Skip to content

Upgrade llama.cpp to b10319: migrate TTS from OuteTTS to Qwen3-TTS #381

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
bernardladenthin merged 29 commits into from
Aug 7, 2026
Merged

Upgrade llama.cpp to b10319: migrate TTS from OuteTTS to Qwen3-TTS #381

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
29 commits
Select commit Hold shift + click to select a range
ac1ccb2
Upgrade llama.cpp from b10217 to b10221
claude Aug 7, 2026
a554f31
Upgrade llama.cpp from b10221 to b10226
claude Aug 7, 2026
f0bf32f
Upgrade llama.cpp from b10226 to b10231
claude Aug 7, 2026
272d3d9
Upgrade llama.cpp from b10231 to b10236
claude Aug 7, 2026
2542459
Upgrade llama.cpp from b10236 to b10241
claude Aug 7, 2026
ee0301a
Upgrade llama.cpp from b10241 to b10246
claude Aug 7, 2026
2e4bb3a
Upgrade llama.cpp from b10246 to b10251
claude Aug 7, 2026
b29bc66
Upgrade llama.cpp from b10251 to b10256
claude Aug 7, 2026
1a7eaf5
Upgrade llama.cpp from b10256 to b10261
claude Aug 7, 2026
562335c
Upgrade llama.cpp from b10261 to b10265
claude Aug 7, 2026
4127e2e
Upgrade llama.cpp from b10265 to b10269
claude Aug 7, 2026
d119cc2
Upgrade llama.cpp from b10269 to b10270; rework TextToSpeech onto Qwe…
claude Aug 7, 2026
ced3a99
Upgrade llama.cpp from b10270 to b10275
claude Aug 7, 2026
3de20c3
Upgrade llama.cpp from b10275 to b10280
claude Aug 7, 2026
d893122
Bump pin files to b10280 and fix eval_llama_cmpl_schema n_ctx_slot drop
claude Aug 7, 2026
264f7eb
Upgrade llama.cpp from b10280 to b10285
claude Aug 7, 2026
9316c4b
Upgrade llama.cpp from b10285 to b10290
claude Aug 7, 2026
fd0ddf5
Upgrade llama.cpp from b10290 to b10295
claude Aug 7, 2026
c7cec86
Confirm b10295 checkpoint: full build + ctest 482/482 passing
claude Aug 7, 2026
7c9128b
Upgrade llama.cpp from b10295 to b10299
claude Aug 7, 2026
deffcd2
Upgrade llama.cpp from b10299 to b10303
claude Aug 7, 2026
74d40a4
Upgrade llama.cpp from b10303 to b10308
claude Aug 7, 2026
f7519d7
Confirm b10308 checkpoint: full build + ctest 482/482 passing
claude Aug 7, 2026
8c99102
Upgrade llama.cpp from b10308 to b10310 (final target of this walk)
claude Aug 7, 2026
8e9bd72
Confirm b10310 final target: full build + ctest 482/482 passing
claude Aug 7, 2026
6edb708
Wire Qwen3-TTS into CI: TtsIntegrationTest runs on every platform now
claude Aug 7, 2026
1af3d82
Upgrade llama.cpp from b10310 to b10313
claude Aug 7, 2026
50afb5e
Upgrade llama.cpp from b10313 to b10318
claude Aug 7, 2026
79df1e6
Upgrade llama.cpp from b10318 to b10319 (final target)
claude Aug 7, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions .github/models.csv
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,5 +17,5 @@ Qwen2.5-1.5B-Instruct-Q4_K_M.gguf,https://huggingface.co/bartowski/Qwen2.5-1.5B-
nomic-embed-text-v1.5.f16.gguf,https://huggingface.co/nomic-ai/nomic-embed-text-v1.5-GGUF/resolve/main/nomic-embed-text-v1.5.f16.gguf
