Peter Lord

Sure ... this is what I've got ( hackish ... testing idea only ) - ``` diff --git a/_example/ssd/main.go b/_example/ssd/main.go index 47a9e7d..914ba7e 100644 --- a/_example/ssd/main.go +++ b/_example/ssd/main.go @@ -12,7 +12,6...

I re-built tensorflow locally ... this time used 2.5.3 rather than the latest. So far no SEGV.

I see newer versions of tensorflow make use of cmake. So this worked for me ( linux arm64 ) : ``` git clone https://github.com/tensorflow/tensorflow.git tensorflow_src cd tensorflow_src git checkout v2.10.0...

Many thanks for checking. Yes, we are suspecting a fault on the zappi.

> Charge target is implemented as well, but it's super-broken in the libbi firmware. 😞 Nice :-) I'll try it with the app tonight. BTW do you have 5.414 installed...

> For those who are foolish and reckless.... I present you the latest bleeding edge version of the integration. > New operation mode for your libbi: '**Export**'. Let me know...

> As this version adds the OAuth endpoints you need to remove your integration and re-add it as it requires both API key/password and App email/app password parameters. Ah, got...

I see this didn't make 0.0.25

> @plord12 The 0.1.0 version of pymyenergi (released yesterday) has all the necessary features in there, so once I rebase my local dev copy to 0.0.25, this will be in...

I hate to pester, given so many give their time for projects like this, but any news ? Once this gets in I hope to look at the libbi dual...