Skip to content
Repo

DeepSeek_V4_1_Flash

from modal_dojo import DeepSeek_V4_1_Flash

DeepSeek-V4.1-Flash sparse-attention MoE model, 40 layers and 384 routed experts.

download() -> None

Download or materialize weights into the model volume.

parse_response(text: str) -> ParsedResponse

Parse model text with response_parser.

Without a configured parser, the model text becomes ParsedResponse.content.

Returns

Parsed model output.