Skip to content

Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.

llm: Fix debug logging for memory estimates

Jesse Grossf4f0992b6ea5
diff --git a/llm/memory.go b/llm/memory.go
index ac830ee84..86694d06d 100644
--- a/llm/memory.go
+++ b/llm/memory.go
@@ -374,7 +374,7 @@ func (m MemoryEstimate) LogValue() slog.Value {
slog.Group(
"weights",
// memory of the weights
- "total", format.HumanBytes2(m.memoryWeights),
+ "total", format.HumanBytes2(m.memoryWeights+m.memoryLayerOutput),
// memory of repeating layers
"repeating", format.HumanBytes2(m.memoryWeights),
// memory of non-repeating layers