loading_adapters.md 17.8 KB
Newer Older
1
<!--Copyright 2024 The HuggingFace Team. All rights reserved.
Steven Liu's avatar
Steven Liu committed
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
-->

# Load adapters

[[open-in-colab]]

17
There are several [training](../training/overview) techniques for personalizing diffusion models to generate images of a specific subject or images in certain styles. Each of these training methods produces a different type of adapter. Some of the adapters generate an entirely new model, while other adapters only modify a smaller set of embeddings or weights. This means the loading process for each adapter is also different.
Steven Liu's avatar
Steven Liu committed
18
19
20
21
22

This guide will show you how to load DreamBooth, textual inversion, and LoRA weights.

<Tip>

23
Feel free to browse the [Stable Diffusion Conceptualizer](https://huggingface.co/spaces/sd-concepts-library/stable-diffusion-conceptualizer), [LoRA the Explorer](https://huggingface.co/spaces/multimodalart/LoraTheExplorer), and the [Diffusers Models Gallery](https://huggingface.co/spaces/huggingface-projects/diffusers-gallery) for checkpoints and embeddings to use.
Steven Liu's avatar
Steven Liu committed
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39

</Tip>

## DreamBooth

[DreamBooth](https://dreambooth.github.io/) finetunes an *entire diffusion model* on just several images of a subject to generate images of that subject in new styles and settings. This method works by using a special word in the prompt that the model learns to associate with the subject image. Of all the training methods, DreamBooth produces the largest file size (usually a few GBs) because it is a full checkpoint model.

Let's load the [herge_style](https://huggingface.co/sd-dreambooth-library/herge-style) checkpoint, which is trained on just 10 images drawn by Hergé, to generate images in that style. For it to work, you need to include the special word `herge_style` in your prompt to trigger the checkpoint:

```py
from diffusers import AutoPipelineForText2Image
import torch

pipeline = AutoPipelineForText2Image.from_pretrained("sd-dreambooth-library/herge-style", torch_dtype=torch.float16).to("cuda")
prompt = "A cute herge_style brown bear eating a slice of pizza, stunning color scheme, masterpiece, illustration"
image = pipeline(prompt).images[0]
40
image
Steven Liu's avatar
Steven Liu committed
41
42
43
44
45
46
47
48
```

<div class="flex justify-center">
    <img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/load_dreambooth.png" />
</div>

## Textual inversion

49
[Textual inversion](https://textual-inversion.github.io/) is very similar to DreamBooth and it can also personalize a diffusion model to generate certain concepts (styles, objects) from just a few images. This method works by training and finding new embeddings that represent the images you provide with a special word in the prompt. As a result, the diffusion model weights stay the same and the training process produces a relatively tiny (a few KBs) file.
Steven Liu's avatar
Steven Liu committed
50
51
52
53
54
55
56

Because textual inversion creates embeddings, it cannot be used on its own like DreamBooth and requires another model.

```py
from diffusers import AutoPipelineForText2Image
import torch

57
pipeline = AutoPipelineForText2Image.from_pretrained("stable-diffusion-v1-5/stable-diffusion-v1-5", torch_dtype=torch.float16).to("cuda")
Steven Liu's avatar
Steven Liu committed
58
59
60
61
62
63
64
65
```

Now you can load the textual inversion embeddings with the [`~loaders.TextualInversionLoaderMixin.load_textual_inversion`] method and generate some images. Let's load the [sd-concepts-library/gta5-artwork](https://huggingface.co/sd-concepts-library/gta5-artwork) embeddings and you'll need to include the special word `<gta5-artwork>` in your prompt to trigger it:

```py
pipeline.load_textual_inversion("sd-concepts-library/gta5-artwork")
prompt = "A cute brown bear eating a slice of pizza, stunning color scheme, masterpiece, illustration, <gta5-artwork> style"
image = pipeline(prompt).images[0]
66
image
Steven Liu's avatar
Steven Liu committed
67
68
69
70
71
72
```

<div class="flex justify-center">
    <img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/load_txt_embed.png" />
</div>

73
Textual inversion can also be trained on undesirable things to create *negative embeddings* to discourage a model from generating images with those undesirable things like blurry images or extra fingers on a hand. This can be an easy way to quickly improve your prompt. You'll also load the embeddings with [`~loaders.TextualInversionLoaderMixin.load_textual_inversion`], but this time, you'll need two more parameters:
Steven Liu's avatar
Steven Liu committed
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92

- `weight_name`: specifies the weight file to load if the file was saved in the 🤗 Diffusers format with a specific name or if the file is stored in the A1111 format
- `token`: specifies the special word to use in the prompt to trigger the embeddings

Let's load the [sayakpaul/EasyNegative-test](https://huggingface.co/sayakpaul/EasyNegative-test) embeddings:

```py
pipeline.load_textual_inversion(
    "sayakpaul/EasyNegative-test", weight_name="EasyNegative.safetensors", token="EasyNegative"
)
```

Now you can use the `token` to generate an image with the negative embeddings:

```py
prompt = "A cute brown bear eating a slice of pizza, stunning color scheme, masterpiece, illustration, EasyNegative"
negative_prompt = "EasyNegative"

image = pipeline(prompt, negative_prompt=negative_prompt, num_inference_steps=50).images[0]
93
image
Steven Liu's avatar
Steven Liu committed
94
95
96
97
98
99
100
101
102
103
104
105
```

<div class="flex justify-center">
    <img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/load_neg_embed.png" />
</div>

## LoRA

[Low-Rank Adaptation (LoRA)](https://huggingface.co/papers/2106.09685) is a popular training technique because it is fast and generates smaller file sizes (a couple hundred MBs). Like the other methods in this guide, LoRA can train a model to learn new styles from just a few images. It works by inserting new weights into the diffusion model and then only the new weights are trained instead of the entire model. This makes LoRAs faster to train and easier to store.

<Tip>

Steven Liu's avatar
Steven Liu committed
106
LoRA is a very general training technique that can be used with other training methods. For example, it is common to train a model with DreamBooth and LoRA. It is also increasingly common to load and merge multiple LoRAs to create new and unique images. You can learn more about it in the in-depth [Merge LoRAs](merge_loras) guide since merging is outside the scope of this loading guide.
Steven Liu's avatar
Steven Liu committed
107
108
109
110
111
112
113
114
115
116
117
118

</Tip>

LoRAs also need to be used with another model:

```py
from diffusers import AutoPipelineForText2Image
import torch

pipeline = AutoPipelineForText2Image.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16).to("cuda")
```

119
Then use the [`~loaders.StableDiffusionLoraLoaderMixin.load_lora_weights`] method to load the [ostris/super-cereal-sdxl-lora](https://huggingface.co/ostris/super-cereal-sdxl-lora) weights and specify the weights filename from the repository:
Steven Liu's avatar
Steven Liu committed
120
121
122
123
124

```py
pipeline.load_lora_weights("ostris/super-cereal-sdxl-lora", weight_name="cereal_box_sdxl_v1.safetensors")
prompt = "bears, pizza bites"
image = pipeline(prompt).images[0]
125
image
Steven Liu's avatar
Steven Liu committed
126
127
128
129
130
131
```

<div class="flex justify-center">
    <img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/load_lora.png" />
</div>

132
The [`~loaders.StableDiffusionLoraLoaderMixin.load_lora_weights`] method loads LoRA weights into both the UNet and text encoder. It is the preferred way for loading LoRAs because it can handle cases where:
Steven Liu's avatar
Steven Liu committed
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148

- the LoRA weights don't have separate identifiers for the UNet and text encoder
- the LoRA weights have separate identifiers for the UNet and text encoder

But if you only need to load LoRA weights into the UNet, then you can use the [`~loaders.UNet2DConditionLoadersMixin.load_attn_procs`] method. Let's load the [jbilcke-hf/sdxl-cinematic-1](https://huggingface.co/jbilcke-hf/sdxl-cinematic-1) LoRA:

```py
from diffusers import AutoPipelineForText2Image
import torch

pipeline = AutoPipelineForText2Image.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16).to("cuda")
pipeline.unet.load_attn_procs("jbilcke-hf/sdxl-cinematic-1", weight_name="pytorch_lora_weights.safetensors")

# use cnmt in the prompt to trigger the LoRA
prompt = "A cute cnmt eating a slice of pizza, stunning color scheme, masterpiece, illustration"
image = pipeline(prompt).images[0]
149
image
Steven Liu's avatar
Steven Liu committed
150
151
152
153
154
155
```

<div class="flex justify-center">
    <img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/load_attn_proc.png" />
</div>

156
To unload the LoRA weights, use the [`~loaders.StableDiffusionLoraLoaderMixin.unload_lora_weights`] method to discard the LoRA weights and restore the model to its original weights:
Steven Liu's avatar
Steven Liu committed
157
158
159
160
161

```py
pipeline.unload_lora_weights()
```

UmerHA's avatar
UmerHA committed
162
163
### Adjust LoRA weight scale

164
For both [`~loaders.StableDiffusionLoraLoaderMixin.load_lora_weights`] and [`~loaders.UNet2DConditionLoadersMixin.load_attn_procs`], you can pass the `cross_attention_kwargs={"scale": 0.5}` parameter to adjust how much of the LoRA weights to use. A value of `0` is the same as only using the base model weights, and a value of `1` is equivalent to using the fully finetuned LoRA.
UmerHA's avatar
UmerHA committed
165

