Update Y01E01 - Make an image/Y01E01 - Make an image.txt
Browse files
Y01E01 - Make an image/Y01E01 - Make an image.txt
CHANGED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
Huggingface for noobs - Year 01 Episode 01
|
| 2 |
+
Make an image! - Written for noobs, by a noob.
|
| 3 |
+
|
| 4 |
+
Today I write you this little text to teach you how to make an image.
|
| 5 |
+
|
| 6 |
+
You can do several things with Hugging Face, one of these things are related to something that HF calls 'models'.
|
| 7 |
+
I don't really know how to explain it but for me it is a sort of 'black box'. I can't explain you yet what is inside.
|
| 8 |
+
However you can put something in it, and get something out it. I have played a bit around and was able to put text/images/others inside the box,
|
| 9 |
+
and depending on the mysterious content inside, I got some really nice outputs: an image, a mp3-file or even a text!
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
This tutorial will be focussing on using an existing 'black box' or model where you put some text in and get an image out.
|
| 13 |
+
This kind of models are called 'Diffusers' by HF. You can find a list of all sorts of Diffusers here (https://huggingface.co/models?library=diffusers).
|
| 14 |
+
Many of the uploaded Diffusers doesn't work, so don't panick when you try one and nothing happens.
|
| 15 |
+
|
| 16 |
+
I have tested several Diffusers and one that works really nice is the following one:
|
| 17 |
+
black-forest-labs/FLUX.1-schnell (url: https://huggingface.co/black-forest-labs/FLUX.1-schnell)
|
| 18 |
+
*If this Diffuser isn't availible again, please let me know by placing a comment on this turial page.
|
| 19 |
+
|
| 20 |
+
So at this webpage, you would like to go to the 'Inference API'-part. This will let you allow to write a sentense and press 'Compute'.
|
| 21 |
+
This is the place where you write or describe your image. You are writing your description or prompt (see also prompt engineering).
|
| 22 |
+
As example you can type 'Photo of neon sunflowers', next press compute*.
|
| 23 |
+
*In some cases the inputfield and/or the computebutton are dissabled. If so, please find another Diffuser.
|
| 24 |
+
|
| 25 |
+
As soon you pressed compute, the 'black box' starts with their mysterious calculations and starts working on your image.
|
| 26 |
+
As soon as your image is ready, it will displayed.
|
| 27 |
+
|
| 28 |
+
So congratulations, you now have successfully made your first image. It is maybe still ugly or bad, but you can increase it quality by writing a more complex input/prompt.
|
| 29 |
+
Search online for 'Prompt Engineering' and you will learn what you can do more.
|
| 30 |
+
|
| 31 |
+
In some cases your prompts are just fine, and it is just the model that is acting weird.
|
| 32 |
+
|
| 33 |
+
'
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
|