SmolVLM-500M-Instruct-Q8_0.gguf,https://huggingface.co/ggml-org/SmolVLM-500M-Instruct-GGUF/resolve/main/SmolVLM-500M-Instruct-Q8_0.gguf
mmproj-SmolVLM-500M-Instruct-Q8_0.gguf,https://huggingface.co/ggml-org/SmolVLM-500M-Instruct-GGUF/resolve/main/mmproj-SmolVLM-500M-Instruct-Q8_0.gguf
OuteTTS-0.2-500M-Q4_K_M.gguf,https://huggingface.co/second-state/OuteTTS-0.2-500M-GGUF/resolve/main/OuteTTS-0.2-500M-Q4_K_M.gguf
WavTokenizer-Large-75-F16.gguf,https://huggingface.co/ggml-org/WavTokenizer/resolve/main/WavTokenizer-Large-75-F16.gguf
Qwen3-TTS-12Hz-1.7B-Base-Q4_K_M.gguf,https://huggingface.co/ggml-org/Qwen3-TTS-12Hz-1.7B-Base-GGUF/resolve/main/Qwen3-TTS-12Hz-1.7B-Base-Q4_K_M.gguf
mmproj-Qwen3-TTS-12Hz-1.7B-Base-Q8_0.gguf,https://huggingface.co/ggml-org/Qwen3-TTS-12Hz-1.7B-Base-GGUF/resolve/main/mmproj-Qwen3-TTS-12Hz-1.7B-Base-Q8_0.gguf
33 changes: 18 additions & 15 deletions .github/workflows/publish.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,9 +36,12 @@ env:
# under the upstream mtmd pipeline.
VISION_MODEL_NAME: "SmolVLM-500M-Instruct-Q8_0.gguf"
VISION_MMPROJ_NAME: "mmproj-SmolVLM-500M-Instruct-Q8_0.gguf"
# Text-to-speech models for AudioInputIntegrationTest's sibling TtsIntegrationTest (OuteTTS pipeline).
TTS_MODEL_NAME: "OuteTTS-0.2-500M-Q4_K_M.gguf"
TTS_VOCODER_NAME: "WavTokenizer-Large-75-F16.gguf"
# Qwen3-TTS backbone + mmproj for TtsIntegrationTest (the OuteTTS+WavTokenizer pair this replaced
# was retired when upstream #26254 replaced the whole TTS pipeline — see
# docs/history/llama-cpp-breaking-changes.md, b10269-b10270). Smallest available quants:
# Q4_K_M backbone (0.96 GiB), Q8_0 mmproj (0.42 GiB; no smaller mmproj quant is published).
TTS_MODEL_NAME: "Qwen3-TTS-12Hz-1.7B-Base-Q4_K_M.gguf"
TTS_MMPROJ_NAME: "mmproj-Qwen3-TTS-12Hz-1.7B-Base-Q8_0.gguf"
# Test image used by MultimodalIntegrationTest is committed to the repo
# at src/test/resources/images/test-image.jpg (see the README in that
# directory for licensing). No download step is needed; CI just points
Expand Down Expand Up @@ -2241,8 +2244,8 @@ jobs:
-Dnet.ladenthin.llama.vision.model=models/${VISION_MODEL_NAME} \
-Dnet.ladenthin.llama.vision.mmproj=models/${VISION_MMPROJ_NAME} \
-Dnet.ladenthin.llama.vision.image=${VISION_IMAGE_PATH} \
-Dnet.ladenthin.llama.tts.ttc.model=models/${TTS_MODEL_NAME} \
-Dnet.ladenthin.llama.tts.vocoder.model=models/${TTS_VOCODER_NAME}
-Dnet.ladenthin.llama.tts.model=models/${TTS_MODEL_NAME} \
-Dnet.ladenthin.llama.tts.mmproj=models/${TTS_MMPROJ_NAME}
- uses: actions/upload-artifact@v7
if: success()
with:
Expand Down Expand Up @@ -2362,8 +2365,8 @@ jobs:
-Dnet.ladenthin.llama.vision.model=models/${VISION_MODEL_NAME} \
-Dnet.ladenthin.llama.vision.mmproj=models/${VISION_MMPROJ_NAME} \
-Dnet.ladenthin.llama.vision.image=${VISION_IMAGE_PATH} \
