#0_DjM_Gallery_

#0_DjM_Gallery_

Golden Easter Egg - Most Run
Blue Easter Egg - 30 Run
Blue Easter Egg - Post
Poached Egg
im is me | №™ | Don't fool yourself with numbers | Listen to my album Bre spotify DJM230x
455
Followers
16
Following
96.2K
Runs
9
Downloads
7K
Likes
3.8K
Stars

Models

View All
855849125722218366
LORA IllustriousUpdated
EXCLUSIVE

DjM_ | gome style [FLUX|ILXL]-ilxl a2v1

187 6

VIDEO WAN_2_1
EARLY ACCESS

action-a1

32 2
854749213588763284
LORA FluxUpdated
EXCLUSIVE

DjM_ | #0_sci-fi_mode [FLUX]-v2

313 12
854723267691333234
LORA XLUpdated
EXCLUSIVE

DjM_ | ocxdjm [FLUX] [SD3.5M] [SD3.5L] [SDXL]-v2-a1-sdxl

456 12
855670949003924094
LORA IllustriousUpdated
EXCLUSIVE

DjM_ | DARKDJM [ILXL|SDXL]-a2.5

1.4K 14
853861046564204680
CHECKPOINT XLUpdated
EXCLUSIVE

DjM_ | dR-jm [SDXL]-v3

239 17

LORA XLUpdated
EXCLUSIVE

DjM_ | dBEAUTIFUL GIRLS [FLUX|SDXL]-sdxl v10

4.3K 84
853316315862041722
LORA Illustrious
EXCLUSIVE

DjM_ | dMMXXIV-i [ILXL|FLUX|SD3.5M|SD3.5L]-ilxl-v2x

898 32
851028364234880700
CHECKPOINT Illustrious
EXCLUSIVE

DjM_ | dmin-v1.5 nA

1.3K 51
852797058610913628
CHECKPOINT Pony
EXCLUSIVE

DjM_ | dpo-1

112 6
829802585402698998
CHECKPOINT Illustrious
EXCLUSIVE

DjM_ | dilustjm-A

14K 123
847110471445145999
CHECKPOINT Illustrious
EXCLUSIVE

DjM_ | dAn-v1-inva

694 30
835627196711224677
CHECKPOINT Illustrious
EXCLUSIVE

DjM_ | dReal [ilxl]-v1.5

909 33
831876427541558054
CHECKPOINT XL
EXCLUSIVE

DjM_ | dilusiJmXL-m1

261 12
828295167850273741
CHECKPOINT XL

DjM_ | dXjmX-illus-x-

563 14
824453006071303322
CHECKPOINT Pony
EXCLUSIVE

DjM_ | radjm-re-alpony [PNXL] UNET-album 1

803 26
823970732783598356
CHECKPOINT Illustrious
EXCLUSIVE

DjM_ | noGdjm [ILXL] UNET-album x3

5.5K 94
823979107965537031
CHECKPOINT Illustrious
EXCLUSIVE

DjM_ | mdjmIllustrioux [ILXL] UNET-album 1

2.8K 39
853236660324845803
LORA Illustrious
EXCLUSIVE

DjM_ | dbyou-1.0

217 32
850533466964545535
LORA Illustrious
EXCLUSIVE

DjM_ | MSKDJM [illustrious]-v1.0

78 4

Articles

View All
Explanation of Parameters and Settings in Model Training

Explanation of Parameters and Settings in Model Training

Explanation of Parameters and Settings in Model Training1. Trigger WordsTrigger words are keywords or phrases used in the text prompt to activate or guide the model in generating the desired image. For example, if we want to generate an image with a specific theme, we might use trigger words like "sunset," "robot," or "vintage."2. Image Processing ParametersRepeat: This parameter determines how many times the model will process or iterate over an image during training. A higher repeat value means more iterations for processing the image.Epoch: The number of cycles the model goes through during training, processing the entire dataset. Each epoch consists of one pass over all the images in the dataset.Save Every N Epochs: This specifies how frequently the model should save a checkpoint during training. For example, if set to 5, the model will save its progress every 5 epochs.Resolution: The size of the image used during training. This determines how large the images processed by the model will be, such as 512x512 or 1024x1024.3. Training ParametersSeed: A random value used to ensure consistent results across different runs. This is important to guarantee that the model can reproduce the same output.Clip Skip: This controls how many layers are skipped during training. Skipping layers can impact the final output quality and training speed.Text Encoder Learning Rate: The learning rate for the text encoder, which is responsible for converting text into vector representations that the model can understand.Unet Learning Rate: The learning rate for the U-Net, an important architecture in image processing for models like Stable Diffusion.LR Scheduler: An algorithm used to adjust the learning rate during training. This helps prevent the model from training too fast or too slow.lr_scheduler_num_cycles: The number of cycles in the learning rate adjustment process during training. This determines how the learning rate fluctuates over time.Num Warmup Steps: The number of steps in which the learning rate gradually increases before