Update file names

build.py

@@ -74,11 +74,11 @@
     "release_version": "2.60.0dev",
     "triton_container_version": "25.08dev",
     "upstream_container_version": "25.07",
-    "ort_version": "1.22.0",
+    "ort_version": "1.23.0",
     "ort_openvino_version": "2025.2.0",
     "standalone_openvino_version": "2025.2.0",
     "dcgm_version": "4.2.3-2",
-    "vllm_version": "0.9.0.1",
+    "vllm_version": "0.9.2",
     "rhel_py_version": "3.12.3",
 }
 
@@ -1259,7 +1259,7 @@ def create_dockerfile_linux(
     # stage of the PyTorch backend
     if not FLAGS.enable_gpu and ("pytorch" in backends):
         df += """
-RUN patchelf --add-needed /usr/local/cuda/lib64/stubs/libcublasLt.so.12 backends/pytorch/libtorch_cuda.so
+RUN patchelf --add-needed /usr/local/cuda/lib64/stubs/libcublasLt.so.13 backends/pytorch/libtorch_cuda.so
 """
     if "tensorrtllm" in backends:
         df += """
@@ -1494,7 +1494,7 @@ def dockerfile_prepare_container_linux(argmap, backends, enable_gpu, target_mach
             cp -r nvpl_slim_24.04/include/* /usr/local/include && \\
             rm -rf nvpl_slim_24.04.tar nvpl_slim_24.04; \\
         fi \\
-        && pip3 install --no-cache-dir --progress-bar on --index-url $VLLM_INDEX_URL -r /run/secrets/requirements \\
+        && pip3 install --no-cache-dir --extra-index-url $VLLM_INDEX_URL -r /run/secrets/requirements \\
         # Need to install in-house build of pytorch-triton to support triton_key definition used by torch 2.5.1
         && cd /tmp \\
         && wget $PYTORCH_TRITON_URL \\
@@ -1554,18 +1554,18 @@ def add_cpu_libs_to_linux_dockerfile(backends, target_machine):
         df += """
 RUN mkdir -p /usr/local/cuda/lib64/stubs
 COPY --from=min_container /usr/local/cuda/lib64/stubs/libcusparse.so /usr/local/cuda/lib64/stubs/libcusparse.so.12
-COPY --from=min_container /usr/local/cuda/lib64/stubs/libcusolver.so /usr/local/cuda/lib64/stubs/libcusolver.so.11
+COPY --from=min_container /usr/local/cuda/lib64/stubs/libcusolver.so /usr/local/cuda/lib64/stubs/libcusolver.so.12
 COPY --from=min_container /usr/local/cuda/lib64/stubs/libcurand.so /usr/local/cuda/lib64/stubs/libcurand.so.10
-COPY --from=min_container /usr/local/cuda/lib64/stubs/libcufft.so /usr/local/cuda/lib64/stubs/libcufft.so.11
-COPY --from=min_container /usr/local/cuda/lib64/stubs/libcublas.so /usr/local/cuda/lib64/stubs/libcublas.so.12
-COPY --from=min_container /usr/local/cuda/lib64/stubs/libcublasLt.so /usr/local/cuda/lib64/stubs/libcublasLt.so.12
-COPY --from=min_container /usr/local/cuda/lib64/stubs/libcublasLt.so /usr/local/cuda/lib64/stubs/libcublasLt.so.11
+COPY --from=min_container /usr/local/cuda/lib64/stubs/libcufft.so /usr/local/cuda/lib64/stubs/libcufft.so.12
+COPY --from=min_container /usr/local/cuda/lib64/stubs/libcublas.so /usr/local/cuda/lib64/stubs/libcublas.so.13
+COPY --from=min_container /usr/local/cuda/lib64/stubs/libcublasLt.so /usr/local/cuda/lib64/stubs/libcublasLt.so.13
 
 RUN mkdir -p /usr/local/cuda/targets/{cuda_arch}-linux/lib
-COPY --from=min_container /usr/local/cuda/lib64/libcudart.so.12 /usr/local/cuda/targets/{cuda_arch}-linux/lib/.
-COPY --from=min_container /usr/local/cuda/lib64/libcupti.so.12 /usr/local/cuda/targets/{cuda_arch}-linux/lib/.
-COPY --from=min_container /usr/local/cuda/lib64/libnvJitLink.so.12 /usr/local/cuda/targets/{cuda_arch}-linux/lib/.
+COPY --from=min_container /usr/local/cuda/lib64/libcudart.so.13 /usr/local/cuda/targets/{cuda_arch}-linux/lib/.
+COPY --from=min_container /usr/local/cuda/lib64/libcupti.so.13 /usr/local/cuda/targets/{cuda_arch}-linux/lib/.
+COPY --from=min_container /usr/local/cuda/lib64/libnvJitLink.so.13 /usr/local/cuda/targets/{cuda_arch}-linux/lib/.
 COPY --from=min_container /usr/local/cuda/lib64/libcufile.so.0 /usr/local/cuda/targets/{cuda_arch}-linux/lib/.
+COPY --from=min_container /usr/local/cuda/lib64/libnvrtc.so.13 /usr/local/cuda/targets/{cuda_arch}-linux/lib/.
 
