Sync master with 'imz-r2.4'

Signed-off-by: Abolfazl Shahbazi <abolfazl.shahbazi@intel.com>

.gitignore

@@ -11,3 +11,5 @@ test_data/
 download_glue_data.py
 data/
 output/
+**/**.whl
+**/**.tar.gz

CODEOWNERS

@@ -1,17 +1,18 @@
 # Lines starting with '#' are comments.
 # Each line is a file pattern followed by one or more owners.
 
-# These owners will be the default owners for everything in the repo.
-*   @mlukaszewski @claynerobison @chuanqi129 @agramesh1 @justkw
+# These owners will be the default owners for everything in the repo,
+# but PR owner should be able to assign other contributors when appropriate
+*    @ashahba @claynerobison @dmsuehir
+datasets @ashahba @claynerobison @dzungductran
+docs @claynerobison @mhbuehler
+k8s  @ashahba @dzungductran @kkasravi
+models @agramesh1 @ashraf-bhuiyan @riverliuintel @wei-v-wang
 
 # Order is important. The last matching pattern has the most precedence.
 # So if a pull request only touches javascript files, only these owners
 # will be requested to review.
 #*.js    @octocat @github/js
 
 # You can also use email addresses if you prefer.
-#docs/*  docs@example.com
 
-# paddlepaddle
-**/paddlepaddle/** @kbinias @sfraczek @Sand3r- @lidanqing-intel @ddokupil @pmajchrzak @wojtuss
-**/PaddlePaddle/** @kbinias @sfraczek @Sand3r- @lidanqing-intel @ddokupil @pmajchrzak @wojtuss

README.md

@@ -3,8 +3,6 @@
 This repository contains **links to pre-trained models, sample scripts, best practices, and step-by-step tutorials** for many popular open-source machine learning models optimized by Intel to run on Intel® Xeon® Scalable processors. 
 
 Model packages and containers for running the Model Zoo's workloads can be found at the [Intel® oneContainer Portal](https://software.intel.com/containers).
-Intel Model Zoo is also bundled as a part of
-[Intel® oneAPI AI Analytics Toolkit](https://software.intel.com/content/www/us/en/develop/tools/oneapi/ai-analytics-toolkit.html) (AI Kit).
 
 ## Purpose of the Model Zoo
 
@@ -17,20 +15,21 @@ For any performance and/or benchmarking information on specific Intel platforms,
 
 ## How to Use the Model Zoo
 
-### Getting Started
-
+### Getting Started using AI Kit
+- The Intel Model Zoo is released as a part of the [Intel® AI Analytics Toolkit](https://software.intel.com/content/www/us/en/develop/tools/oneapi/ai-analytics-toolkit.html)
+  which provides a consolidated package of Intel’s latest deep and machine learning optimizations
+  all in one place for ease of development. Along with Model Zoo, the toolkit also includes Intel
+  optimized versions of deep learning frameworks (TensorFlow, PyTorch) and high performing Python
+  libraries to streamline end-to-end data science and AI workflows on Intel architectures.
+- The [documentation here](/docs/general/tensorflow/AIKit.md) has instructions on how to get to
+  the Model Zoo's conda environments and code directory within AI Kit.
+- There is a table of TensorFlow models with links to instructions on how to run the models [here](/benchmarks/README.md).
+- To get started you can refer to the [ResNet50 FP32 Inference code sample.](https://github.com/oneapi-src/oneAPI-samples/tree/master/AI-and-Analytics/Features-and-Functionality/IntelTensorFlow_ModelZoo_Inference_with_FP32_Int8)
+
+### Getting Started without AI Kit
 - If you know what model you are interested in, or if you want to see a full list of models in the Model Zoo, start **[here](/benchmarks)**.
 - For framework best practice guides, and step-by-step tutorials for some models in the Model Zoo, start **[here](/docs)**.
 
-- AI Kit provides a consolidated package of Intel’s latest deep and machine
-  learning optimizations all in one place for ease of development. Along with
-  Model Zoo,  the toolkit also includes Intel optimized versions of deep
-  learning frameworks (TensorFlow, PyTorch) and high performing Python libraries
-  to streamline end-to-end data science and AI workflows on Intel architectures.
-
-  |[Download AI Kit](https://software.intel.com/content/www/us/en/develop/tools/oneapi/ai-analytics-toolkit/) |[AI Kit Get Started Guide](https://software.intel.com/content/www/us/en/develop/documentation/get-started-with-ai-linux/top.html) |
-  |---|---|
-
 ### Directory Structure
 The Model Zoo is divided into four main directories:
 - **[benchmarks](/benchmarks)**: Look here for sample scripts and complete instructions on downloading and running each Intel-optimized pre-trained model.

benchmarks/common/base_model_init.py

@@ -21,6 +21,7 @@
 import glob
 import json
 import os
+import sys
 import time
 
 
@@ -93,7 +94,7 @@ def __init__(self, args, custom_args=[], platform_util=None):
                     + " --map-by ppr:" + str(pps) + ":socket:pe=" + split_a_socket + " --cpus-per-proc " \
                     + split_a_socket + " " + self.python_exe
 
-    def run_command(self, cmd):
+    def run_command(self, cmd, replace_unique_output_dir=None):
         """
         Prints debug messages when verbose is enabled, and then runs the
         specified command.
@@ -118,7 +119,8 @@ def run_command(self, cmd):
                       "the list of cpu nodes could not be retrieved. Please ensure "
                       "that your system has numa nodes and numactl is installed.")
             else:
-                self.run_numactl_multi_instance(cmd)
+                self.run_numactl_multi_instance(
+                    cmd, replace_unique_output_dir=replace_unique_output_dir)
         else:
             if self.args.verbose:
                 print("Running: {}".format(str(cmd)))
@@ -136,7 +138,7 @@ def group_cores(self, cpu_cores_list, cores_per_instance):
         end_list.append(cpu_cores_list[-count:]) if count != 0 else end_list
         return end_list
 
-    def run_numactl_multi_instance(self, cmd):
+    def run_numactl_multi_instance(self, cmd, replace_unique_output_dir=None):
         """
         Generates a series of commands that call the specified cmd with multiple
         instances, where each instance uses the a specified number of cores. The
@@ -195,7 +197,15 @@ def run_numactl_multi_instance(self, cmd):
                       "numactl --localalloc --physcpubind={1}").format(
                 len(core_list), ",".join(core_list))
             instance_logfile = log_filename_format.format("instance" + str(instance_num))
-            instance_command = "{} {}".format(prefix, cmd)
+
+            unique_command = cmd
+            if replace_unique_output_dir:
+                # Swap out the output dir for a unique dir
+                unique_dir = os.path.join(replace_unique_output_dir,
+                                          "instance_{}".format(instance_num))
+                unique_command = unique_command.replace(replace_unique_output_dir, unique_dir)
+
+            instance_command = "{} {}".format(prefix, unique_command)
             multi_instance_command += "{} >> {} 2>&1 & \\\n".format(
                 instance_command, instance_logfile)
             instance_logfiles.append(instance_logfile)
@@ -209,6 +219,7 @@ def run_numactl_multi_instance(self, cmd):
 
         # Run the multi-instance command
         print("\nMulti-instance run:\n" + multi_instance_command)
+        sys.stdout.flush()
         os.system(multi_instance_command)
 
         # Wait to ensure that log files have been written