mtmd: mtmd_audio_streaming_istft for audio output - #18645
Merged
Merged
Conversation
Change is decoupled from ggml-org#18641. [LFM2.5-Audio-1.5B](https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B) needs streaming istft for generating output audio. * add streaming ISTFT class (`mtmd_audio_streaming_istft`) with overlap-add for audio reconstruction * replace global audio cache with per-instance cache, the model requires two independent caches, for preprocessing (audio input) and for istft (audio output). * unified templated FFT/IFFT implementation supporting both forward and inverse transforms
4 of 6 tasks
ngxson
approved these changes
Jan 6, 2026
ngxson
left a comment
ngxson
left a comment
Collaborator
There was a problem hiding this comment.
nice, thanks! I ran the test and confirmed that this doesn't break other models
Seunghhon
pushed a commit
to Seunghhon/llama.cpp
that referenced
this pull request
Apr 26, 2026
Change is decoupled from ggml-org#18641. [LFM2.5-Audio-1.5B](https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B) needs streaming istft for generating output audio. * add streaming ISTFT class (`mtmd_audio_streaming_istft`) with overlap-add for audio reconstruction * replace global audio cache with per-instance cache, the model requires two independent caches, for preprocessing (audio input) and for istft (audio output). * unified templated FFT/IFFT implementation supporting both forward and inverse transforms
my-other-github-account
pushed a commit
to my-other-github-account/llama.cpp
that referenced
this pull request
May 15, 2026
Change is decoupled from ggml-org#18641. [LFM2.5-Audio-1.5B](https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B) needs streaming istft for generating output audio. * add streaming ISTFT class (`mtmd_audio_streaming_istft`) with overlap-add for audio reconstruction * replace global audio cache with per-instance cache, the model requires two independent caches, for preprocessing (audio input) and for istft (audio output). * unified templated FFT/IFFT implementation supporting both forward and inverse transforms
fukuro-kun
pushed a commit
to fukuro-kun/fukuro-llama-cpp-turboquant
that referenced
this pull request
Jul 5, 2026
Change is decoupled from ggml-org#18641. [LFM2.5-Audio-1.5B](https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B) needs streaming istft for generating output audio. * add streaming ISTFT class (`mtmd_audio_streaming_istft`) with overlap-add for audio reconstruction * replace global audio cache with per-instance cache, the model requires two independent caches, for preprocessing (audio input) and for istft (audio output). * unified templated FFT/IFFT implementation supporting both forward and inverse transforms
MrLordCat
referenced
this pull request
in MrLordCat/llama.cpp-rdna-lab
Jul 16, 2026
Change is decoupled from ggml-org/llama.cpp#18641. [LFM2.5-Audio-1.5B](https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B) needs streaming istft for generating output audio. * add streaming ISTFT class (`mtmd_audio_streaming_istft`) with overlap-add for audio reconstruction * replace global audio cache with per-instance cache, the model requires two independent caches, for preprocessing (audio input) and for istft (audio output). * unified templated FFT/IFFT implementation supporting both forward and inverse transforms
zommiommy
pushed a commit
to zommiommy/llama.cpp
that referenced
this pull request
Aug 18, 2026
Change is decoupled from ggml-org#18641. [LFM2.5-Audio-1.5B](https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B) needs streaming istft for generating output audio. * add streaming ISTFT class (`mtmd_audio_streaming_istft`) with overlap-add for audio reconstruction * replace global audio cache with per-instance cache, the model requires two independent caches, for preprocessing (audio input) and for istft (audio output). * unified templated FFT/IFFT implementation supporting both forward and inverse transforms
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Change is decoupled from #18641.
LFM2.5-Audio-1.5B needs streaming istft for generating output audio.
mtmd_audio_streaming_istft) with overlap-add for audio reconstructionMake sure to read the contributing guidelines before submitting a PR