Because it is heavy and requires more stuff than SD.
And we have new way to run Flux ez with 1 click: https://civitai.com/models/628682/flux-1-checkpoint-easy-to-use
Check out more detailed instructions here: https://maitruclam.com/flux-ai-la-gi/
- This is the SFW model ðŸ˜ðŸ˜ðŸ˜
- Really sticks to the prompt
- Recognizes and displays text very well
- Variety of content can be created without Lora or additional training
- Cool cool cool!
- Write a simple prompt like you don't know MidJourney or Stable Diffusion (bye bye a bunch of techniques, brackets, strange codes and negative prompts)
- Just a simple description like: a young woman lying on a bed wearing a shirt with the words "touch me".
- The model is super heavy! 24GB
- It takes a long time to create but it's cool: 30s/ 1024x1024.
- Should use VPS or if you are not afraid of burning your laptop (or if you are rich, it depends on you :) )
Note:
- If you have less than 30GB of storage and 12GB of vRam, you should probably consider using it.
- If you download and it doesn't work, change .safetensors to .sft.
- Currently it only works with Comfy UI, simple workflow here:
- A1111 will probably update... maybe.
Before:
- Is your computer configuration strong enough?
- Minimum 12GB vRam, Nvidia 3060 card or higher (this is from me, not their official information, because my laptop running Nvidia 3060 6GB vram is fine even though it goes up to 105 degrees C when running)
- Minimum 30GB of space to store models, VAE and CLIP?
- Do you choose dev (quality) or schnell (fast) to run?
Also, you should check my tutorial to understand how to install and some things around: https://maitruclam.com/flux-ai-la-gi/
While:
- Make sure you have downloaded all the required files into UNET, VAE AND CLIP
- Make sure you have upgraded ComfyUI to the latest version
After:
- On average, the reports show that the minimum time when running with schnell is 10s for 1 photo 1024x1024px.
- For weak devices, the time for each 1024x1024 photo is 2p - 45p (wtf this guy is really patient when waiting for 45p :))
- If you have unique workflows to process photos such as: creating a series of different result prompts, can process skin, color, change style,... Contact me and I am ready to share my VPS to discuss more in: https://discord.gg/PYefENGNRY
My setting VPS:
- I am using VPS on AWS.
- VPS g5.4xlarge (GPU A10 24GB vRam)
- 800GB of storage
- If you want to share and use VPS, you can contact me via the comment, Facebook fb.com/maitruclam4real or email: me@maitruclam.com.
- I DO NOT SHARE FREE VPS, but will exchange with appropriate value (other software, services, knowledge of using VPS, creating videos on VPS Ubuntu, how to optimize AWS...) contact and exchange with each other about what you have and what I have so that both sides can happily share resources.
And here is the description of the publisher:
FLUX.1 [dev]
is a 12 billion parameter rectified flow transformer capable of generating images from text descriptions. For more information, please read our blog post.
Cutting-edge output quality, second only to our state-of-the-art model FLUX.1 [pro]
.
Competitive prompt following, matching the performance of closed source alternatives .
Trained using guidance distillation, making FLUX.1 [dev]
more efficient.
Open weights to drive new scientific research, and empower artists to develop innovative workflows.
Generated outputs can be used for personal, scientific, and commercial purposes as described in the flux-1-dev-non-commercial-license.
We provide a reference implementation of FLUX.1 [dev]
, as well as sampling code, in a dedicated github repository. Developers and creatives looking to build on top of FLUX.1 [dev]
are encouraged to use this as a starting point.
The FLUX.1 models are also available via API from the following sources
bfl.ml (currently FLUX.1 [pro]
)
FLUX.1 [dev]
is also available in Comfy UI for local inference with a node-based workflow.
This model is not intended or able to provide factual information.
As a statistical model this checkpoint might amplify existing societal biases.
The model may fail to generate output that matches the prompts.
Prompt following is heavily influenced by the prompting-style.
The model and its derivatives may not be used
In any way that violates any applicable national, federal, state, local or international law or regulation.
For the purpose of exploiting, harming or attempting to exploit or harm minors in any way; including but not limited to the solicitation, creation, acquisition, or dissemination of child exploitative content.
To generate or disseminate verifiably false information and/or content with the purpose of harming others.
To generate or disseminate personal identifiable information that can be used to harm an individual.
To harass, abuse, threaten, stalk, or bully individuals or groups of individuals.
To create non-consensual nudity or illegal pornographic content.
For fully automated decision making that adversely impacts an individual's legal rights or otherwise creates or modifies a binding, enforceable obligation.
Generating or facilitating large-scale disinformation campaigns.
This model falls under the FLUX.1 [dev]
Non-Commercial License.