[Skip to content](https://github.com/ggml-org/llama.cpp/discussions/25167#start-of-content) You signed in with another tab or window. [Reload](https://github.com/ggml-org/llama.cpp/discussions/25167) to refresh your session.You signed out in another tab or window. [Reload](https://github.com/ggml-org/llama.cpp/discussions/25167) to refresh your session.You switched accounts on another tab or window. [Reload](https://github.com/ggml-org/llama.cpp/discussions/25167) to refresh your session.Dismiss alert {{ message }} ### Uh oh! There was an error while loading. [Please reload this page](https://github.com/ggml-org/llama.cpp/discussions/25167). [ggml-org](https://github.com/ggml-org)/ **[llama.cpp](https://github.com/ggml-org/llama.cpp)** Public - [Notifications](https://github.com/login?return_to=%2Fggml-org%2Fllama.cpp) You must be signed in to change notification settings - [Fork\\ 20.2k](https://github.com/login?return_to=%2Fggml-org%2Fllama.cpp) - [Star\\ 119k](https://github.com/login?return_to=%2Fggml-org%2Fllama.cpp) # Support DeepSeek DSpark speculative decoding \#25167 [bendoverplz](https://github.com/bendoverplz) started this conversation in [Ideas](https://github.com/ggml-org/llama.cpp/discussions/categories/ideas) [Support DeepSeek DSpark speculative decoding](https://github.com/ggml-org/llama.cpp/discussions/25167#top)#25167 [![@bendoverplz](https://avatars.githubusercontent.com/u/161540363?s=40&v=4)\\ bendoverplz](https://github.com/bendoverplz) 4 days agoJun 30, 2026· 1 comment [Return to top](https://github.com/ggml-org/llama.cpp/discussions/25167#top) Discussion options ### Uh oh! There was an error while loading. [Please reload this page](https://github.com/ggml-org/llama.cpp/discussions/25167). # {{title}} Quote reply ## [![](https://avatars.githubusercontent.com/u/161540363?s=64&v=4)\ bendoverplz](https://github.com/bendoverplz) [4 days agoJun 30, 2026](https://github.com/ggml-org/llama.cpp/discussions/25167\#discussion-10345729) - | | | --- | | DeepSeek recently introduced DSpark, which improves speculative decoding by increasing acceptance rates and inference throughput. It would be valuable to add support for DSpark in llama.cpp so compatible GGUF models can benefit without relying on custom forks. I have an rtx 3070 8gb and willing to run tests or provide feedback however I can't implement or code anything.
DeepSpec repo: [https://github.com/deepseek-ai/DeepSpec](https://github.com/deepseek-ai/DeepSpec)
DSpark paper: [https://github.com/deepseek-ai/DeepSpec/blob/main/DSpark\_paper.pdf](https://github.com/deepseek-ai/DeepSpec/blob/main/DSpark_paper.pdf) | BetaWas this translation helpful? [Give feedback.](https://github.com/ggml-org/llama.cpp/discussions/25167#) 5You must be logged in to vote All reactions ## Replies: 1 comment Comment options ### Uh oh! There was an error while loading. [Please reload this page](https://github.com/ggml-org/llama.cpp/discussions/25167). # {{title}} Quote reply ### [![](https://avatars.githubusercontent.com/u/1180939?s=64&v=4)\ giveen](https://github.com/giveen) [2 days agoJul 1, 2026](https://github.com/ggml-org/llama.cpp/discussions/25167\#discussioncomment-17500568) - | | | --- | | [#25173](https://github.com/ggml-org/llama.cpp/pull/25173) ;) | BetaWas this translation helpful? [Give feedback.](https://github.com/ggml-org/llama.cpp/discussions/25167#) 4You must be logged in to vote All reactions 0 replies [Sign up for free](https://github.com/join?source=comment-repo) **to join this conversation on GitHub**. Already have an account? [Sign in to comment](https://github.com/login?return_to=https%3A%2F%2Fgithub.com%2Fggml-org%2Fllama.cpp%2Fdiscussions%2F25167) Category [💡\\ \\ Ideas](https://github.com/ggml-org/llama.cpp/discussions/categories/ideas) Labels None yet 2 participants [![@bendoverplz](https://avatars.githubusercontent.com/u/161540363?s=48&v=4)](https://github.com/bendoverplz)[![@giveen](https://avatars.githubusercontent.com/u/1180939?s=48&v=4)](https://github.com/giveen) Heading Bold Italic Quote Code Link * * * Numbered list Unordered list Task list * * * Attach files Mention Reference # Select a reply Loading [Create a new saved reply](https://github.com/ggml-org/llama.cpp/discussions/25167) 👍1 reacted with thumbs up emoji👎1 reacted with thumbs down emoji😄1 reacted with laugh emoji🎉1 reacted with hooray emoji😕1 reacted with confused emoji❤️1 reacted with heart emoji🚀1 reacted with rocket emoji👀1 reacted with eyes emoji You can’t perform that action at this time.