[model] add support for qwen3vl series #16780
ngxson
merged
24 commits into
ggml-org:master
from
JJJYmmm:add_qwen3vl
support qwen3vl series.
1e4fd194
bugfix: fix the arch check for qwen3vl-moe.
f84bd67c
ngxson
commented
on 2025-10-28
use build_ffn
0443a098
ngxson
commented
on 2025-10-28
optimize deepstack structure
32718772
optimize deepstack feature saving
f321b9fd
Merge branch 'ggml-org:master' into add_qwen3vl
b0169b78
Revert "optimize deepstack feature saving" for temporal fix
794481e4
code clean
0b37fffb
use fused qkv in clip
2be9279f
clean up / rm is_deepstack_layers for simplification
2a669261
Merge branch 'master' into add_qwen3vl
52e3dac8
ngxson
approved these changes
on 2025-10-29
add test model
96037a89
move test model to "big" section
473ee903
ggerganov
commented
on 2025-10-30
fix imrope check
e9a167b2
remove trailing whitespace
69e26ffe
ggerganov
commented
on 2025-10-30
fix rope fail
0518b0a9
metal : add imrope support
6a0191a7
add imrope support for sycl
10ce7fb1
CISC
approved these changes
on 2025-10-30
vulkan: add imrope w/o check
cbca610e
fix vulkan
950c7646
ngxson
commented
on 2025-10-30
Merge branch 'master' into add_qwen3vl
19a458ff
CISC
commented
on 2025-10-30
webgpu: add imrope w/o check
0bed5d87
Update gguf-py/gguf/tensor_mapping.py
b3380106
fix tensor mapping
7d9c149e
ggerganov
approved these changes
on 2025-10-30
Reviewers
Assignees
No one assigned
Labels
testing
Nvidia GPU
Vulkan
examples
python
ggml
SYCL
Apple Metal
Milestone
No milestone
Login to write a write a comment.
Login via GitHub