File size: 1,144 Bytes
479a20d
 
 
 
34c0cfa
 
 
 
479a20d
 
 
478c52d
479a20d
aa30071
 
 
254bfc7
aa30071
479a20d
 
 
 
 
 
 
 
 
aa30071
 
 
 
 
 
a9f5eb2
 
 
 
 
 
 
 
 
 
 
 
479a20d
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
---
license: gpl
language:
- en
datasets:
- Differentiable/processed_chats
tags:
- twitch
---
# Description

This is my first attempt to simulate the Filian's Twitch stream chat based on scraped data. The model is a fine-tuning of GPT-Neo for 125M parameters. The model was able to be trained within 30 minutes on a GTX 1650 with Pytorch's memory pinning feature to share the local RAM with the GPU.

# Usage

Make sure to include the prefix "STREAMER: " at the end of the prompt or anywhere the response of the streamer is supposed to go. "Chat messages" are prepend with the "CHAT: " keyword.
In case the Hosted Inference API (text box on the side) gets too little text, just keep pressing "Compute" until the sentence ends.

## Prompt examples
```
CHAT: You're cringe
CHAT: KEKW
STREAMER:
```
```
CHAT: Are you a Filipino boy?
STREAMER: 
```

```
CHAT: Hello
CHAT: KEKW Filian is cringe
STREAMER: 
```

Using the keyword "STREAM TITLE" will attempt to replicate the name of a stream:

```
STREAM TITLE: 
```

You can also induce the streamer to say something by adding a few words of your own at the beginning:

```
STREAM: Today we will
```