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.
82 lines
2.5 KiB
82 lines
2.5 KiB
WEBVTT |
|
|
|
00:00.980 --> 00:06.620 |
|
I'm so very happy that you've reached this epic moment in the course and that you're hanging in there. |
|
|
|
00:06.620 --> 00:12.110 |
|
This is where things have gotten really crucial, and we're learning some of the most important things |
|
|
|
00:12.110 --> 00:13.640 |
|
about building an LMS. |
|
|
|
00:13.760 --> 00:20.930 |
|
And I'm I'm just very grateful that you've been staying along and hopefully enjoying it as much as I |
|
|
|
00:20.960 --> 00:21.530 |
|
am. |
|
|
|
00:21.770 --> 00:24.590 |
|
And yeah, the best is yet to come. |
|
|
|
00:24.590 --> 00:28.490 |
|
So, uh, for today, it's going to be few slides. |
|
|
|
00:28.490 --> 00:30.320 |
|
It's going to be all action. |
|
|
|
00:30.410 --> 00:35.360 |
|
Uh, you're going to be looking at the actual run happening and weights and biases, and I'll be showing |
|
|
|
00:35.360 --> 00:40.070 |
|
you the sorts of things to look for and how we can use weights and biases to get some intelligence about |
|
|
|
00:40.070 --> 00:40.970 |
|
what's happening. |
|
|
|
00:41.270 --> 00:46.610 |
|
Uh, I'm going to show you the Huggingface hub with models in it, and you get a sense of what's going |
|
|
|
00:46.610 --> 00:47.210 |
|
on. |
|
|
|
00:47.420 --> 00:54.050 |
|
Um, there's also one thing that's been on my mind that I mentioned last time that I, I don't want |
|
|
|
00:54.050 --> 00:58.220 |
|
to be flippant about the importance of keeping costs low. |
|
|
|
00:58.310 --> 01:04.190 |
|
The idea of this class is not to build up a big bill for you from, uh, from Google, who will happily |
|
|
|
01:04.190 --> 01:07.190 |
|
take your, your money for for running their boxes. |
|
|
|
01:07.310 --> 01:10.010 |
|
Um, I have a lot of fun running this training. |
|
|
|
01:10.010 --> 01:12.740 |
|
It's, um, running the training process. |
|
|
|
01:12.800 --> 01:19.820 |
|
Um, it's perfectly achievable to do that and get good results with the spending a matter of cents. |
|
|
|
01:19.820 --> 01:23.810 |
|
And I did want to take just a quick moment to talk about that, because it's very important. |
|
|
|
01:23.840 --> 01:28.340 |
|
And I don't want to, uh, lead you down the wrong path of thinking you have to spend lots of money |
|
|
|
01:28.340 --> 01:32.180 |
|
to get good results and to learn particularly, which is the main objective here. |
|
|
|
01:32.480 --> 01:35.540 |
|
Um, and so, in fact, we're going to start with that. |
|
|
|
01:35.600 --> 01:42.530 |
|
Um, and I want to, uh, take you straight over to JupyterLab, not not to Google Colab, but JupyterLab, |
|
|
|
01:42.530 --> 01:47.780 |
|
where I'm going to show you something to illustrate my point and to help set you up should you wish |
|
|
|
01:47.780 --> 01:50.090 |
|
to be training at a lower cost.
|
|
|