From 740299bd9d04db641a5efb054132340c1fc8b312 Mon Sep 17 00:00:00 2001 From: rob thijssen Date: Wed, 27 May 2026 12:36:18 +0300 Subject: [PATCH] chore(neuron/beast): switch default-model quant from q5k to q6k q5k produced NaN logits on Qwen/Qwen3.6-27B under candle TP=2 (sampler fell over with "logits unhealthy nan: 248320/248320"). q6k is the quant that worked well in production under mistral.rs on the same hardware, so it's the right baseline for verifying the mempool-trim fix. Co-Authored-By: Claude Opus 4.7 (1M context) --- asset/neuron/beast.toml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/asset/neuron/beast.toml b/asset/neuron/beast.toml index 432a660..5def13e 100644 --- a/asset/neuron/beast.toml +++ b/asset/neuron/beast.toml @@ -19,6 +19,6 @@ name = "candle" [[default_models]] model_id = "Qwen/Qwen3.6-27B" harness = "candle" -quant = "q5k" +quant = "q6k" tensor_parallel = 2 devices = [0, 1]