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.
 
 

61 lines
2.0 KiB

WEBVTT
00:00.740 --> 00:05.720
Well, well, well, it's been a long day, but congratulations, you've made it.
00:05.750 --> 00:12.260
We've gone through and curated a pristine data set, working very hard to make sure that it's got a
00:12.260 --> 00:16.310
good sample representation of the data we want to train by.
00:16.340 --> 00:23.450
And at the end, we of course turned it into a hugging face data set, a data set dict with the training
00:23.450 --> 00:28.310
and test parts to it, and we uploaded it to the Hugging Face Hub.
00:28.310 --> 00:31.880
And if you've gone through all of these instructions, which I know you will have done, you've been
00:31.880 --> 00:36.410
following along in JupyterLab, getting comfortable with the different things that I've been doing then
00:36.410 --> 00:44.210
now you'll be rewarded with your own data set, sitting there in the hub that you will be able to use
00:44.210 --> 00:46.820
in the subsequent sessions.
00:46.820 --> 00:50.060
So congratulations on on on getting that far.
00:50.270 --> 00:56.660
Uh, so we've added to the skills that you've acquired, the understanding of the five step strategy
00:56.660 --> 01:03.560
to solving commercial business problems with Llms, uh, weighing up the three different optimization
01:03.560 --> 01:11.900
techniques and some real detail in data set curation, including some thorny, uh, bits of code there
01:11.900 --> 01:18.020
that I do hope you'll look through and understand them and then use them in your projects, like sampling
01:18.050 --> 01:20.180
from existing data sets.
01:20.630 --> 01:25.550
So next time, next time we're going to be talking about baseline models.
01:25.550 --> 01:29.210
We're going to be creating a traditional machine learning solution.
01:29.210 --> 01:36.320
And we're going to be applying some traditional and advanced techniques to see what gives us good results.
01:36.320 --> 01:38.120
And I'm excited for it.
01:38.120 --> 01:39.440
And I will see you there.