junrushao opened a new pull request, #334: URL: https://github.com/apache/tvm-ffi/pull/334
Should fix #264, #322. Supersedes #331. This PR introduces a flag `keep_module_alive: bool = True` to the following APIs: - **`load_module`** - **`cpp.load_inline`** - **`cpp.load`** By default, the flag is set to `True`, meaning it is bound to `ModuleGlobals` inside `libtvm_ffi.so`. It retains a reference count to those modules, and Python side will not unload the module. Two private APIs are added: - `tvm_ffi._ffi_api.ModuleGlobalsAdd`: adds a module to global registry `ModuleGlobals` - `tvm_ffi._ffi_api.ModuleGlobalsRemove`: removes the module from global registry `ModuleGlobals` -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
