models : add DeepSeek V4 Flash conversions - #11
Merged
Conversation
MTP sidecar conversion is commented out for now, will be re-enabled later. Assisted-by: llama.cpp:Qwen3.6-27B
ggerganov
force-pushed
the
gg/deepseek-v4-flash
branch
from
August 2, 2026 14:06
d2f0070 to
660cc5d
Compare
- deepseek-v4-flash: MTP extracted from same model (PATH_PRIMARY) - deepseek-v4-flash-0731: MTP sourced from deepseek-v4-flash (DEP_MTP/PATH_MTP) Assisted-by: llama.cpp:Qwen3.6-27B
Assisted-by: llama.cpp:Qwen3.6-27B
Assisted-by: llama.cpp:Qwen3.6-27B
Assisted-by: llama.cpp:Qwen3.6-27B
Assisted-by: llama.cpp:Qwen3.6-27B
ggerganov
force-pushed
the
gg/deepseek-v4-flash
branch
from
August 2, 2026 16:28
90e328c to
79c0e20
Compare
ggerganov
changed the title
ggerganov
marked this pull request as ready for review
August 2, 2026 16:41
Member
Author
|
The base model currently fails to convert: ggml-org/llama.cpp#26452 (comment) |
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Add conversion scripts for three DeepSeek V4 Flash variants:
Notes
MXFP4_MOEquantization (BF16 not uploaded, only used as intermediate)--mtpflag--dspark --target-model-dirflags