File size: 16,778 Bytes
d02c2a6 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 | import React from 'react';
import { ConfigDoc } from '@/types';
import { IoFlaskSharp } from 'react-icons/io5';
const docs: { [key: string]: ConfigDoc } = {
'config.name': {
title: 'Training Name',
description: (
<>
The name of the training job. This name will be used to identify the job in the system and will the the filename
of the final model. It must be unique and can only contain alphanumeric characters, underscores, and dashes. No
spaces or special characters are allowed.
</>
),
},
gpuids: {
title: 'GPU ID',
description: (
<>
This is the GPU that will be used for training. Only one GPU can be used per job at a time via the UI currently.
However, you can start multiple jobs in parallel, each using a different GPU.
</>
),
},
'config.process[0].trigger_word': {
title: 'Trigger Word',
description: (
<>
Optional: This will be the word or token used to trigger your concept or character.
<br />
<br />
When using a trigger word, If your captions do not contain the trigger word, it will be added automatically the
beginning of the caption. If you do not have captions, the caption will become just the trigger word. If you
want to have variable trigger words in your captions to put it in different spots, you can use the{' '}
<code>{'[trigger]'}</code> placeholder in your captions. This will be automatically replaced with your trigger
word.
<br />
<br />
Trigger words will not automatically be added to your test prompts, so you will need to either add your trigger
word manually or use the
<code>{'[trigger]'}</code> placeholder in your test prompts as well.
</>
),
},
'config.process[0].model.name_or_path': {
title: 'Name or Path',
description: (
<>
The name of a diffusers repo on Huggingface or the local path to the base model you want to train from. The
folder needs to be in diffusers format for most models. For some models, such as SDXL and SD1, you can put the
path to an all in one safetensors checkpoint here.
</>
),
},
'datasets.control_path': {
title: 'Control Dataset',
description: (
<>
The control dataset needs to have files that match the filenames of your training dataset. They should be
matching file pairs. These images are fed as control/input images during training. The control images will be
resized to match the training images.
</>
),
},
'datasets.multi_control_paths': {
title: 'Multi Control Dataset',
description: (
<>
The control dataset needs to have files that match the filenames of your training dataset. They should be
matching file pairs. These images are fed as control/input images during training.
<br />
<br />
For multi control datasets, the controls will all be applied in the order they are listed. If the model does not
require the images to be the same aspect ratios, such as with Qwen/Qwen-Image-Edit-2509, then the control images
do not need to match the aspect size or aspect ratio of the target image and they will be automatically resized
to the ideal resolutions for the model / target images.
</>
),
},
'datasets.num_frames': {
title: 'Number of Frames',
description: (
<>
This sets the number of frames to shrink videos to for a video dataset. If this dataset is images, set this to 1
for one frame. If your dataset is only videos, frames will be extracted evenly spaced from the videos in the
dataset.
<br />
<br />
It is best to trim your videos to the proper length before training. Wan is 16 frames a second. Doing 81 frames
will result in a 5 second video. So you would want all of your videos trimmed to around 5 seconds for best
results.
<br />
<br />
Example: Setting this to 81 and having 2 videos in your dataset, one is 2 seconds and one is 90 seconds long,
will result in 81 evenly spaced frames for each video making the 2 second video appear slow and the 90second
video appear very fast.
</>
),
},
'datasets.do_i2v': {
title: 'Do I2V',
description: (
<>
For video models that can handle both I2V (Image to Video) and T2V (Text to Video), this option sets this
dataset to be trained as an I2V dataset. This means that the first frame will be extracted from the video and
used as the start image for the video. If this option is not set, the dataset will be treated as a T2V dataset.
</>
),
},
'datasets.do_audio': {
title: 'Do Audio',
description: (
<>
For models that support audio with video, this option will load the audio from the video and resize it to match
the video sequence. Since the video is automatically resized, the audio may drop or raise in pitch to match the
new speed of the video. It is important to prep your dataset to have the proper length before training.
</>
),
},
'datasets.audio_normalize': {
title: 'Audio Normalize',
description: (
<>
When loading audio, this will normalize the audio volume to the max peaks. Useful if your dataset has varying
audio volumes. Warning, do not use if you have clips with full silence you want to keep, as it will raise the
volume of those clips.
