hashbrown

ExperimentalWebLlmImageTransportOptions

Configuration asserting the application engine supports image input.

engine
{
        chat: {
            completions: {
                create: (request: WebLlmRequest<WebLlmImageContent>) => Promise<AsyncIterable<WebLlmChunk>> | AsyncIterable<WebLlmChunk>;
            };
        };
        interruptGenerate(): void;
    }

Loaded image-capable engine; Hashbrown never loads, resets or unloads it.

model
string

Selects an already loaded model.