intel
diff --git a/‎.gitignore
Lines changed: 1 addition & 0 deletions b/‎.gitignore
Lines changed: 1 addition & 0 deletions
diff --git a/‎CODEOWNERS
Lines changed: 5 additions & 3 deletions b/‎CODEOWNERS
Lines changed: 5 additions & 3 deletions
diff --git a/‎README.md
Lines changed: 42 additions & 16 deletions b/‎README.md
Lines changed: 42 additions & 16 deletions
diff --git a/‎benchmarks/README.md
Lines changed: 4 additions & 4 deletions b/‎benchmarks/README.md
Lines changed: 4 additions & 4 deletions
diff --git a/‎benchmarks/common/base_benchmark_util.py
Lines changed: 10 additions & 3 deletions b/‎benchmarks/common/base_benchmark_util.py
Lines changed: 10 additions & 3 deletions
diff --git a/‎benchmarks/common/tensorflow/start.sh
Lines changed: 52 additions & 12 deletions b/‎benchmarks/common/tensorflow/start.sh
Lines changed: 52 additions & 12 deletions
@@ -17,3 +17,4 @@ tools/docker/models*
 .ipynb_checkpoints
 nc_workspace
 benchmarks/horovod
+data_connector/credentials.json
@@ -3,12 +3,14 @@
 
 # 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
+* [email protected] @ashahba @claynerobison 
 datasets @ashahba @claynerobison @dzungductran
 docs @claynerobison @mhbuehler
-k8s  @ashahba @dzungductran @kkasravi
-models @agramesh1 @ashraf-bhuiyan @riverliuintel @wei-v-wang
+k8s  @ashahba @dzungductran
+models @ashraf-bhuiyan @riverliuintel
+models @riverliuintel
 models/**/pytorch/ @leslie-fang-intel @jiayisunx @zhuhaozhe
+quickstart [email protected]
 quickstart/**/pytorch/ @leslie-fang-intel @jiayisunx @zhuhaozhe
 