</>
),
},
'datasets.audio_preserve_pitch': {
title: 'Audio Preserve Pitch',
description: (
<>
When loading audio to match the number of frames requested, this option will preserve the pitch of the audio if
the length does not match training target. It is recommended to have a dataset that matches your target length,
as this option can add sound distortions.
</>
),
},
'datasets.flip': {
title: 'Flip X and Flip Y',
description: (
<>
You can augment your dataset on the fly by flipping the x (horizontal) and/or y (vertical) axis. Flipping a
single axis will effectively double your dataset. It will result it training on normal images, and the flipped
versions of the images. This can be very helpful, but keep in mind it can also be destructive. There is no
reason to train people upside down, and flipping a face can confuse the model as a person's right side does not
look identical to their left side. For text, obviously flipping text is not a good idea.
<br />
<br />
Control images for a dataset will also be flipped to match the images, so they will always match on the pixel
level.
</>
),
},
'train.unload_text_encoder': {
title: 'Unload Text Encoder',
description: (
<>
Unloading text encoder will cache the trigger word and the sample prompts and unload the text encoder from the
GPU. Captions in for the dataset will be ignored
</>
),
},
'train.cache_text_embeddings': {
title: 'Cache Text Embeddings',
description: (
<>
<small>(experimental)</small>
<br />
Caching text embeddings will process and cache all the text embeddings from the text encoder to the disk. The
text encoder will be unloaded from the GPU. This does not work with things that dynamically change the prompt
such as trigger words, caption dropout, etc.
</>
),
},
'model.multistage': {
title: 'Stages to Train',
description: (
<>
Some models have multi stage networks that are trained and used separately in the denoising process. Most
common, is to have 2 stages. One for high noise and one for low noise. You can choose to train both stages at
once or train them separately. If trained at the same time, The trainer will alternate between training each
model every so many steps and will output 2 different LoRAs. If you choose to train only one stage, the trainer
will only train that stage and output a single LoRA.
</>
),
},
'train.switch_boundary_every': {
title: 'Switch Boundary Every',
description: (
<>
When training a model with multiple stages, this setting controls how often the trainer will switch between
training each stage.
<br />
<br />
For low vram settings, the model not being trained will be unloaded from the gpu to save memory. This takes some
time to do, so it is recommended to alternate less often when using low vram. A setting like 10 or 20 is
recommended for low vram settings.
<br />
<br />
The swap happens at the batch level, meaning it will swap between a gradient accumulation steps. To train both
stages in a single step, set them to switch every 1 step and set gradient accumulation to 2.
</>
),
},
'train.force_first_sample': {
title: 'Force First Sample',
description: (
<>
This option will force the trainer to generate samples when it starts. The trainer will normally only generate a
first sample when nothing has been trained yet, but will not do a first sample when resuming from an existing
checkpoint. This option forces a first sample every time the trainer is started. This can be useful if you have
changed sample prompts and want to see the new prompts right away.
</>
),
},
'model.layer_offloading': {
title: (
<>
Layer Offloading{' '}
<span className="text-yellow-500">
( <IoFlaskSharp className="inline text-yellow-500" name="Experimental" /> Experimental)
</span>
</>
),
description: (
<>
This is an experimental feature based on{' '}
<a className="text-blue-500" href="https://github.com/lodestone-rock/RamTorch" target="_blank">
RamTorch
</a>
. This feature is early and will have many updates and changes, so be aware it may not work consistently from
one update to the next. It will also only work with certain models.
<br />
<br />
Layer Offloading uses the CPU RAM instead of the GPU ram to hold most of the model weights. This allows training
a much larger model on a smaller GPU, assuming you have enough CPU RAM. This is slower than training on pure GPU
RAM, but CPU RAM is cheaper and upgradeable. You will still need GPU RAM to hold the optimizer states and LoRA
weights, so a larger card is usually still needed.
<br />
<br />
You can also select the percentage of the layers to offload. It is generally best to offload as few as possible
(close to 0%) for best performance, but you can offload more if you need the memory.
</>
),
},
'model.qie.match_target_res': {
title: 'Match Target Res',
description: (
<>
This setting will make the control images match the resolution of the target image. The official inference
example for Qwen-Image-Edit-2509 feeds the control image is at 1MP resolution, no matter what size you are
generating. Doing this makes training at lower res difficult because 1MP control images are fed in despite how
large your target image is. Match Target Res will match the resolution of your target to feed in the control
images allowing you to use less VRAM when training with smaller resolutions. You can still use different aspect
ratios, the image will just be resizes to match the amount of pixels in the target image.
</>
),
},
'train.diff_output_preservation': {
title: 'Differential Output Preservation',
description: (
<>
Differential Output Preservation (DOP) is a technique to help preserve class of the trained concept during
training. For this, you must have a trigger word set to differentiate your concept from its class. For instance,
You may be training a woman named Alice. Your trigger word may be "Alice". The class is "woman", since Alice is
a woman. We want to teach the model to remember what it knows about the class "woman" while teaching it what is
different about Alice. During training, the trainer will make a prediction with your LoRA bypassed and your
trigger word in the prompt replaced with the class word. Making "photo of Alice" become "photo of woman". This
prediction is called the prior prediction. Each step, we will do the normal training step, but also do another
step with this prior prediction and the class prompt in order to teach our LoRA to preserve the knowledge of the
class. This should not only improve the performance of your trained concept, but also allow you to do things
like "Alice standing next to a woman" and not make both of the people look like Alice.
</>
),
},
'train.blank_prompt_preservation': {
title: 'Blank Prompt Preservation',
description: (
<>
Blank Prompt Preservation (BPP) is a technique to help preserve the current models knowledge when unprompted.
This will not only help the model become more flexible, but will also help the quality of your concept during
inference, especially when a model uses CFG (Classifier Free Guidance) on inference. At each step during
training, a prior prediction is made with a blank prompt and with the LoRA disabled. This prediction is then
used as a target on an additional training step with a blank prompt, to preserve the model's knowledge when no
prompt is given. This helps the model to not overfit to the prompt and retain its generalization capabilities.
</>
),
},
'train.do_differential_guidance': {
title: 'Differential Guidance',
description: (
<>
Differential Guidance will amplify the difference of the model prediction and the target during training to make
a new target. Differential Guidance Scale will be the multiplier for the difference. This is still experimental,
but in my tests, it makes the model train faster, and learns details better in every scenario I have tried with
it.
<br />
<br />
The idea is that normal training inches closer to the target but never actually gets there, because it is
limited by the learning rate. With differential guidance, we amplify the difference for a new target beyond the
actual target, this would make the model learn to hit or overshoot the target instead of falling short.
<br />
<br />
<img src="/imgs/diff_guidance.png" alt="Differential Guidance Diagram" className="max-w-full mx-auto" />
</>
),
},
'dataset.num_repeats': {
title: 'Num Repeats',
description: (
<>
Number of Repeats will allow you to repeate the items in a dataset multiple times. This is useful when you are
using multiple datasets and want to balance the number of samples from each dataset. For instance, if you have a
small dataset of 10 images and a large dataset of 100 images, you can set the small dataset to have 10 repeats
to effectively make it 100 images, making the two datasets occour equally during training.
</>
),
},
'train.audio_loss_multiplier': {
title: 'Audio Loss Multiplier',
description: (
<>
When training audio and video, sometimes the video loss is so great that it outweights the audio loss, causing
the audio to become distorted. If you are noticing this happen, you can increase the audio loss multiplier to
give more weight to the audio loss. You could try something like 2.0, 10.0 etc. Warning, setting this too high
could overfit and damage the model.
</>
),
},
'datasets.auto_frame_count': {
title: 'Auto Frame Count',
description: (
<>
This will automatically determine the number of frames to use for each video in your dataset instead of relying
on a fixed num_frames. This allows you to include videos of different lengths in the dataset, and each video
will be processed without speeding up or slowing down. Be careful about adding long videos into your dataset, as
they use up more VRAM. This currently will not work with a batch size greater than 1.
</>
),
},
};
export const getDoc = (key: string | null | undefined): ConfigDoc | null => {
if (key && key in docs) {
return docs[key];
}
return null;
};
export default docs;
|