How to run Dall e on PC and Google Colab?

Many people have a query in their mind Can we run Dall-e run locally? The answer is Yes. We have a new model called OpenDalle which can generate image art like what we can inOpenAI's Dalle but using Stable Diffusion. 

Not, only this, there are multiple issues related to the policies and restrictions of Dalle which sometimes problematic and can't get the overall advantage of the black box model. But, using this model as open source you can get rid of such issues, widen your image generation leverage the power of Dalle into it.

We are showing this illustration on Automatic1111, but you can use this model in other WebUi or applications like ComfyUI, and Fooocus as well.

Install and run OpenDalle locally:

0. Make sure you have done installation of Python and Automatic1111 for Stable Diffusion if not yet.

1. Go to the Hugging Face platform by using the link provided below:

https://huggingface.co/dataautogpt3/OpenDalleV1.1 

2. Click on "File and Versions".

3. You will be moved to the next page. Just scroll a little downward and search for "OpenDalle-Safetensors" which will be around 6GB file. Click on the download button to download it.

4. Now, save the downloaded model into your model's folder. Restart the Automatic1111 to effect. 

5. Here, after loading the OpenDalle model, you will see a new model in Stable Diffusion Checkpoint on Automatic1111.

    Use the following settings:

    -CFG Scale: CFG scale of 8 to 7

    -Steps: 60 to 70 steps for more detail, 35 steps for faster results.

    -Sampler: DPM2

    -Scheduler: Normal or Karras

6. Now, let's try with different prompts:

The prompts used to generate this art style:
portrait of a black man, standing in a disco, GTA 5 concept art style

The results are really great. Observe the background with some disco lighting effects. Quite good. Now, let's try for GTA5 girl. 

The prompts used to generate this art style:

portrait of a woman, red long hair, standing in a disco, GTA 5 concept art style

Now the next one that we have tried is something in cartoonist style, and here is the result. 

 

The image-generated effect is really cartoonist, but the logo embossed into it is used by Photoshop which is not a tough job for you. But you can try to add logos into an image using the img2img function of Stable Diffusion as well. By the way, you can try Stable Diffusion on Photoshop as well using a simple plugin which will make your task easier.

              

Now, we tried something more related to cartoon style and these are the results.

     

The prompts used to generate this art:

portrait of a 20 years girl, wearing jeans, a jacket holding on her hand,  standing on the side of the road, Gta5 concept style art

Now, if you want to learn in-depth then you can check on how to do prompting with stable diffusion models as well.

Run on Google Colab Using Diffusers Library:

Coders who are using a diffuser library for art generation can take advantage of Dalle into Stable Diffusion using code.

1. Import AutoPipelineForText2Image from diffusers:

from diffusers import AutoPipelineForText2Image

2. Import torch

import torch       

3. Now, pass "dataautogpt3/OpenDalleV1.1" model as argument like this:

pipeline = AutoPipelineForText2Image.from_pretrained('dataautogpt3/OpenDalleV1.1', torch_dtype=torch.float16).to('cuda')        

4. Input your prompts as an argument into a created object:

image = pipeline('black fluffy gorgeous dangerous cat animal creature, large orange eyes, big fluffy ears, piercing gaze, full moon, dark ambiance, best quality, extremely detailed').images[0]

5. For displaying output on terminal, using python variable:

image

Conclusion:

Running Dalle3 locally or on cloud is simply impossible because it's a closed project, which is run and managed by OpenAI. It works on the concept of deep learning. But, OpenDalle is a new model released by the community by which one can take advantage of it with stable diffusion.