 # Order is important. The last matching pattern has the most precedence.
 
@@ -31,13 +31,13 @@ For information on running more advanced use cases using the workload containers
 | Image Segmentation      | [3D U-Net MLPerf*](https://arxiv.org/pdf/1606.06650.pdf) | Inference |  | [FP32 BFloat16 Int8](image_segmentation/tensorflow/3d_unet_mlperf/inference/README.md) | [BRATS 2019](https://www.med.upenn.edu/cbica/brats2019/data.html) |
 | Image Segmentation      | [MaskRCNN](https://arxiv.org/abs/1703.06870) | Inference | Model Containers: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/mask-rcnn-fp32-inference-tensorflow-container.html) <br> Model Packages: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/mask-rcnn-fp32-inference-tensorflow-model.html)  | [FP32](image_segmentation/tensorflow/maskrcnn/inference/fp32/README.md) | [MS COCO 2014](https://github.com/IntelAI/models/tree/master/benchmarks/image_segmentation/tensorflow/maskrcnn/inference/fp32#datasets-and-pretrained-model) |
 | Image Segmentation      | [UNet](https://arxiv.org/pdf/1606.06650.pdf) | Inference | Model Containers: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/unet-fp32-inference-tensorflow-container.html) <br> Model Packages: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/unet-fp32-inference-tensorflow-model.html)  | [FP32](image_segmentation/tensorflow/unet/inference/fp32/README.md) |
-| Language Modeling       | [BERT](https://arxiv.org/pdf/1810.04805.pdf) | Inference | Model Containers: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-fp32-inference-tensorflow-container.html) [BFloat16](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-bfloat16-inference-tensorflow-container.html) <br> Model Packages: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-fp32-inference-tensorflow-model.html) [BFloat16](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-bfloat16-inference-tensorflow-model.html) |  [FP32](language_modeling/tensorflow/bert_large/inference/fp32/README.md) [BFloat16](language_modeling/tensorflow/bert_large/inference/bfloat16/README.md) | [SQuAD](https://github.com/IntelAI/models/tree/master/datasets/bert_data/README.md#inference) |
-| Language Modeling       | [BERT](https://arxiv.org/pdf/1810.04805.pdf) | Training | Model Containers: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-fp32-training-tensorflow-container.html) [BFloat16](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-bfloat16-training-tensorflow-container.html) <br> Model Packages: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-fp32-training-tensorflow-model.html) [BFloat16](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-bfloat16-training-tensorflow-model.html)  | [FP32](language_modeling/tensorflow/bert_large/training/fp32/README.md) [BFloat16](language_modeling/tensorflow/bert_large/training/bfloat16/README.md) | [SQuAD](https://github.com/IntelAI/models/tree/master/datasets/bert_data/README.md#fine-tuning-with-bert-using-squad-data) and [MRPC](https://github.com/IntelAI/models/tree/master/datasets/bert_data/README.md#classification-training-with-bert) |
+| Language Modeling       | [BERT](https://arxiv.org/pdf/1810.04805.pdf) | Inference | Model Containers: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-fp32-inference-tensorflow-container.html) [BFloat16](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-bfloat16-inference-tensorflow-container.html) <br> Model Packages: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-fp32-inference-tensorflow-model.html) [BFloat16](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-bfloat16-inference-tensorflow-model.html) | [Int8](language_modeling/tensorflow/bert_large/inference/int8/README.md) [FP32](language_modeling/tensorflow/bert_large/inference/fp32/README.md) [BFloat16](language_modeling/tensorflow/bert_large/inference/bfloat16/README.md) | [SQuAD](https://github.com/IntelAI/models/tree/master/datasets/bert_data/README.md#inference) |
+| Language Modeling       | [BERT](https://arxiv.org/pdf/1810.04805.pdf) | Training | Model Containers: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-fp32-training-tensorflow-container.html) [BFloat16](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-bfloat16-training-tensorflow-container.html) <br> Model Packages: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-fp32-training-tensorflow-model.html) [BFloat16](https://software.intel.com/content/www/us/en/develop/articles/containers/bert-large-bfloat16-training-tensorflow-model.html)  | [FP32](language_modeling/tensorflow/bert_large/training/fp32/Advanced.md) [BFloat16](language_modeling/tensorflow/bert_large/training/bfloat16/Advanced.md) [FP16](language_modeling/tensorflow/bert_large/training/fp16/Advanced.md) | [SQuAD](https://github.com/IntelAI/models/tree/master/datasets/bert_data/README.md#fine-tuning-with-bert-using-squad-data) and [MRPC](https://github.com/IntelAI/models/tree/master/datasets/bert_data/README.md#classification-training-with-bert) |
 | Language Modeling | [distilBERT](https://arxiv.org/abs/1910.01108) | Inference | Model Containers: | [FP32 BFloat16](https://github.com/IntelAI/models/benchmarks/language_modeling/tensorflow/distilbert_base/inference/README.md) | [SST-2](https://huggingface.co/datasets/sst2) |
 | Language Translation    | [BERT](https://arxiv.org/pdf/1810.04805.pdf) | Inference |  | [FP32](language_translation/tensorflow/bert/inference/README.md) | [MRPC](https://github.com/IntelAI/models/tree/master/datasets/bert_data/README.md#classification-training-with-bert) |