stabilizing at a higher value.Optimizer: The algorithm used to optimize the model weights during training, such as AdamW and othersNetwork Dim: The dimension of the neural network used by the model, typically referring to the size of layers and the number of neurons.Network Alpha: A scaling factor for controlling the learning process during training, adjusting the rate at which weights are updated.Gradient Accumulation Steps: The number of gradient accumulation steps taken before updating the weights. This is useful when working with limited memory or large batch sizes.4. Label ParametersShuffle Caption: Controls whether the caption labels should be shuffled during training. This can help prevent the model from becoming too dependent on the order of the text.Keep N Tokens: Specifies how many tokens should be retained in the caption, which can influence the efficiency of text processing.5. Advanced ParametersNoise Offset: Adjusts the noise level applied to the model, which can affect how the model generates images.Multires Noise Discount: Controls the reduction of multiresolution noise during training to improve stability.Multires Noise Iterations: Specifies how many iterations of multiresolution noise should be applied to the images.Conv Dim: The dimension of the convolutional kernels used by the model during image processing.Conv Alpha: Adjusts the strength of the convolutional influence in the network, which can affect the quality of the generated images.6. Sample Image SettingsPrompt: The text description or input provided to the model to generate an image. This is the main part of the input for text-based models.Image Size: The size of the image output by the model. This is related to the input resolution used during training and influences the output image dimensions.Sampler: The algorithm used to select points in the solution space during image generation. Different samplers can produce images with varying styles or quality.ConclusionEach of these parameters influences how the model operates, from training to image generation. A solid understanding of these settings allows for better customization of the model’s training process to generate images that are more accurate and meet specific needs.
🤸‍♂️Guide/Tutorial How to Use Radio Button for AI Tool in TensorArt. Step by step 🏃‍♂️‍➡️

🤸‍♂️Guide/Tutorial How to Use Radio Button for AI Tool in TensorArt. Step by step 🏃‍♂️‍➡️

Hello🙋‍♂️ This is just a brief explanation of how to use the Radio Button for the AI tool I use. Here’s the 🔗LINK to the AI tool I’m referring to. Alright, let’s jump right in, as I don’t have a proper introduction, haha.😂 I’m using TA Node - PromptText to create the Radio Button. Here’s how to set it up. First, grab the node from this icon: Then, look for TA Node - PromptText, like this: Once found, you’ll see the settings appear on the right, like this: Next, click Advanced settings, like this: The Description field will look like this when you use it: Choose the input type Radio Button, then click Add in the prompt settings. The display will look like this: You can choose whichever options you like. I’ll pick Action as an example, like this: I’ll take one option as an example. Here’s how it looks: Then, I edit it by clicking Add first, and it looks like this: You can adjust the Label and Prompt settings as you like. I’ve set them up like this: After that, click Add. The result will look like this: You can add as many options as you want by repeating the same steps. Once you’ve added everything you need, you can delete tags that you don’t need. For example, I don’t need this one: Here’s a tip if you want to follow my method, add first before deleting anything. Alright, moving on You can add more options later if you need them, using the same process. Here’s what the final result looks like: OK, as an addition, don't forget to connect TA Node - PromptText to the CLIP Text like this Alright, that’s it! Give it a try...You can leave a comment if you have any questions. I’ll answer them to the best of my knowledge.Also, let me know what you create using this method. I’d be thrilled if this helps you.I hope this is useful. Thank you!...
2

Posts