llama.cpp
c02b0a8a - cann: support q4_0 model (#8822)

Commit

343 days ago

cann: support q4_0 model (#8822)

References

#8822 - [CANN] Support Q4_0 for Ascend NPU

Author

wangshuai09

wangshuai09

Parents

Files7

ggml/src
- ggml-cann.cpp
- ggml-cann
  - acl_tensor.cpp
  - acl_tensor.h
  - aclnn_ops.cpp
  - kernels
    - CMakeLists.txt
    - ascendc_kernels.h
    - quantize_float_to_q4_0.cpp