 | Language Translation    | [GNMT*](https://arxiv.org/pdf/1609.08144.pdf) | Inference | Model Containers: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/gnmt-fp32-inference-tensorflow-container.html) <br> Model Packages: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/gnmt-fp32-inference-tensorflow-model.html) | [FP32](language_translation/tensorflow/mlperf_gnmt/inference/README.md) | [MLPerf GNMT model benchmarking dataset](https://github.com/IntelAI/models/tree/master/benchmarks/language_translation/tensorflow/mlperf_gnmt/inference/fp32#datasets) |
-| Language Translation    | [Transformer_LT_mlperf*](https://arxiv.org/pdf/1706.03762.pdf) | Training | Model Containers: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/transformer-lt-mlperf-fp32-training-tensorflow-container.html) [BFloat16](https://software.intel.com/content/www/us/en/develop/articles/containers/transformer-lt-mlperf-bfloat16-training-tensorflow-container.html) <br> Model Packages: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/transformer-lt-mlperf-fp32-training-tensorflow-model.html) [BFloat16](https://software.intel.com/content/www/us/en/develop/articles/containers/transformer-lt-mlperf-bfloat16-training-tensorflow-model.html) | [FP32](language_translation/tensorflow/transformer_mlperf/training/fp32/README.md) [BFloat16](language_translation/tensorflow/transformer_mlperf/training/bfloat16/README.md) | [WMT English-German dataset](https://github.com/IntelAI/models/tree/master/datasets/transformer_data#transformer-language-mlperf-dataset) |
-| Language Translation    | [Transformer_LT_mlperf*](https://arxiv.org/pdf/1706.03762.pdf) | Inference | | [FP32](language_translation/tensorflow/transformer_mlperf/inference/fp32/README.md) [BFloat16](language_translation/tensorflow/transformer_mlperf/inference/bfloat16/README.md) [Int8](language_translation/tensorflow/transformer_mlperf/inference/int8/README.md) | [WMT English-German data](https://github.com/IntelAI/models/tree/master/datasets/transformer_data#transformer-language-mlperf-dataset) |
+| Language Translation    | [Transformer_LT_mlperf*](https://arxiv.org/pdf/1706.03762.pdf) | Training | Model Containers: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/transformer-lt-mlperf-fp32-training-tensorflow-container.html) [BFloat16](https://software.intel.com/content/www/us/en/develop/articles/containers/transformer-lt-mlperf-bfloat16-training-tensorflow-container.html) <br> Model Packages: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/transformer-lt-mlperf-fp32-training-tensorflow-model.html) [BFloat16](https://software.intel.com/content/www/us/en/develop/articles/containers/transformer-lt-mlperf-bfloat16-training-tensorflow-model.html) | [FP32 BFloat16](language_translation/tensorflow/transformer_mlperf/training/README.md) | [WMT English-German dataset](https://github.com/IntelAI/models/tree/master/datasets/transformer_data#transformer-language-mlperf-dataset) |
+| Language Translation    | [Transformer_LT_mlperf*](https://arxiv.org/pdf/1706.03762.pdf) | Inference | | [FP32 BFloat16  Int8](language_translation/tensorflow/transformer_mlperf/inference/README.md) | [WMT English-German data](https://github.com/IntelAI/models/tree/master/datasets/transformer_data#transformer-language-mlperf-dataset) |
 | Language Translation    | [Transformer_LT_Official](https://arxiv.org/pdf/1706.03762.pdf) | Inference | Model Containers: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/transformer-lt-official-fp32-inference-tensorflow-container.html) <br> Model Packages: [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/transformer-lt-official-fp32-inference-tensorflow-model.html)  | [FP32](language_translation/tensorflow/transformer_lt_official/inference/README.md) | [WMT English-German dataset](https://github.com/IntelAI/models/tree/master/datasets/transformer_data#transformer-language-mlperf-dataset) |
 | Object Detection        | [Faster R-CNN](https://arxiv.org/pdf/1506.01497.pdf) | Inference | Model Containers: [Int8](https://software.intel.com/content/www/us/en/develop/articles/containers/faster-rcnn-int8-inference-tensorflow-container.html) [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/faster-rcnn-fp32-inference-tensorflow-container.html) <br> Model Packages: [Int8](https://software.intel.com/content/www/us/en/develop/articles/containers/faster-rcnn-int8-inference-tensorflow-model.html) [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/faster-rcnn-fp32-inference-tensorflow-model.html)  | [Int8](object_detection/tensorflow/faster_rcnn/inference/int8/README.md) [FP32](object_detection/tensorflow/faster_rcnn/inference/fp32/README.md) | [COCO 2017 validation dataset](https://github.com/IntelAI/models/tree/master/datasets/coco#download-and-preprocess-the-coco-validation-images) |
 | Object Detection        | [R-FCN](https://arxiv.org/pdf/1605.06409.pdf) | Inference | Model Containers: [Int8](https://software.intel.com/content/www/us/en/develop/articles/containers/rfcn-int8-inference-tensorflow-container.html) [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/rfcn-fp32-inference-tensorflow-container.html) <br> Model Packages: [Int8](https://software.intel.com/content/www/us/en/develop/articles/containers/rfcn-int8-inference-tensorflow-model.html) [FP32](https://software.intel.com/content/www/us/en/develop/articles/containers/rfcn-fp32-inference-tensorflow-model.html)  | [Int8 FP32](object_detection/tensorflow/rfcn/inference/README.md) | [COCO 2017 validation dataset](https://github.com/IntelAI/models/tree/master/datasets/coco#download-and-preprocess-the-coco-validation-images) |
 
