Closed jatinarora15 closed 6 months ago
cache/vector/repo is supported to contain embedding vectors of repo windows. In run_pipeline.py, one has to manually call the https://github.com/microsoft/CodeT/blob/35f54d60b152cc31d134b788e702878ad613d9f7/RepoCoder/build_vector.py#L46 before starting the search process.
The cache files under "vector/repos/" are vectorized code fragments built from the code files within a repo. During code search, the vectorized code fragments are used for similarity comparison. So before running RG1 or repocoder method, we need to vectorize the code windows produced from each repo by calling the vectorize_repo_windows() function. I have updated the code in this commit (https://github.com/microsoft/CodeT/commit/6a6ef6359a3587e134d8350c3eebb0e639e7789a). Sorry for the inconvenience.
While running
run_pipeline.py
I'm encountering the below error:Additional Info:
Can someone please help with this issue.