tensorflow / benchmarks

A benchmark framework for Tensorflow

tensorflow/benchmarks Issues

4090 multi gpu not support
Closed 5 months ago
ImportError: cannot import name 'variable_v1' from 'tensorflow.python.ops'
Closed 6 months ago2
VariableV1 does not exist in /tensorflow/python/ops/variables.py
Closed 7 months ago2
Docker 22.04 ubuntu and 12.1 cuda?
Updated 9 months ago1
perfzero resnet benchmark is outdated
Updated 10 months ago3
What to use as replacement for tf_cnn_benchmark in the official tensorflow models
Closed a year ago2
PerfZero Dataset for RetinaNet is not Avaiable
Closed a year ago1
Getting Error "ModuleNotFoundError: No module named 'official.resnet'" in PerfZero benchmark
Closed 4 years ago6
Impossible to test perfzero without docker
Updated 2 years ago2
Alternative/current state of tf_cnn_benchmark
Updated 2 years ago3
How to evaluate worker performance independently on a distributed training
Updated 2 years ago
Alexnet benchmark 1.12 performance poor on V100S 2GPU
Updated 2 years ago1
Perfzero support for Openshift on RHEL
Updated 2 years ago
OP_REQUIRES failed at save_restore_v2_ops.cc:205 : Not found: Key grouping/TCN/res_0_1/layer_normalization_1/beta not found in checkpoint
Closed 3 years ago
resnet50 --use_fp16 error: cuDNN launch failure : input shape ([128,112,112,64])
Updated 3 years ago
The accuracy of the program running by horovod is low
Updated 3 years ago
TypeError: visualize_boxes_and_labels_on_image_array() missing 1 required positional argument: 'category_index'
Updated 3 years ago
Fails to run tf_cnn_benchmarks.py in branch cnn_tf_v1.13_compatible with parameter server mode
Closed 4 years ago3
not using GCE instance
Updated 3 years ago2
libcudart.so.10.1 (and others) are not in the built docker image
Updated 3 years ago6
ImportError: cannot import name multi_device_iterator_ops
Closed 3 years ago1
Failure when running models with tf_cnn_benchmarks (threadpool in preprocessing.py)
Closed 3 years ago1
NotFoundError: No CPU devices are available in this process
Closed 3 years ago1
How do I get TF2.0 benchmark.
Closed 4 years ago1
data_flow_ops.RecordInput outperforms tf.data
Updated 4 years ago
ModuleNotFoundError: No module named 'official.benchmark'
Closed 4 years ago3
Imagenet (synthetic) data generation methodology?
Closed 4 years ago1
distributed_replicated is worse than parameter_server？
Closed 4 years ago
some question about server_protocol parameter in tf_cnn_benchmarks.
Closed 4 years ago
Link to estimator_benchmark.py in PerfZero readme is broken
Updated 4 years ago
PerfZero without Docker fails with AccessDeniedException
Updated 4 years ago3
resnet50 training rasie "tensorflow.python.framework.errors_impl.InvalidArgumentError: Retval[0] does not have value" on aarch64
Closed 4 years ago3
Default setup doesn't build image with all required libraries
Closed 4 years ago1
ImportError of nccl
Closed 4 years ago1
Page not found
Closed 4 years ago1
Why need L2_loss?
Closed 4 years ago1
Gradients are not averaged in AllReduceSpec=nccl (variable_update=replicated) mode
Closed 4 years ago1
How to replace AllReduce with Reduce in parameter server mode?
Updated 4 years ago4
Invalid argument: Default MaxPoolingOp only supports NHWC on device type CPU
Closed 4 years ago
The meaning of the flag "use_datasets" is confusing.
Updated 4 years ago3
Error in calculating image ratio
Closed 4 years ago1
Unable to run tf_cnn_benchmarks.py
Updated 4 years ago
How to set num_eval_epochs?
Closed 4 years ago3
Are the tf_cnn_benchmarks deprecated and replaced by MLPerf?
Closed 4 years ago8
Mismatched BROADCAST CPU/GPU
Updated 4 years ago1
Stable branch for TF1.15?
Closed 4 years ago1
Doesn't work with cuDNN v7.6.5 (November 18th, 2019), for CUDA 10.2
Closed 4 years ago1
which branch is compatiable with TF 1.14?
Closed 4 years ago2
no attribute 'function_buffering_resource'
Closed 4 years ago1
PerfZero README not accurate in the "PerfZero without Docker" section
Closed 4 years ago2