-Dnet.ladenthin.llama.tts.ttc.model=models/${TTS_MODEL_NAME} \
-Dnet.ladenthin.llama.tts.vocoder.model=models/${TTS_VOCODER_NAME}
-Dnet.ladenthin.llama.tts.model=models/${TTS_MODEL_NAME} \
-Dnet.ladenthin.llama.tts.mmproj=models/${TTS_MMPROJ_NAME}
- name: Memory after tests
if: always()
run: vm_stat && sysctl hw.memsize hw.physmem
Expand Down Expand Up @@ -2431,8 +2434,8 @@ jobs:
-Dnet.ladenthin.llama.vision.model=models/${VISION_MODEL_NAME} \
-Dnet.ladenthin.llama.vision.mmproj=models/${VISION_MMPROJ_NAME} \
-Dnet.ladenthin.llama.vision.image=${VISION_IMAGE_PATH} \
-Dnet.ladenthin.llama.tts.ttc.model=models/${TTS_MODEL_NAME} \
-Dnet.ladenthin.llama.tts.vocoder.model=models/${TTS_VOCODER_NAME}
-Dnet.ladenthin.llama.tts.model=models/${TTS_MODEL_NAME} \
-Dnet.ladenthin.llama.tts.mmproj=models/${TTS_MMPROJ_NAME}
- name: Memory after tests
if: always()
run: vm_stat && sysctl hw.memsize hw.physmem
Expand Down Expand Up @@ -2500,8 +2503,8 @@ jobs:
-Dnet.ladenthin.llama.vision.model=models/${VISION_MODEL_NAME} \
-Dnet.ladenthin.llama.vision.mmproj=models/${VISION_MMPROJ_NAME} \
-Dnet.ladenthin.llama.vision.image=${VISION_IMAGE_PATH} \
-Dnet.ladenthin.llama.tts.ttc.model=models/${TTS_MODEL_NAME} \
-Dnet.ladenthin.llama.tts.vocoder.model=models/${TTS_VOCODER_NAME}
-Dnet.ladenthin.llama.tts.model=models/${TTS_MODEL_NAME} \
-Dnet.ladenthin.llama.tts.mmproj=models/${TTS_MMPROJ_NAME}
- name: Memory after tests
if: always()
run: vm_stat && sysctl hw.memsize hw.physmem
Expand Down Expand Up @@ -2587,8 +2590,8 @@ jobs:
"-Dnet.ladenthin.llama.vision.model=models/$env:VISION_MODEL_NAME" `
"-Dnet.ladenthin.llama.vision.mmproj=models/$env:VISION_MMPROJ_NAME" `
"-Dnet.ladenthin.llama.vision.image=$env:VISION_IMAGE_PATH" `
"-Dnet.ladenthin.llama.tts.ttc.model=models/$env:TTS_MODEL_NAME" `
"-Dnet.ladenthin.llama.tts.vocoder.model=models/$env:TTS_VOCODER_NAME"
"-Dnet.ladenthin.llama.tts.model=models/$env:TTS_MODEL_NAME" `
"-Dnet.ladenthin.llama.tts.mmproj=models/$env:TTS_MMPROJ_NAME"
- name: Memory after tests
if: always()
run: Get-CimInstance Win32_OperatingSystem | Select-Object FreePhysicalMemory,TotalVisibleMemorySize | Format-List
Expand Down Expand Up @@ -2681,8 +2684,8 @@ jobs:
"-Dnet.ladenthin.llama.vision.model=models/$env:VISION_MODEL_NAME" `
"-Dnet.ladenthin.llama.vision.mmproj=models/$env:VISION_MMPROJ_NAME" `
"-Dnet.ladenthin.llama.vision.image=$env:VISION_IMAGE_PATH" `
"-Dnet.ladenthin.llama.tts.ttc.model=models/$env:TTS_MODEL_NAME" `
"-Dnet.ladenthin.llama.tts.vocoder.model=models/$env:TTS_VOCODER_NAME"
"-Dnet.ladenthin.llama.tts.model=models/$env:TTS_MODEL_NAME" `
"-Dnet.ladenthin.llama.tts.mmproj=models/$env:TTS_MMPROJ_NAME"
- name: Memory after tests
if: always()
run: Get-CimInstance Win32_OperatingSystem | Select-Object FreePhysicalMemory,TotalVisibleMemorySize | Format-List
Expand Down
Loading
Loading