Add support for CogVLM model #15002
Tianyue-Zhao
marked this pull request as ready for review 141 days ago
CISC
commented
on 2025-08-02
CISC
commented
on 2025-08-08
CISC
commented
on 2025-08-18
Added GGUF mappings for CogVLM model
da81af4f
Add tensor mapping for CogVLM visual encoder
6de9d169
Add CogVLM to conversion script, no vision part yet
11fac0c0
Added CogVLM vision model to conversion script
a3adac12
Add graph for CogVLM CLIP model
30267810
Add graph for CogVLM
183ca2eb
Fixes for CogVLM. Now compiles.
ac3e3489
Model now runs
0634dc94
Fixes for cogvlm graph
4c080bf5
Account for graph context change after rebase
bc3f0847
Changes for whitespace
bcbd6ef4
Changes in convert script according to comments
c65f5aa9
Switch CogVLM LLM graph to merged QKV tensor
00af7ee5
Use rope_type variable instead of direct definition
79e5640c
Change CogVLM CLIP encoder to use SWIGLU
86a10bc1
Switch CogVLM CLIP to use merged QKV
a959a1fa
Apply rebase edits and remove ggml_cont call that is now unnecessary
06a07194
Merge branch 'master' into cogvlm_support
f5cb848b
clean up
85764bd8
ngxson
approved these changes
on 2025-10-29
ngxson
merged
bacddc04
into master 50 days ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub