WEBVTT

00:00.040 --> 00:00.520
All right.

00:00.560 --> 00:04.040
So now let's go and let's debug it.

00:04.040 --> 00:08.440
And let's examine some objects here so we can get a better understanding.

00:08.680 --> 00:11.840
So I'm going to put a breakpoint and run it in debug mode.

00:11.960 --> 00:16.080
And let me fast forward everything until we hit this breakpoint here.

00:19.720 --> 00:22.280
Let's go and check out the response.

00:22.520 --> 00:26.000
And the response is of type a message.

00:26.120 --> 00:33.880
And a message is a simple wrapper class on what the LM is going to return us, and it's going to be

00:33.880 --> 00:37.000
located in the field of content.

00:37.160 --> 00:42.360
So in content here you can see that we have here the answer that the LM generated.

00:42.600 --> 00:50.480
Now I message is containing a lot more information like tool calling like how much tokens did we consume.

00:50.520 --> 00:51.880
Like how much did it cost.

00:51.920 --> 00:53.240
And all of those things.

00:53.520 --> 00:56.680
However, we're not going to review all of these features right now.

00:56.680 --> 01:02.680
If you want, you can check out the messages video I made where I elaborately go and discuss all the

01:02.680 --> 01:04.320
message types in link chain.

01:04.960 --> 01:10.880
All right, so let's go and just poke around the I message object.

01:11.040 --> 01:15.200
So what we can also find here is that here we can see that the type here is I.

01:15.880 --> 01:22.480
And we can also go and check out the response metadata which we're going to have some information about.

01:22.480 --> 01:23.920
Which model did we use.

01:23.960 --> 01:29.920
What was the finish reason for the LM and will be elaborating a lot about this when we're talking about

01:29.920 --> 01:36.320
agents, and we can get the number of tokens we consumed, and a lot of metadata that is useful for

01:36.320 --> 01:39.280
debugging and for monitoring and analyzing.