@@ -1,7 +1,7 @@
 #
 # -*- coding: utf-8 -*-
 #
-# Copyright (c) 2018-2021 Intel Corporation
+# Copyright (c) 2018-2023 Intel Corporation
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -68,8 +68,8 @@ def _define_args(self):
 
         self._common_arg_parser.add_argument(
             "-p", "--precision",
-            help="Specify the model precision to use: fp32, fp16, int8, or bfloat16",
-            required=required_arg, choices=["fp32", "fp16", "int8", "bfloat16"],
+            help="Specify the model precision to use: fp32, int8, bfloat16 or fp16",
+            required=required_arg, choices=["fp32", "int8", "bfloat16", "fp16"],
             dest="precision")
 
         self._common_arg_parser.add_argument(
@@ -228,6 +228,13 @@ def _define_args(self):
             dest="experimental_gelu", choices=["True", "False"],
             default=False)
 
+        self._common_arg_parser.add_argument(
+            "--amp",
+            help="use grappler auto-mixed precision as opposed to \
+            keras mixed precision",
+            dest="amp", choices=["True", "False"],
+            default=False)
+
         # Note this can't be a normal boolean flag, because we need to know when the user
         # does not explicitly set the arg value so that we can apply the appropriate
         # default value, depending on the the precision.
 
@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 #
-# Copyright (c) 2018-2019 Intel Corporation
+# Copyright (c) 2018-2023 Intel Corporation
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -160,7 +160,7 @@ if [[ ${NOINSTALL} != "True" ]]; then
       export HOROVOD_WITHOUT_PYTORCH=1
       export HOROVOD_WITHOUT_MXNET=1
       export HOROVOD_WITH_TENSORFLOW=1
-      export HOROVOD_VERSION=35b27e9
+      export HOROVOD_VERSION=b1d0ce8
 
       # Install GCC 7 from devtoolset-7
       if [[ ${OS_VERSION} =~ "7".* ]]; then
@@ -239,7 +239,16 @@ if [[ ${NOINSTALL} != "True" ]]; then
       # In case installing released versions of Horovod fail,and there is
       # a working commit replace next set of commands with something like:
       apt-get install -y --no-install-recommends --fix-missing cmake git
-      python3 -m pip install --no-cache-dir git+https://github.com/horovod/horovod.git@${HOROVOD_VERSION}
+      # TODO: Once this PR https://github.com/horovod/horovod/pull/3864 is merged, we can install horovod as before.
+      # python3 -m pip install --no-cache-dir git+https://github.com/horovod/horovod.git@${HOROVOD_VERSION}
+      git clone https://github.com/horovod/horovod.git
+      cd horovod
+      git reset --hard ${HOROVOD_VERSION}
+      git submodule update --init --recursive
+      git fetch origin pull/3864/head:ashahba/issue-3861-fix
+      git checkout ashahba/issue-3861-fix
+      python3 -m pip install --no-cache-dir -v -e .
+
       horovodrun --check-build
     fi
   fi
@@ -559,6 +568,9 @@ function bert_options() {
   if [[ -n "${OPTIMIZED_SOFTMAX}" && ${OPTIMIZED_SOFTMAX} != "" ]]; then
     CMD=" ${CMD} --optimized-softmax=${OPTIMIZED_SOFTMAX}"
   fi
+  if [[ -n "${AMP}" && ${AMP} != "" ]]; then
+    CMD=" ${CMD} --amp=${AMP}"
+  fi
 
   if [[ -n "${MPI_WORKERS_SYNC_GRADIENTS}" && ${MPI_WORKERS_SYNC_GRADIENTS} != "" ]]; then
     CMD=" ${CMD} --mpi_workers_sync_gradients=${MPI_WORKERS_SYNC_GRADIENTS}"
@@ -991,7 +1003,11 @@ function resnet101_inceptionv3() {
 function resnet50() {
     export PYTHONPATH=${PYTHONPATH}:$(pwd):${MOUNT_BENCHMARK}
     is_model_gpu_supported="True"
-
+    if [ ${GPU} == "True" ]; then
+      PYTHONPATH=${PYTHONPATH}:${MOUNT_INTELAI_MODELS_SOURCE}/${MODE}/gpu
+    else
+      PYTHONPATH=${PYTHONPATH}:${MOUNT_INTELAI_MODELS_SOURCE}/${MODE}/cpu
+    fi
     # For accuracy, dataset location is required.
     if [ "${DATASET_LOCATION_VOL}" == "None" ] && [ ${ACCURACY_ONLY} == "True" ]; then
       echo "No Data directory specified, accuracy will not be calculated."
@@ -1002,10 +1018,6 @@ function resnet50() {
         CMD="${CMD} $(add_steps_args) $(add_calibration_arg)"
         PYTHONPATH=${PYTHONPATH} CMD=${CMD} run_model
     elif [ ${PRECISION} == "fp32" ] || [ ${PRECISION} == "bfloat16" ] || [ ${PRECISION} == "fp16" ]; then
-        if  [ ${PRECISION} == "fp16" ] && [ ${GPU} == "False" ]; then
-          echo "PRECISION=${PRECISION} is not supported without --gpu."
-          exit 1
-        fi
       CMD="${CMD} $(add_steps_args)"
       PYTHONPATH=${PYTHONPATH} CMD=${CMD} run_model
     else
@@ -1187,7 +1199,7 @@ function ssd-resnet34() {
           $(add_arg "--num_warmup_batches" ${NUM_WARMUP_BATCHES})"
           local old_pythonpath=${PYTHONPATH}
           export PYTHONPATH=${PYTHONPATH}:${MOUNT_EXTERNAL_MODELS_SOURCE}
-          export PYTHONPATH=${PYTHONPATH}:${TF_MODELS_DIR}:${TF_MODELS_DIR}/research:"/tmp/benchmark_ssd_resnet34/scripts/tf_cnn_benchmarks"
+          export PYTHONPATH=${PYTHONPATH}:${MOUNT_EXTERNAL_MODELS_SOURCE}/research:"/tmp/benchmark_ssd_resnet34/scripts/tf_cnn_benchmarks"
           CMD=${CMD} run_model
           PYTHONPATH=${old_pythonpath}
         else
@@ -1466,8 +1478,7 @@ function bert_large() {
       bert_options
       CMD=${CMD} run_model
     else
-      # Change if to support fp32
-      if [ ${PRECISION} == "fp32" ]  || [ ${PRECISION} == "int8" ] || [ ${PRECISION} == "bfloat16" ]; then
+      if [ ${PRECISION} == "fp32" ]  || [ $PRECISION == "int8" ] || [ $PRECISION == "bfloat16" ] || [ $PRECISION == "fp16" ]; then
         export PYTHONPATH=${PYTHONPATH}:${MOUNT_EXTERNAL_MODELS_SOURCE}
         bert_options
         CMD=${CMD} run_model
@@ -1507,6 +1518,35 @@ function distilbert_base() {
     fi
 }
 
+# distilBERT base model
+function distilbert_base() {
+    if ([ ${PRECISION} == "fp32" ] || [ ${PRECISION} == "bfloat16" ] ||
+       [ ${PRECISION} == "int8" ] || [ ${PRECISION} == "fp16" ]); then
+      export PYTHONPATH=${PYTHONPATH}:${MOUNT_EXTERNAL_MODELS_SOURCE}
+      CMD="${CMD} $(add_arg "--warmup-steps" ${WARMUP_STEPS})"
+      CMD="${CMD} $(add_arg "--steps" ${STEPS})"
+
+      if [ ${NUM_INTER_THREADS} != "None" ]; then
+        CMD="${CMD} $(add_arg "--num-inter-threads" ${NUM_INTER_THREADS})"
+      fi
+
+      if [ ${NUM_INTRA_THREADS} != "None" ]; then
+        CMD="${CMD} $(add_arg "--num-intra-threads" ${NUM_INTRA_THREADS})"
+      fi
+
+      if [ -z ${STEPS} ]; then
+        CMD="${CMD} $(add_arg "--steps" ${STEPS})"
+      fi
+
+      if [ -z $MAX_SEQ_LENGTH ]; then
+        CMD="${CMD} $(add_arg "--max-seq-length" ${MAX_SEQ_LENGTH})"
+      fi
+      CMD=${CMD} run_model
+    else
+      echo "PRECISION=${PRECISION} not supported for ${MODEL_NAME} in this repo."
+      exit 1
+    fi
+}
 
 # Wide & Deep model
 function wide_deep() {
@@ -1650,7 +1690,7 @@ elif [ ${MODEL_NAME} == "bert_large" ]; then
 elif [ ${MODEL_NAME} == "dien" ]; then
   dien
 elif [ ${MODEL_NAME} == "distilbert_base" ]; then
-  distilbert_base 
+  distilbert_base  
 else
   echo "Unsupported model: ${MODEL_NAME}"
   exit 1