swetha097

Results 7 comments of swetha097

> @swetha097 I think we are using too many parameters for tensor implementation. The idea is to derive as many from the vx_tensor class. This will cause lot of additional...

> please address the review comments Addressed the review comments

> @swetha097 why was the exit label added back? @LakshmiKumar23 The exit label seems to be present in internal_rpp.h file for error check, if removed in the Slice.cpp / NSR.cpp,...

> the patch version in vx_rpp_ext_version file All comments have been addressed except patch version updation in vx_rpp_ext_version file.

> Added few comments Resolved the comments

Tested on 2 commits - the PR branch commit and the base commit PR commit: [https://github.com/QingtaoLi1/tmac_llama.cpp/commit/983c3ecad6588185d89ead87266876b273e0348a](https://github.com/QingtaoLi1/tmac_llama.cpp/commit/983c3ecad6588185d89ead87266876b273e0348a) Base commit: [https://github.com/QingtaoLi1/tmac_llama.cpp/commit/526739b879c9d6a702ed681138611197fa4d3f18](https://github.com/QingtaoLi1/tmac_llama.cpp/commit/526739b879c9d6a702ed681138611197fa4d3f18) While testing the same, we were not able to see the...

> @swetha097 Which model and data type do you test? I want to test at my side to check the reason. We are working with [Meta Llama 2 7B model...