Misc updates (pytorch#1868)

Summary:
Pull Request resolved: pytorch#1868

- Add better logging for failed Python import test

- Add more regex patterns to .gitignore

Reviewed By: sryap

Differential Revision: D47345742

fbshipit-source-id: be2ca19c1c03d8b6c14836434c79f6ebeea7c3a5

.github/scripts/utils_base.bash

@@ -83,7 +83,7 @@ test_python_import () {
   if conda run -n "${env_name}" python -c "import ${python_import}"; then
     echo "[CHECK] Python package ${python_import} found."
   else
-    echo "[CHECK] Python package ${python_import} not found!"
+    echo "[CHECK] Python package ${python_import} was not found or is broken!"
     return 1
   fi
 }

.gitignore

@@ -1,2 +1,88 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+
+# Configuration is based on recommendations for C++, CMake, CUDA, and Python
+# found in:
+# https://github.com/github/gitignore/
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# Distribution / packaging
+.Python
 build/
 _skbuild/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# Prerequisites
+*.d
+
+# Compiled Object files
+*.slo
+*.lo
+*.o
+*.obj
+
+# Precompiled Headers
+*.gch
+*.pch
+
+# Compiled Dynamic libraries
+*.so
+*.dylib
+*.dll
+
+# Fortran module files
+*.mod
+*.smod
+
+# Compiled Static libraries
+*.lai
+*.la
+*.a
+*.lib
+
+# Executables
+*.exe
+*.out
+*.app
+
+# CUDA outputs
+*.i
+*.ii
+*.gpu
+*.ptx
+*.cubin
+*.fatbin
+
+# CMake outputs
+CMakeLists.txt.user
+CMakeCache.txt
+CMakeFiles
+CMakeScripts
+Testing
+cmake_install.cmake
+install_manifest.txt
+compile_commands.json
+CTestTestfile.cmake
+_deps

fbgemm_gpu/codegen/embedding_backward_split_template.cu

@@ -493,6 +493,7 @@ Tensor split_embedding{{ "_nobag" if nobag else "" }}_backward_codegen_{{ optimi
             }
             {%- endif %}
 
+
             {%- if not dense and optimizer != "none" %}
             at::PhiloxCudaState rng_engine_inputs;
             if (stochastic_rounding && !std::is_same<emb_t, float>::value) {
@@ -503,6 +504,8 @@ Tensor split_embedding{{ "_nobag" if nobag else "" }}_backward_codegen_{{ optimi
                         ->philox_cuda_state(4);
             }
             {%- endif %}
+
+
             // kMaxElemPerThread is # of elements handled by thread if we use a full warp for a row
             // We consider kMaxElemPerThread 1 and 2, and then a multiple of 4.
             {%- for kMaxElemPerThread in range(1, max_embedding_dim // (items_per_warp // 4) + 1) %}