nsp | node security platform command-line tool | Runtime Evironment library

Use:

You forgot to declare the protocol in your URL, try http://192.168.1.101:1337 or ws://192.168.1.101:1337 (If your server does not use SSL)

And be sure the server side also implements socket.io server not just TCP protocol.

Setting {transports: ["websocket"]} on the client seems to have done the trick.

There is no such thing as a "JMS channel," per se. The channel referenced here is the underlying Netty channel used in the JMS client implementation provided by ActiveMQ Artemis.

The stack-trace simply indicates that the channel has failed. It has been disconnected. Therefore, the client is no longer connected to the broker. There is no specific indication of why the channel has been disconnected. The disconnect could have been caused by the broker crashing, a network failure of some kind, etc.

I solved this problem by doing the following steps:

1. Giving up socket.io-client and moving to ngx-socket-io. Guess we should not use socket.io-client directly if we are using an Angular frontend.

After doing so, I found that the console on the browser starts popping up cors errors. And

1. I added the white list for the server-side socket io initialization.

I ended up solving this by listening for the 'audio' event in the socket connection handler, then emitting the events on a separate EventEmitter instance created for the purpose. I was then able to listen on that instance in my get handler.

Regarding the AbstractMethod error, that typically happens when you invoke an interface method which is at a higher JDBC spec level than the implementation actually implements. The java.sql.Connection.createArrayOf method was added in JDBC 4.0. Are you using an older spec level JDBC driver JAR that implements an older version such as JDBC 3.0 which would lack the createArrayOf method? That seems like a good possibility based on the error you reported. If so, just move up to a newer spec level version of the JDBC driver.

I would suggest doing the following:

1. First pre-train BERT on the MLM objective. HuggingFace provides a script especially for training BERT on the MLM objective on your own data. You can find it here. As you can see in the run_mlm.py script, they use AutoModelForMaskedLM, and you can specify any architecture you want.

2. Second, if want to train on the next sentence prediction task, you can define a BertForPretraining model (which has both the MLM and NSP heads on top), then load in the weights from the model you trained in step 1, and then further pre-train it on a next sentence prediction task.

UPDATE: apparently the next sentence prediction task did help improve performance of BERT on some GLUE tasks. See this talk by the author of BERT.

As mentioned in the comment, you can try the following implementation of MLP-BERT TensorFlow. It's a simplified version and easy to follow comparatively.