tekkonetes
commited on
Commit
•
71780e7
1
Parent(s):
2cbf387
Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Train-your-own GPT
|
2 |
+
A very simple script on how to train a GPT model with just a text file. Created by @Tekkonetes (Hugging Face) or @pxlmastrXD (Replit).
|
3 |
+
|
4 |
+
## Setup
|
5 |
+
You will need some packages installed to run this. Make sure you have a recent (supported) python version, and then install some packages.
|
6 |
+
I would recommend using a Virtual environment to do this, so you can remove it easily after training and deployment:
|
7 |
+
```shell
|
8 |
+
python -m venv venv
|
9 |
+
source venv/bin/activate
|
10 |
+
```
|
11 |
+
Then, install some required packages:
|
12 |
+
```shell
|
13 |
+
pip install transformers numpy torch
|
14 |
+
```
|
15 |
+
That should be all you need.
|
16 |
+
## Usage
|
17 |
+
First, put all your data into the train.txt file. By tweaking the code, you can use multiple.
|
18 |
+
|
19 |
+
Second, select your GPT model. By default, the script uses `gpt2-medium`, but you can replace it with most GPT-based models.
|
20 |
+
You might have to do some tweaking, however.
|
21 |
+
|
22 |
+
Then, run the script.
|
23 |
+
```shell
|
24 |
+
python main.py
|
25 |
+
```
|
26 |
+
|
27 |
+
That is pretty much it. Feel free to use this for whatever you want!!!
|
28 |
+
- pxlmastrXD
|