 RUN mkdir -p /opt/hpcx/ucc/lib/ /opt/hpcx/ucx/lib/
 COPY --from=min_container /opt/hpcx/ucc/lib/libucc.so.1 /opt/hpcx/ucc/lib/libucc.so.1

qa/common/gen_jetson_trt_models

@@ -142,6 +142,7 @@ docker pull $TENSORRT_IMAGE
 
 docker run $DOCKER_GPU_ARGS \
    --rm -v $DOCKER_VOLUME:/mnt \
+   -e TRT_VERBOSE \
   $TENSORRT_IMAGE bash -xe $VOLUME_SRCDIR/$TRT_MODEL_SCRIPT
 
 # Copy generated models to /tmp/ if not running in CI

qa/common/gen_qa_dyna_sequence_implicit_models.py

@@ -357,7 +357,11 @@ def create_onnx_modelconfig(models_dir, model_version, max_batch, dtype, shape):
 
 def create_plan_modelfile(models_dir, model_version, max_batch, dtype, shape):
     trt_dtype = np_to_trt_dtype(dtype)
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 
@@ -492,7 +496,11 @@ def create_plan_rf_modelfile(models_dir, model_version, max_batch, dtype, shape)
     trt_dtype = np_to_trt_dtype(dtype)
     trt_memory_format = trt.TensorFormat.LINEAR
 
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 

qa/common/gen_qa_dyna_sequence_models.py

@@ -59,7 +59,11 @@ def create_plan_shape_tensor_modelfile(
     trt_shape_dtype = np_to_trt_dtype(shape_tensor_input_dtype)
     trt_memory_format = trt.TensorFormat.LINEAR
 
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 
@@ -202,7 +206,11 @@ def create_plan_modelfile(models_dir, model_version, max_batch, dtype, shape):
     # Create the model. For now don't implement a proper accumulator
     # just return 0 if not-ready and 'INPUT'+'START'*('END'*'CORRID')
     # otherwise...  the tests know to expect this.
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 
@@ -310,7 +318,11 @@ def create_plan_rf_modelfile(models_dir, model_version, max_batch, dtype, shape)
     # Create the model. For now don't implement a proper accumulator
     # just return 0 if not-ready and 'INPUT'+'START'*('END'*'CORRID')
     # otherwise...  the tests know to expect this.
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 

qa/common/gen_qa_identity_models.py

@@ -545,7 +545,11 @@ def create_plan_dynamic_rf_modelfile(
     models_dir, model_version, io_cnt, max_batch, dtype, shape, profile_max_size
 ):
     # Create the model
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 
@@ -644,7 +648,11 @@ def create_plan_shape_tensor_modelfile(
     # Note that values of OUTPUT tensor must be identical
     # to INPUT values
 
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 
@@ -748,7 +756,11 @@ def create_plan_dynamic_modelfile(
     models_dir, model_version, io_cnt, max_batch, dtype, shape, profile_max_size
 ):
     # Create the model
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 

qa/common/gen_qa_implicit_models.py

@@ -899,7 +899,11 @@ def create_onnx_modelconfig(
 
 def create_plan_modelfile(models_dir, model_version, max_batch, dtype, shape):
     trt_dtype = np_to_trt_dtype(dtype)
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 
@@ -1005,7 +1009,11 @@ def create_plan_rf_modelfile(models_dir, model_version, max_batch, dtype, shape)
     trt_dtype = np_to_trt_dtype(dtype)
     trt_memory_format = trt.TensorFormat.LINEAR
 
