From 10f14815dce23ff68b66fda03ef070ee349da848 Mon Sep 17 00:00:00 2001
From: Arjun Suresh <arjunsuresh1987@gmail.com>
Date: Tue, 24 Dec 2024 14:56:55 +0530
Subject: [PATCH 1/3] Fixed stable-diffusion-xl name in SUT configs

---
 .../pytorch-framework/default-config.yaml     | 38 -------------------
 .../tensorrt-framework/default-config.yaml    | 30 ---------------
 .../default-config.yaml                       |  5 +--
 .../tensorrt-framework/default-config.yaml    | 38 -------------------
 .../default-config.yaml                       |  2 +-
 5 files changed, 3 insertions(+), 110 deletions(-)
 delete mode 100644 script/get-mlperf-inference-sut-configs/configs/IntelSPR.24c/intel-implementation/cpu-device/pytorch-framework/default-config.yaml
 delete mode 100644 script/get-mlperf-inference-sut-configs/configs/RTX4090x1/nvidia_original-implementation/gpu-device/tensorrt-framework/default-config.yaml
 delete mode 100644 script/get-mlperf-inference-sut-configs/configs/RTX4090x2/nvidia_original-implementation/gpu-device/tensorrt-framework/default-config.yaml

diff --git a/script/get-mlperf-inference-sut-configs/configs/IntelSPR.24c/intel-implementation/cpu-device/pytorch-framework/default-config.yaml b/script/get-mlperf-inference-sut-configs/configs/IntelSPR.24c/intel-implementation/cpu-device/pytorch-framework/default-config.yaml
deleted file mode 100644
index 040ccbdcc..000000000
--- a/script/get-mlperf-inference-sut-configs/configs/IntelSPR.24c/intel-implementation/cpu-device/pytorch-framework/default-config.yaml
+++ /dev/null
@@ -1,38 +0,0 @@
----
-  resnet50:
-    Offline:
-      target_qps: 1000.0
-    Server:
-      target_qps: 500.0
-  retinanet:
-    Offline:
-      target_qps: 50.0
-    Server:
-      target_qps: 30
-  bert-99:
-    Offline:
-      target_qps: 100
-  bert-99.9:
-    Offline:
-      target_qps: 100
-  3d-unet-99:
-    Offline:
-      target_qps: 1.0
-  3d-unet-99.9:
-    Offline:
-      target_qps: 1.0
-  gptj-99.9:
-    Offline:
-      target_qps: 0.5
-    Server:
-      target_qps: 0.3
-  gptj-99:
-    Offline:
-      target_qps: 0.5
-    Server:
-      target_qps: 0.3
-  sdxl:
-    Offline:
-      target_qps: 0.1
-    Server:
-      target_qps: 0.1
diff --git a/script/get-mlperf-inference-sut-configs/configs/RTX4090x1/nvidia_original-implementation/gpu-device/tensorrt-framework/default-config.yaml b/script/get-mlperf-inference-sut-configs/configs/RTX4090x1/nvidia_original-implementation/gpu-device/tensorrt-framework/default-config.yaml
deleted file mode 100644
index f7a8477a8..000000000
--- a/script/get-mlperf-inference-sut-configs/configs/RTX4090x1/nvidia_original-implementation/gpu-device/tensorrt-framework/default-config.yaml
+++ /dev/null
@@ -1,30 +0,0 @@
-3d-unet-99:
-  Offline:
-    target_qps: 4.0
-  SingleStream:
-    target_latency: 400
-3d-unet-99.9:
-  Offline:
-    target_qps: 4.0
-  SingleStream:
-    target_latency: 400
-bert-99:
-  Offline:
-    target_qps: 4000.0
-  Server:
-    target_qps: 3800.0
-bert-99.9:
-  Offline:
-    target_qps: 2000.0
-  Server:
-    target_qps: 1600.0
-resnet50:
-  Offline:
-    target_qps: '42959.4'
-  Server:
-    target_qps: 35000.0
-retinanet:
-  Offline:
-    target_qps: 850.0
-  Server:
-    target_qps: 630.0
diff --git a/script/get-mlperf-inference-sut-configs/configs/RTX4090x1/nvidia_original-implementation/gpu-device/tensorrt-framework/framework-version-default/default-config.yaml b/script/get-mlperf-inference-sut-configs/configs/RTX4090x1/nvidia_original-implementation/gpu-device/tensorrt-framework/framework-version-default/default-config.yaml
index a9ad05a50..d764ab24d 100644
--- a/script/get-mlperf-inference-sut-configs/configs/RTX4090x1/nvidia_original-implementation/gpu-device/tensorrt-framework/framework-version-default/default-config.yaml
+++ b/script/get-mlperf-inference-sut-configs/configs/RTX4090x1/nvidia_original-implementation/gpu-device/tensorrt-framework/framework-version-default/default-config.yaml
@@ -28,9 +28,8 @@ retinanet:
     target_qps: 850.0
   Server:
     target_qps: 630.0
-sdxl:
+stable-diffusion-xl:
   Offline:
     target_qps: 0.7
   Server:
-    target_qps: 0.3
-  
+    target_qps: 0.3  
diff --git a/script/get-mlperf-inference-sut-configs/configs/RTX4090x2/nvidia_original-implementation/gpu-device/tensorrt-framework/default-config.yaml b/script/get-mlperf-inference-sut-configs/configs/RTX4090x2/nvidia_original-implementation/gpu-device/tensorrt-framework/default-config.yaml
deleted file mode 100644
index 7b24138ac..000000000
--- a/script/get-mlperf-inference-sut-configs/configs/RTX4090x2/nvidia_original-implementation/gpu-device/tensorrt-framework/default-config.yaml
+++ /dev/null
@@ -1,38 +0,0 @@
----
-  resnet50:
-    Offline:
-      target_qps: 43000.0
-    Server:
-      target_qps: 37000.0
-  retinanet:
-    Offline:
-      target_qps: 650.0
-    Server:
-      target_qps: 600
-  bert-99:
-    Offline:
-      target_qps: 4000
-  bert-99.9:
-    Offline:
-      target_qps: 4000
-  3d-unet-99:
-    Offline:
-      target_qps: 2.0
-  3d-unet-99.9:
-    Offline:
-      target_qps: 2.0
-  gptj-99.9:
-    Offline:
-      target_qps: 4
-    Server:
-      target_qps: 3.5
-  gptj-99:
-    Offline:
-      target_qps: 4
-    Server:
-      target_qps: 3.5
-  sdxl:
-    Offline:
-      target_qps: 2
-    Server:
-      target_qps: 1
diff --git a/script/get-mlperf-inference-sut-configs/configs/RTX4090x2/nvidia_original-implementation/gpu-device/tensorrt-framework/framework-version-default/default-config.yaml b/script/get-mlperf-inference-sut-configs/configs/RTX4090x2/nvidia_original-implementation/gpu-device/tensorrt-framework/framework-version-default/default-config.yaml
index 4820e8b52..294b2eda7 100644
--- a/script/get-mlperf-inference-sut-configs/configs/RTX4090x2/nvidia_original-implementation/gpu-device/tensorrt-framework/framework-version-default/default-config.yaml
+++ b/script/get-mlperf-inference-sut-configs/configs/RTX4090x2/nvidia_original-implementation/gpu-device/tensorrt-framework/framework-version-default/default-config.yaml
@@ -35,7 +35,7 @@
       target_qps: 8
     Server:
       target_qps: 7
-  sdxl:
+  stable-diffusion-xl:
     Offline:
       target_qps: 1.3
     Server:

From a10c616b40671a2ba7c4a7d24296913c4bc65cd1 Mon Sep 17 00:00:00 2001
From: Arjun Suresh <arjunsuresh1987@gmail.com>
Date: Tue, 24 Dec 2024 13:19:04 +0000
Subject: [PATCH 2/3] Fix tensorrt detect on aarch64

---
 script/get-tensorrt/customize.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/script/get-tensorrt/customize.py b/script/get-tensorrt/customize.py
index addc7322c..263c645e8 100644
--- a/script/get-tensorrt/customize.py
+++ b/script/get-tensorrt/customize.py
@@ -13,7 +13,7 @@ def preprocess(i):
 
     # Not enforcing dev requirement for now
     if env.get('CM_TENSORRT_TAR_FILE_PATH', '') == '' and env.get(
-            'CM_TENSORRT_REQUIRE_DEV1', '') != 'yes' and env.get('CM_HOST_PLATFORM_FLAVOR', '') != 'aarch64':
+            'CM_TENSORRT_REQUIRE_DEV1', '') != 'yes' and env.get('CM_HOST_PLATFORM_FLAVOR_', '') != 'aarch64':
 
         if os_info['platform'] == 'windows':
             extra_pre = ''

From 83d9d2d4f005ec76dfcac8b5ec906b9bab13f20d Mon Sep 17 00:00:00 2001
From: Arjun Suresh <arjunsuresh1987@gmail.com>
Date: Tue, 24 Dec 2024 20:38:22 +0530
Subject: [PATCH 3/3] Added torch deps for get-ml-model-gptj-nvidia

---
 script/get-ml-model-gptj/_cm.yaml      | 4 ++++
 script/get-ml-model-gptj/run-nvidia.sh | 2 +-
 2 files changed, 5 insertions(+), 1 deletion(-)

diff --git a/script/get-ml-model-gptj/_cm.yaml b/script/get-ml-model-gptj/_cm.yaml
index 7d645976a..25b2ef981 100644
--- a/script/get-ml-model-gptj/_cm.yaml
+++ b/script/get-ml-model-gptj/_cm.yaml
@@ -161,6 +161,10 @@ variations:
       - python3
       tags: get,python3
     - tags: get,generic-python-lib,_package.safetensors
+    - tags: get,generic-python-lib,_torch
+      names:
+      - torch
+      - pytorch
   rclone:
     add_deps_recursive:
       dae:
diff --git a/script/get-ml-model-gptj/run-nvidia.sh b/script/get-ml-model-gptj/run-nvidia.sh
index 27e5a675c..b16ee45da 100644
--- a/script/get-ml-model-gptj/run-nvidia.sh
+++ b/script/get-ml-model-gptj/run-nvidia.sh
@@ -17,5 +17,5 @@ export DOCKER_RUN_ARGS=" -v ${CM_NVIDIA_MLPERF_SCRATCH_PATH}:/mnt"
 make -C docker run LOCAL_USER=1
 test $? -eq 0 || exit $?
 
-${CM_PYTHON_BIN_WITH_PATH} ${CM_MLPERF_INFERENCE_NVIDIA_CODE_PATH}/code/gptj/tensorrt/onnx_tune.py --fp8-scalers-path=${CM_NVIDIA_MLPERF_SCRATCH_PATH}/models/GPTJ-6B/fp8-quantized-ammo/GPTJ-FP8-quantized/rank0.safetensors --scaler 1.005 --index 15
+PYTHONPATH='' ${CM_PYTHON_BIN_WITH_PATH} ${CM_MLPERF_INFERENCE_NVIDIA_CODE_PATH}/code/gptj/tensorrt/onnx_tune.py --fp8-scalers-path=${CM_NVIDIA_MLPERF_SCRATCH_PATH}/models/GPTJ-6B/fp8-quantized-ammo/GPTJ-FP8-quantized/rank0.safetensors --scaler 1.005 --index 15
 test $? -eq 0 || exit $?