WEBVTT

1
00:00:00.000 --> 00:00:01.920
<v ->[Maximilian Schwarzmuller] So with help of Ollama,</v>

2
00:00:01.920 --> 00:00:05.790
we got a model up and running and we can chat with it.

3
00:00:05.790 --> 00:00:07.540
As I explained in the previous lecture,

4
00:00:07.540 --> 00:00:11.190
we don't get a graphical user interface with Ollama.

5
00:00:11.190 --> 00:00:14.670
If you want one LM Studio is the better idea.

6
00:00:14.670 --> 00:00:16.500
Or alternatively,

7
00:00:16.500 --> 00:00:20.490
you could explore options like Open WebUI.

8
00:00:20.490 --> 00:00:24.600
Open WebUI is a tool you can install

9
00:00:24.600 --> 00:00:26.075
and run on your system,

10
00:00:26.075 --> 00:00:28.286
and you'll find installation instructions

11
00:00:28.286 --> 00:00:29.850
on their GitHub page.

12
00:00:29.850 --> 00:00:32.398
It helps if you got some programming experience

13
00:00:32.398 --> 00:00:35.520
or some experience with using Docker though,

14
00:00:35.520 --> 00:00:37.106
because that's one of the main ways

15
00:00:37.106 --> 00:00:39.625
of running this tool on your system.

16
00:00:39.625 --> 00:00:43.217
It's not really a tool that's meant to be used by users

17
00:00:43.217 --> 00:00:45.906
that don't have any development experience.

18
00:00:45.906 --> 00:00:50.520
I do recommend LM Studio if you want a hassle-free way

19
00:00:50.520 --> 00:00:53.850
of running open models locally without having to deal

20
00:00:53.850 --> 00:00:56.398
with Docker or advanced setup steps.

21
00:00:56.398 --> 00:00:58.920
But if you do have that programming knowledge,

22
00:00:58.920 --> 00:01:02.888
you could explore installing and running Open WebUI

23
00:01:02.888 --> 00:01:07.888
because what you'll get is a browser-based AI chatbot

24
00:01:08.040 --> 00:01:09.990
that's running on your system though,

25
00:01:09.990 --> 00:01:13.256
not on some remote server that can connect

26
00:01:13.256 --> 00:01:17.370
to proprietary models like the open AI models.

27
00:01:17.370 --> 00:01:19.997
You will need to pay for the usage of that though.

28
00:01:19.997 --> 00:01:23.370
But that can also connect to locally running models,

29
00:01:23.370 --> 00:01:27.900
for example, to models that are running through Ollama,

30
00:01:27.900 --> 00:01:29.059
like here, where I connected

31
00:01:29.059 --> 00:01:32.517
to that locally running Gemma 3 model, which is running

32
00:01:32.517 --> 00:01:36.030
through Ollama, which I started a couple of minutes ago,

33
00:01:36.030 --> 00:01:37.604
and with which I could now chat here

34
00:01:37.604 --> 00:01:42.150
in this beautiful web-based chat user interface.

35
00:01:42.150 --> 00:01:45.229
So that would give me a graphical user interface

36
00:01:45.229 --> 00:01:48.690
running on my system to chat with that model

37
00:01:48.690 --> 00:01:51.948
that's also running on my system with help of Ollama.

38
00:01:51.948 --> 00:01:54.600
Using that is totally optional though,

39
00:01:54.600 --> 00:01:57.060
and not the focus of this section.

40
00:01:57.060 --> 00:02:00.810
Since if you want a graphical user interface to work

41
00:02:00.810 --> 00:02:03.480
and interact with locally running open models,

42
00:02:03.480 --> 00:02:05.941
I recommend using LM Studio.

43
00:02:05.941 --> 00:02:10.593
I just wanted to mention Open WebUI as a great alternative.