There are the steps that I executed (thanks Halysson Carvalho for the instructions)
$ mkdir TensorFlow$ git clone [email protected]:WildernessLabs/Meadow.Tensorflow.git
$ cd TensorFlow/Meadow.Tensorflow/Tensorflow
$ meadow file write -f Tensorflow.so
Writing 1 file to device...
Writing 'Tensorflow.so' as 'Tensorflow.so' to device
Writing Tensorflow.so: 100%
$ cd -
$ cd TensorFlow/Meadow.Tensorflow/Source/Tests/HelloWorld/
$ dotnet build
MSBuild version 17.8.5+b5265ef37 for .NET
Determining projects to restore...
Restored /WL/TensorFlow/Meadow.Tensorflow/Source/Tests/HelloWorld/HelloWorld.csproj (in 819 ms).
1 of 2 projects are up-to-date for restore.
TensorFlowMicro -> /WL/TensorFlow/Meadow.Tensorflow/Source/Meadow.Tensorflow/bin/Debug/netstandard2.1/TensorFlowMicro.dll
/WL/TensorFlow/Meadow.Tensorflow/Source/Tests/HelloWorld/MeadowApp.cs(73,36): warning CS1998: This async method lacks 'await' operators and will run synchronously. Consider using the 'await' operator to await non-blocking API calls, or 'await Task.Run(...)' to do CPU-bound work on a background thread. [/WL/TensorFlow/Meadow.Tensorflow/Source/Tests/HelloWorld/HelloWorld.csproj]
HelloWorld -> /WL/TensorFlow/Meadow.Tensorflow/Source/Tests/HelloWorld/bin/Debug/netstandard2.1/App.dll
Build succeeded.
/WL/TensorFlow/Meadow.Tensorflow/Source/Tests/HelloWorld/MeadowApp.cs(73,36): warning CS1998: This async method lacks 'await' operators and will run synchronously. Consider using the 'await' operator to await non-blocking API calls, or 'await Task.Run(...)' to do CPU-bound work on a background thread. [/WL/TensorFlow/Meadow.Tensorflow/Source/Tests/HelloWorld/HelloWorld.csproj]
1 Warning(s)
0 Error(s)
Time Elapsed 00:00:01.99
$ meadow app deploy
Deploying app from /WL/TensorFlow/Meadow.Tensorflow/Source/Tests/HelloWorld/bin/Debug/netstandard2.1...
Generating list of files to deploy...
Deleting file 'cell.config.yaml'...
Deleting file 'meadow.log'...
Deleting file 'dns.conf'...
Deleting file 'Meadow.OS.Runtime.bin'...
Deleting file 'update'...
Deleting file 'Tensorflow.so'...
Writing mscorlib.dll: 100%
Writing System.Core.dll: 100%
Writing System.dll: 100%
Writing Mono.Security.dll: 100%
Writing System.Configuration.dll: 100%
Writing System.Xml.dll: 100%
Writing System.Security.dll: 100%
Writing System.Numerics.dll: 100%
Writing System.Data.dll: 100%
Writing System.Transactions.dll: 100%
Writing System.EnterpriseServices.dll: 100%
Writing System.Data.DataSetExtensions.dll: 100%
Writing System.Drawing.dll: 100%
Writing System.IO.Compression.dll: 100%
Writing System.IO.Compression.FileSystem.dll: 100%
Writing System.ComponentModel.Composition.dll: 100%
Writing System.Net.Http.dll: 100%
Writing System.Runtime.Serialization.dll: 100%
Writing System.ServiceModel.Internals.dll: 100%
Writing System.Web.dll: 100%
Writing System.Runtime.Serialization.Formatters.Soap.dll: 100%
Writing System.ComponentModel.DataAnnotations.dll: 100%
Writing Mono.Data.Sqlite.dll: 100%
Writing System.Xml.Linq.dll: 100%
Writing Meadow.dll: 100%
Writing Meadow.Contracts.dll: 100%
Writing Meadow.Units.dll: 100%
Writing Meadow.Logging.dll: 100%
Writing System.Runtime.CompilerServices.Unsafe.dll: 100%
Writing System.Numerics.Vectors.dll: 100%
Writing MQTTnet.dll: 100%
Writing Meadow.F7.dll: 100%
Writing TensorFlowMicro.dll: 100%
Writing App.dll: 100%
Writing App.pdb: 100%
Writing app.config.yaml: 100%
Writing App.deps.json: 100%
Writing meadow.config.yaml: 100%
$ meadow runtime enable
Enabling runtime...
The operation has timed out.
>>> Reset the board pressing RST button <<<
$ meadow port select
Found device on port:
1: /dev/ttyACM0
Setting route=/dev/ttyACM0
$ meadow listen
Listening for Meadow Console output on '/dev/ttyACM0'. Press Ctrl+C to exit...
stdout> Creating 'F7CoreComputeV2' instance took 2113ms
stdout> Device Initialize starting...
stdout> Device is configured to use WiFi for the network interface
stdout> PlatformOS Initialize starting...
stdout> SD Card is configured to 'off'
stdout> File system Initialize starting...
stdout> Creating instance of MeadowApp...
stdout> Creating 'MeadowApp' instance took 123ms
stdout> Update Service is disabled.
stdout> Health Metrics disabled.
stdout> Initializing App
stdout> Initialize Tensor Flow ...
stdout> Running App
stdout> 0 - (0, 0)
stdout> 1 - (0.3141593, 0.3727683)
stdout> 2 - (0.6283185, 0.5591524)
stdout> 3 - (0.9424779, 0.8387287)
stdout> 4 - (1.256637, 0.9658087)
stdout> 5 - (1.570796, 1.042057)
stdout> 6 - (1.884956, 0.9573368)
stdout> 7 - (2.199115, 0.8217847)
stdout> 8 - (2.513274, 0.5337364)
stdout> 9 - (2.827433, 0.2372162)
stdout> 10 - (3.141593, 0.008472007)
stdout> 11 - (3.455752, -0.3049923)
stdout> 12 - (3.769912, -0.5337364)
stdout> 13 - (4.08407, -0.7794246)
stdout> 14 - (4.39823, -0.9658087)
stdout> 15 - (4.712389, -1.109833)
stdout> 16 - (5.026548, -0.9827528)
stdout> 17 - (5.340708, -0.7455366)
stdout> 18 - (5.654867, -0.5337364)
stdout> 19 - (5.969026, -0.3558243)
stdout> Tensor Flow completed