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(nix): expose LAYA_BODY_READ_TIMEOUT as bodyReadTimeout
thread3dc708bc110968
diff --git a/nix/laya-serve.nix b/nix/laya-serve.nixindex 6835dfc9..a7ef1c1d 100644--- a/nix/laya-serve.nix+++ b/nix/laya-serve.nix@@ -215,6 +215,22 @@ in'';};+ bodyReadTimeout = lib.mkOption {+ type = lib.types.nullOr lib.types.numbers.positive;+ default = null;+ example = 30;+ description = ''+ Seconds a request may spend delivering its body (sets+ `LAYA_BODY_READ_TIMEOUT`). Admission is taken before the body is read, so+ a client that opens a connection, sends a request head and then stops+ writing holds one of `maxConcurrent`'s slots for as long as it stays+ connected; past this deadline the server answers 408 and gives the slot+ back. Raise it for clients on links slower than a few kilobytes a second,+ lower it to turn a stalled upload into a refusal faster. null leaves the+ server's own default (10 seconds).+ '';+ };+maxLoaded = lib.mkOption {type = lib.types.nullOr lib.types.ints.positive;default = null;@@ -381,6 +397,8 @@ inLAYA_LOG_LEVEL = cfg.logLevel;} // lib.optionalAttrs (cfg.maxConcurrent != null) {LAYA_MAX_CONCURRENT = toString cfg.maxConcurrent;+ } // lib.optionalAttrs (cfg.bodyReadTimeout != null) {+ LAYA_BODY_READ_TIMEOUT = toString cfg.bodyReadTimeout;} // lib.optionalAttrs (cfg.cudaAmp != null) {LAYA_CUDA_AMP = cfg.cudaAmp;} // lib.optionalAttrs (cfg.cpuAmp != null) {