filterChannelContentFromKvCache property

bool? get filterChannelContentFromKvCache

Whether channel content is filtered from the KV cache.

If null, uses the model or runtime default. If true, channel content (e.g. reasoning) will be filtered from the KV cache. If false, channel content remains in the KV cache.

Note: This flag is read only when a new Conversation is created. Changing this value will not affect any existing Conversation instances.

Implementation

static bool? get filterChannelContentFromKvCache =>
    LiteRtLmNativeRuntime.instance.filterChannelContentFromKvCache;
set filterChannelContentFromKvCache (bool? value)

Implementation

static set filterChannelContentFromKvCache(bool? value) {
  LiteRtLmNativeRuntime.instance.filterChannelContentFromKvCache = value;
}