an alias and a+ # path together, so this file cannot preselect one and still show the mount.+ # Enable both this setting and the checkpoint mount below. The directory+ # must be writable by UID 10001 for tokenizer compatibility updates.# LAYA_MODEL_PATH: /models/custom# Optional token: enable this setting and its read-only mount below.