Could you please add support for Apple's M1 chip for running Mobilenet SSD V2 training

[getabhishekified]

Prerequisites

Please answer the following question for yourself before submitting an issue.

• [x] I checked to make sure that this feature has not been requested already.

1. The entire URL of the file you are using

https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf2.md

2. Describe the feature you request

I tried installing the Mobilenet SSD V2 on Apple's Mac M1 chip laptop, but getting error on running this command: python -m pip install --use-feature=2020-resolver .

error: ERROR: Could not find a version that satisfies the requirement tensorflow_io (from object-detection) (from versions: none) ERROR: No matching distribution found for tensorflow_io

4. Are you willing to contribute it? (Yes or No)

No. I am a beginner.

[vineetsharma14]

I am having a similar issue. I have installed Tensorflow Mac OS using miniforge3 as per the instruction mentioned on the link : https://developer.apple.com/metal/tensorflow-plugin/

I have verified that the Tensorflow is correctly installed and is using the GPU on the M1 Mac.

I then tried to installed Tensorflow Object Detection API (tried both version 1 and 2); but the installation always gives error on the command "python -m pip install --use-feature=2020-resolver ."

Can someone please clarify - if we can use Tensorflow Object Detection API on Mac with Apple M1 Chip. If yes, it would be great if someone can provide us with the installation steps to follow.

Thanks.

[dusskapark]

Same problem here.

I got error messages below:

// python -m pip install .

ERROR: Could not find a version that satisfies the requirement tensorflow_io (from object-detection) (from versions: none)
ERROR: No matching distribution found for tensorflow_io

[AritroSaha10]

Here's a similar issue on the Tensorflow IO repo: https://github.com/tensorflow/io/issues/1625

[divyansh-kumar]

Same problem here.

I got error messages below:

// python -m pip install .

ERROR: Could not find a version that satisfies the requirement tensorflow_io (from object-detection) (from versions: none)
ERROR: No matching distribution found for tensorflow_io

This might help https://stackoverflow.com/questions/70277737/cant-install-tensorflow-io-on-m1

[phuochung]

Same issue

[damian0815]

Same issue, but there's another problem which has to do with protobuf versions. tensorflow-macos==2.9.2 (latest) and tensorboard 2.9.1 both want protobuf<3.20,>=3.9.2, but protobuf.internal.builder (needed to run object_detection/builders/model_builder_tf2_test.py and I guess other things too) does not exist in protobuf 3.19.

Forcing the upgrade past protobuf 3.20 causes this error:

(tensorflow-object-detection-api) damian@d-mba2 research % python object_detection/builders/model_builder_tf2_test.py
Traceback (most recent call last):
  File "object_detection/builders/model_builder_tf2_test.py", line 21, in <module>
    import tensorflow.compat.v1 as tf
  File "/Users/damian/miniforge3/envs/tensorflow-object-detection-api/lib/python3.8/site-packages/tensorflow/__init__.py", line 37, in <module>
    from tensorflow.python.tools import module_util as _module_util
  File "/Users/damian/miniforge3/envs/tensorflow-object-detection-api/lib/python3.8/site-packages/tensorflow/python/__init__.py", line 37, in <module>
    from tensorflow.python.eager import context
  File "/Users/damian/miniforge3/envs/tensorflow-object-detection-api/lib/python3.8/site-packages/tensorflow/python/eager/context.py", line 29, in <module>
    from tensorflow.core.framework import function_pb2
  File "/Users/damian/miniforge3/envs/tensorflow-object-detection-api/lib/python3.8/site-packages/tensorflow/core/framework/function_pb2.py", line 16, in <module>
    from tensorflow.core.framework import attr_value_pb2 as tensorflow_dot_core_dot_framework_dot_attr__value__pb2
  File "/Users/damian/miniforge3/envs/tensorflow-object-detection-api/lib/python3.8/site-packages/tensorflow/core/framework/attr_value_pb2.py", line 16, in <module>
    from tensorflow.core.framework import tensor_pb2 as tensorflow_dot_core_dot_framework_dot_tensor__pb2
  File "/Users/damian/miniforge3/envs/tensorflow-object-detection-api/lib/python3.8/site-packages/tensorflow/core/framework/tensor_pb2.py", line 16, in <module>
    from tensorflow.core.framework import resource_handle_pb2 as tensorflow_dot_core_dot_framework_dot_resource__handle__pb2
  File "/Users/damian/miniforge3/envs/tensorflow-object-detection-api/lib/python3.8/site-packages/tensorflow/core/framework/resource_handle_pb2.py", line 16, in <module>
    from tensorflow.core.framework import tensor_shape_pb2 as tensorflow_dot_core_dot_framework_dot_tensor__shape__pb2
  File "/Users/damian/miniforge3/envs/tensorflow-object-detection-api/lib/python3.8/site-packages/tensorflow/core/framework/tensor_shape_pb2.py", line 36, in <module>
    _descriptor.FieldDescriptor(
  File "/Users/damian/miniforge3/envs/tensorflow-object-detection-api/lib/python3.8/site-packages/google/protobuf/descriptor.py", line 560, in __new__
    _message.Message._CheckCalledFromGeneratedFile()
TypeError: Descriptors cannot not be created directly.
If this call came from a _pb2.py file, your generated code is out of date and must be regenerated with protoc >= 3.19.0.
If you cannot immediately regenerate your protos, some other possible workarounds are:
 1. Downgrade the protobuf package to 3.20.x or lower.
 2. Set PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python (but this will use pure-Python parsing and will be much slower).

Rebuilding protos by doing protoc object_detection/protos/*.proto --python_out=. as in this guide does not make the issue go away.

The workaround here is awful - manually copy builder.py from a more recent installation of protobuf into <env>/lib/python3.8/site-packages/google/protobuf/internal/ - but it works.

Would be good to not have to do this!

[aino-gautam]

Is there support for tensorflow objection detection on Mac M1 / M2 ?

I am getting this error -

ERROR: Could not find a version that satisfies the requirement tensorflow_io (from object-detection) (from versions: none) ERROR: No matching distribution found for tensorflow_io