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 

qa/common/gen_qa_model_repository

@@ -430,7 +430,7 @@ python3 $VOLUME_SRCDIR/gen_qa_ragged_models.py --tensorrt --models_dir=$VOLUME_R
 chmod -R 777 $VOLUME_RAGGEDDESTDIR
 python3 $VOLUME_SRCDIR/gen_qa_trt_format_models.py --models_dir=$VOLUME_FORMATDESTDIR
 chmod -R 777 $VOLUME_FORMATDESTDIR
-python3 $VOLUME_SRCDIR/gen_qa_trt_data_dependent_shape.py --models_dir=$VOLUME_DATADEPENDENTDIR
+nvidia-smi --query-gpu=compute_cap | grep -qz 11.0 && echo -e '\033[33m[WARNING]\033[0m Skipping model generation for data dependent shape' || python3 $VOLUME_SRCDIR/gen_qa_trt_data_dependent_shape.py --models_dir=$VOLUME_DATADEPENDENTDIR
 chmod -R 777 $VOLUME_DATADEPENDENTDIR
 # Make shared library for custom Hardmax plugin.
 if [ -d "/usr/src/tensorrt" ]; then
@@ -463,6 +463,7 @@ if [ "$MODEL_TYPE" != "igpu" ] ; then
     --label PROJECT_NAME=$PROJECT_NAME \
     $DOCKER_GPU_ARGS \
     -v $DOCKER_VOLUME:/mnt \
+    -e TRT_VERBOSE \
     $TENSORRT_IMAGE \
     bash -xe $VOLUME_SRCDIR/$TRTSCRIPT
 

qa/common/gen_qa_models.py

@@ -66,7 +66,11 @@ def create_plan_dynamic_rf_modelfile(
     trt_memory_format = trt.TensorFormat.LINEAR
 
     # Create the model
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
     if max_batch == 0:
@@ -206,7 +210,11 @@ def create_plan_dynamic_modelfile(
     trt_output1_dtype = np_to_trt_dtype(output1_dtype)
 
     # Create the model
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
     if max_batch == 0:
@@ -372,7 +380,11 @@ def create_plan_fixed_rf_modelfile(
     trt_memory_format = trt.TensorFormat.LINEAR
 
     # Create the model
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
     if max_batch == 0:
@@ -483,7 +495,11 @@ def create_plan_fixed_modelfile(
     trt_output1_dtype = np_to_trt_dtype(output1_dtype)
 
     # Create the model
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
     if max_batch == 0:

qa/common/gen_qa_ragged_models.py

@@ -57,7 +57,11 @@ def create_plan_modelfile(models_dir, model_version, dtype):
     # - BATCH_MAX_ELEMENT_COUNT_AS_SHAPE
     # - BATCH_ITEM_SHAPE_FLATTEN
 
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
     trt_dtype = np_to_trt_dtype(dtype)
@@ -412,7 +416,11 @@ def create_plan_itemshape_modelfile(models_dir, model_version, dtype):
     # generated to have matching batch dimension, the output can be produced
     # via identity op and expect Triton will scatter the output properly.
 
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
     trt_dtype = np_to_trt_dtype(dtype)

qa/common/gen_qa_reshape_models.py

@@ -58,7 +58,11 @@ def create_plan_modelfile(
     io_cnt = len(input_shapes)
 
     # Create the model that copies inputs to corresponding outputs.
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 

qa/common/gen_qa_sequence_models.py

@@ -59,7 +59,11 @@ def create_plan_shape_tensor_modelfile(
     trt_shape_dtype = np_to_trt_dtype(shape_tensor_input_dtype)
     trt_memory_format = trt.TensorFormat.LINEAR
 
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 
@@ -182,7 +186,11 @@ def create_plan_modelfile(models_dir, model_version, max_batch, dtype, shape):
     # Create the model. For now don't implement a proper accumulator
     # just return 0 if not-ready and 'INPUT'+'START' otherwise...  the
     # tests know to expect this.
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 
@@ -271,7 +279,11 @@ def create_plan_rf_modelfile(models_dir, model_version, max_batch, dtype, shape)
     # Create the model. For now don't implement a proper accumulator
     # just return 0 if not-ready and 'INPUT'+'START' otherwise...  the
     # tests know to expect this.
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 

qa/common/gen_qa_trt_data_dependent_shape.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 
-# Copyright 2022-2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# Copyright 2022-2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
 # modification, are permitted provided that the following conditions
@@ -45,7 +45,11 @@ def create_data_dependent_modelfile(
     trt_input_dtype = np_to_trt_dtype(input_dtype)
 
     # Create the model
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
 

qa/common/gen_qa_trt_format_models.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 
-# Copyright 2020-2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# Copyright 2020-2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
 # modification, are permitted provided that the following conditions
@@ -81,7 +81,11 @@ def create_plan_modelfile(
     trt_output_memory_format = output_memory_format
 
     # Create the model
-    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+    TRT_LOGGER = (
+        trt.Logger(trt.Logger.INFO)
+        if os.environ.get("TRT_VERBOSE") != "1"
+        else trt.Logger(trt.Logger.VERBOSE)
+    )
     builder = trt.Builder(TRT_LOGGER)
     network = builder.create_network()
     if max_batch == 0: