mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-09-04 10:47:38 +02:00
tc-mbandSigbjørn Skjæret
2496f9c149
* Support MiniCPM-V 4.6 in new branch Signed-off-by: tc-mb <[email protected]> * fix code bug Signed-off-by: tc-mb <[email protected]> * fix pre-commit Signed-off-by: tc-mb <[email protected]> * fix convert Signed-off-by: tc-mb <[email protected]> * rename clip_graph_minicpmv4_6 Signed-off-by: tc-mb <[email protected]> * use new TYPE_MINICPMV4_6 Signed-off-by: tc-mb <[email protected]> * use build_attn to allow flash attention support Signed-off-by: tc-mb <[email protected]> * no use legacy code, restored here. Signed-off-by: tc-mb <[email protected]> * use the existing tensors name Signed-off-by: tc-mb <[email protected]> * unused ctx->model.hparams.minicpmv_version Signed-off-by: tc-mb <[email protected]> * use n_merge for slice alignment Signed-off-by: tc-mb <[email protected]> * borrow wa_layer_indexes for vit_merger insertion point Signed-off-by: tc-mb <[email protected]> * fix code style Signed-off-by: tc-mb <[email protected]> * Update convert_hf_to_gguf.py Co-authored-by: Sigbjørn Skjæret <[email protected]> * use filter_tensors and add model.vision_tower Signed-off-by: tc-mb <[email protected]> * fix chkhsh Signed-off-by: tc-mb <[email protected]> * fix type check Signed-off-by: tc-mb <[email protected]> --------- Signed-off-by: tc-mb <[email protected]> Co-authored-by: Sigbjørn Skjæret <[email protected]>