This is tests of training a flux model, and all licensing conditions for the Flux Dev model apply. You can find the details here: <a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs/FLUX.1-dev">https://huggingface.co/black-forest-labs/FLUX.1-dev</a>. It's a great model!A huge thank you to <a target="_blank" rel="ugc" href="http://Datacrunch.io">Datacrunch.io</a> for generously providing their resources for training and support. I highly recommend them.This is Checkpoint. CLIP and VAE is already onThis is my model trained on Flux, and it aims to create more diverse and realistic faces. The results show improvements in anatomy, facial features, and overall realism. However, a side effect is that noise appears in the images. This can be corrected with any upscaler - just find one that works for you. The model was trained on approximately 7,000 images over the course of several hours using powerful H100 cards. If the test results are mostly positive, I will train it on a larger dataset.The number of training steps was limited due to budget constraints. Additionally, I used a small portion of my dataset, only 7,000 images, to reduce training time. When I get the opportunity, I plan to train a more refined model. You can use this model as you like, but please remember to adhere to the Black Forest Lab license terms.I welcome any feedback in the comments - feel free to share any errors or shortcomings of the model. This is just an experiment.

Flux_Realistic

This is the NF4 version. This is the checkpoint with Unet, Clip and VAE inside. I've made this version smaller for you. I hope you are happy and can test it.

NF4_Flux_realistic_v1_SaMay.safetensors

v1.0_NF4_Full_Model

Gguf_Flux_Realistic_SaMay

Flux_realistic_SaMay_v2

GGUF_Flux_Realistic_v1_Sa

v1.0_U-NET_ONLY

v1.0

Flux_Realistic_v1

Found a FLUX model on C Station that seems to specialize in portrait categories. It's a flux model based on FLUX DEV for secondary training, unsure of the effect, feel free to use it, paying tribute to lvmin zhang. The sampling method only supports FlowMatchEuler, ignoring euler, coefficient CFG recommended is 3.5, adjustable by yourself. Sampling steps 20, increase as needed.Reprint from https://civitai.com/models/167214/archilabv1This is a test of training the Flux model, all licensing conditions for Flux Dev models apply. You can find more details here: https://huggingface.co/black-forest-labs/FLUX.1-dev This is my model trained on Flux, and it aims to create more diverse and realistic faces. The results show improvements in anatomy, facial features, and overall realism. However, a side effect is that noise appears in the images. This can be corrected with any upscaler - just find one that works for you. The model was trained on approximately 7,000 images over the course of several hours using powerful H100 cards. If the test results are mostly positive, I will train it on a larger dataset.The number of training steps was limited due to budget constraints. Additionally, I used a small portion of my dataset, only 7,000 images, to reduce training time. When I get the opportunity, I plan to train a more refined model. You can use this model as you like, but please remember to adhere to the Black Forest Lab license terms. I welcome any feedback in the comments - feel free to share any errors or shortcomings of the model. This is just an experiment.

Files with SHA256: 34dafe4fed3e9ebbb824ff5c85e79bff4db41fbaa39b074beb31bcd5fc59f86c