From the uDemy course on LLM engineering.
https://www.udemy.com/course/llm-engineering-master-ai-and-large-language-models
You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
67 lines
2.1 KiB
67 lines
2.1 KiB
WEBVTT |
|
|
|
00:00.710 --> 00:05.480 |
|
Well, we took on a lot today and we seem to have been successful. |
|
|
|
00:05.510 --> 00:12.020 |
|
These red icons that you see on this agent workflows diagram have been built, and we have them in the |
|
|
|
00:12.020 --> 00:16.280 |
|
agents directory in the week eight folder. |
|
|
|
00:16.520 --> 00:22.220 |
|
And so do spend some time looking through them and getting some appreciation for what we've built. |
|
|
|
00:22.250 --> 00:31.970 |
|
And with that it brings us to a summary end of end of the day slide, which is inching ever closer to |
|
|
|
00:32.000 --> 00:34.130 |
|
the summit of the mountain. |
|
|
|
00:34.160 --> 00:37.820 |
|
You are at the moment at 92.5% of the way there. |
|
|
|
00:37.820 --> 00:42.260 |
|
And I hope you're starting to really get that sense that you are becoming advanced. |
|
|
|
00:42.290 --> 00:48.560 |
|
You're taking the conversion from being confident with this stuff to being an expert at this point. |
|
|
|
00:48.560 --> 00:54.380 |
|
You can also, along with everything else you can do, you can now deploy models to production, including |
|
|
|
00:54.410 --> 01:02.210 |
|
through modal, to call them remotely on the cloud and also building some pretty heavyweight rag pipelines |
|
|
|
01:02.210 --> 01:08.120 |
|
that look over a big Croma database to build context to send to OpenAI. |
|
|
|
01:08.360 --> 01:15.230 |
|
And combining that with machine learning models as well, to build a robust production solution. |
|
|
|
01:15.710 --> 01:19.970 |
|
Next time we're going to look at some more things, we're going to look at something called structured |
|
|
|
01:20.000 --> 01:24.980 |
|
outputs, which is a way that we can connect with OpenAI and enforce that. |
|
|
|
01:24.980 --> 01:28.640 |
|
The response comes in a particular specification. |
|
|
|
01:28.640 --> 01:34.220 |
|
It's something that's relatively new, and it's good for us to to add that to your skill sets to your |
|
|
|
01:34.220 --> 01:35.000 |
|
tool belt. |
|
|
|
01:35.420 --> 01:42.650 |
|
And it's just going to be about solving some more problems using models as you build and build expertise |
|
|
|
01:42.650 --> 01:44.090 |
|
and keep upskilling. |
|
|
|
01:44.090 --> 01:49.460 |
|
So I'm so excited and so pleased with our progress. |
|
|
|
01:49.490 --> 01:54.080 |
|
I'll see you next time when we look at structured outputs.
|
|
|