Improve OS related lines and add .gitignore

.gitignore

@@ -0,0 +1,142 @@
+model/
+models/
+.vscode/
+dataset/
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/

pytorch/bts_main.py

@@ -17,7 +17,7 @@
 import time
 import argparse
 import datetime
-import sys
+import sys, shutil
 import os
 
 import torch
@@ -434,11 +434,11 @@ def main_worker(gpu, ngpus_per_node, args):
     num_log_images = args.batch_size
     end_learning_rate = args.end_learning_rate if args.end_learning_rate != -1 else 0.1 * args.learning_rate
 
-    var_sum = [var.sum() for var in model.parameters() if var.requires_grad]
-    var_cnt = len(var_sum)
-    var_sum = np.sum(var_sum)
+    # var_sum = [var.sum().detach().numpy() for var in model.parameters() if var.requires_grad]
+    # var_cnt = len(var_sum)
+    # var_sum = np.sum(var_sum)
 
-    print("Initial variables' sum: {:.3f}, avg: {:.3f}".format(var_sum, var_sum/var_cnt))
+    # print("Initial variables' sum: {:.3f}, avg: {:.3f}".format(var_sum, var_sum/var_cnt))
 
     steps_per_epoch = len(dataloader.data)
     num_total_steps = args.num_epochs * steps_per_epoch
@@ -481,7 +481,7 @@ def main_worker(gpu, ngpus_per_node, args):
             if global_step and global_step % args.log_freq == 0 and not model_just_loaded:
                 var_sum = [var.sum() for var in model.parameters() if var.requires_grad]
                 var_cnt = len(var_sum)
-                var_sum = np.sum(var_sum)
+                var_sum = torch.sum(var_sum)
                 examples_per_sec = args.batch_size / duration * args.log_freq
                 duration = 0
                 time_sofar = (time.time() - start_time) / 3600
@@ -570,30 +570,36 @@ def main():
 
     model_filename = args.model_name + '.py'
     args.model_name=args.model_name + '_rank_'+str(args.att_rank)
-    command = 'mkdir ' + args.log_directory + '/' + args.model_name
-    os.system(command)
+    # command = 'mkdir ' + args.log_directory + '/' + args.model_name
+    # os.system(command)
+    os.makedirs(args.log_directory + '/' + args.model_name, exist_ok=True)
 
     args_out_path = args.log_directory + '/' + args.model_name + '/' + sys.argv[1]
-    command = 'cp ' + sys.argv[1] + ' ' + args_out_path
-    os.system(command)
+    # command = 'cp ' + sys.argv[1] + ' ' + args_out_path
+    # os.system(command)
+    shutil.copyfile(sys.argv[1], args_out_path)
 
     if args.checkpoint_path == '':
         model_out_path = args.log_directory + '/' + args.model_name + '/' + model_filename
-        command = 'cp bts.py ' + model_out_path
-        os.system(command)
+        # command = 'cp bts.py ' + model_out_path
+        # os.system(command)
+        shutil.copyfile('bts.py',model_out_path )
         aux_out_path = args.log_directory + '/' + args.model_name + '/.'
-        command = 'cp bts_main.py ' + aux_out_path
-        os.system(command)
-        command = 'cp bts_dataloader.py ' + aux_out_path
-        os.system(command)
+        # command = 'cp bts_main.py ' + aux_out_path
+        # os.system(command)
+        shutil.copyfile('bts_main.py',os.path.join(aux_out_path, 'bts_main.py'))
+        # command = 'cp bts_dataloader.py ' + aux_out_path
+        # os.system(command)
+        shutil.copyfile('bts_dataloader.py',os.path.join(aux_out_path, 'bts_dataloader.py'))
     else:
         loaded_model_dir = os.path.dirname(args.checkpoint_path)
         loaded_model_name = os.path.basename(loaded_model_dir)
         loaded_model_filename = loaded_model_name + '.py'
 
         model_out_path = args.log_directory + '/' + args.model_name + '/' + model_filename
-        command = 'cp ' + loaded_model_dir + '/' + loaded_model_filename + ' ' + model_out_path
-        os.system(command)
+        # command = 'cp ' + loaded_model_dir + '/' + loaded_model_filename + ' ' + model_out_path
+        # os.system(command)
+        shutil.copyfile(loaded_model_dir + '/' + loaded_model_filename, model_out_path)
 
     torch.cuda.empty_cache()
     args.distributed = args.world_size > 1 or args.multiprocessing_distributed