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.
64 lines
2.1 KiB
64 lines
2.1 KiB
WEBVTT |
|
|
|
00:00.920 --> 00:03.530 |
|
Well, another congratulations moment. |
|
|
|
00:03.530 --> 00:13.010 |
|
You have 40% on the way to being an LM engineer at a high level of proficiency at this point, in addition |
|
|
|
00:13.010 --> 00:19.340 |
|
to the stuff that you you know very well about frontier models and open source models, you can also |
|
|
|
00:19.340 --> 00:28.160 |
|
now use the Hugging Face open LM leaderboard to compare different models against a number of hard metrics, |
|
|
|
00:28.160 --> 00:34.010 |
|
and you've got a great understanding of the difference between these metrics, what they score and why, |
|
|
|
00:34.010 --> 00:36.620 |
|
and the limitations of the metrics as well. |
|
|
|
00:36.650 --> 00:42.860 |
|
Hopefully you spent some quality time yourself now with the LM leaderboard, getting familiar with the |
|
|
|
00:42.860 --> 00:44.060 |
|
different models. |
|
|
|
00:44.270 --> 00:47.480 |
|
Next week there is some more on leaderboards to do. |
|
|
|
00:47.480 --> 00:51.050 |
|
I hope you're not bored of leaderboards yet, because we're going to look at a bunch of others that |
|
|
|
00:51.050 --> 00:57.560 |
|
are used to compare more specialized things about models, and we're also going to look at ones that |
|
|
|
00:57.560 --> 01:01.910 |
|
bring together open source and closed source, which will be very interesting. |
|
|
|
01:02.330 --> 01:08.390 |
|
We are going to talk about real world use cases of LM solving commercial problems. |
|
|
|
01:08.390 --> 01:16.220 |
|
I imagine everyone has experienced this in your workplace and in just using products all the time. |
|
|
|
01:16.220 --> 01:22.430 |
|
But I want to go through a few examples and have you think about other places where Llms have been applied |
|
|
|
01:22.460 --> 01:29.300 |
|
to solve hard commercial problems beyond the obvious, and that should equip you to be able to confidently |
|
|
|
01:29.300 --> 01:36.530 |
|
choose the right LM for your task at hand, so that you know when you're facing a commercial project, |
|
|
|
01:36.560 --> 01:43.670 |
|
how do you go about navigating the huge number of possible models and select the right couple to build |
|
|
|
01:43.670 --> 01:44.690 |
|
a prototype for? |
|
|
|
01:44.690 --> 01:47.060 |
|
So with that, I will see you next time. |
|
|
|
01:47.060 --> 01:48.110 |
|
We've got lots to do.
|
|
|