166
For more granular control on the amount of LoRA weights used per layer, you can use [`~loaders.StableDiffusionLoraLoaderMixin.set_adapters`] and pass a dictionary specifying by how much to scale the weights in each layer by.
UmerHA's avatar
UmerHA committed
167
168
```python
pipe = ... # create pipeline
169
pipe.load_lora_weights(..., adapter_name="my_adapter")
UmerHA's avatar
UmerHA committed
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
scales = {
    "text_encoder": 0.5,
    "text_encoder_2": 0.5,  # only usable if pipe has a 2nd text encoder
    "unet": {
        "down": 0.9,  # all transformers in the down-part will use scale 0.9
        # "mid"  # in this example "mid" is not given, therefore all transformers in the mid part will use the default scale 1.0
        "up": {
            "block_0": 0.6,  # all 3 transformers in the 0th block in the up-part will use scale 0.6
            "block_1": [0.4, 0.8, 1.0],  # the 3 transformers in the 1st block in the up-part will use scales 0.4, 0.8 and 1.0 respectively
        }
    }
}
pipe.set_adapters("my_adapter", scales)
```

This also works with multiple adapters - see [this guide](https://huggingface.co/docs/diffusers/tutorials/using_peft_for_inference#customize-adapters-strength) for how to do it.

<Tip warning={true}>

189
Currently, [`~loaders.StableDiffusionLoraLoaderMixin.set_adapters`] only supports scaling attention weights. If a LoRA has other parts (e.g., resnets or down-/upsamplers), they will keep a scale of 1.0.
UmerHA's avatar
UmerHA committed
190
191
192

</Tip>

Steven Liu's avatar
Steven Liu committed
193
194
195
196
### Kohya and TheLastBen

Other popular LoRA trainers from the community include those by [Kohya](https://github.com/kohya-ss/sd-scripts/) and [TheLastBen](https://github.com/TheLastBen/fast-stable-diffusion). These trainers create different LoRA checkpoints than those trained by 🤗 Diffusers, but they can still be loaded in the same way.

Steven Liu's avatar
Steven Liu committed
197
198
199
200
<hfoptions id="other-trainers">
<hfoption id="Kohya">

To load a Kohya LoRA, let's download the [Blueprintify SD XL 1.0](https://civitai.com/models/150986/blueprintify-sd-xl-10) checkpoint from [Civitai](https://civitai.com/) as an example:
Steven Liu's avatar
Steven Liu committed
201

202
```sh
Steven Liu's avatar
Steven Liu committed
203
204
205
!wget https://civitai.com/api/download/models/168776 -O blueprintify-sd-xl-10.safetensors
```

206
Load the LoRA checkpoint with the [`~loaders.StableDiffusionLoraLoaderMixin.load_lora_weights`] method, and specify the filename in the `weight_name` parameter:
Steven Liu's avatar
Steven Liu committed
207
208
209
210
211

```py
from diffusers import AutoPipelineForText2Image
import torch

212
pipeline = AutoPipelineForText2Image.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16).to("cuda")
Steven Liu's avatar
Steven Liu committed
213
214
215
216
217
218
219
220
221
pipeline.load_lora_weights("path/to/weights", weight_name="blueprintify-sd-xl-10.safetensors")
```

Generate an image:

```py
# use bl3uprint in the prompt to trigger the LoRA
prompt = "bl3uprint, a highly detailed blueprint of the eiffel tower, explaining how to build all parts, many txt, blueprint grid backdrop"
image = pipeline(prompt).images[0]
222
image
Steven Liu's avatar
Steven Liu committed
223
224
225
226
227
228
```

<Tip warning={true}>

Some limitations of using Kohya LoRAs with 🤗 Diffusers include:

229
- Images may not look like those generated by UIs - like ComfyUI - for multiple reasons, which are explained [here](https://github.com/huggingface/diffusers/pull/4287/#issuecomment-1655110736).
230
- [LyCORIS checkpoints](https://github.com/KohakuBlueleaf/LyCORIS) aren't fully supported. The [`~loaders.StableDiffusionLoraLoaderMixin.load_lora_weights`] method loads LyCORIS checkpoints with LoRA and LoCon modules, but Hada and LoKR are not supported.
Steven Liu's avatar
Steven Liu committed
231
232
233

</Tip>

Steven Liu's avatar
Steven Liu committed
234
235
236
</hfoption>
<hfoption id="TheLastBen">

Steven Liu's avatar
Steven Liu committed
237
238
239
240
241
242
243
244
245
246
247
248
Loading a checkpoint from TheLastBen is very similar. For example, to load the [TheLastBen/William_Eggleston_Style_SDXL](https://huggingface.co/TheLastBen/William_Eggleston_Style_SDXL) checkpoint:

```py
from diffusers import AutoPipelineForText2Image
import torch

pipeline = AutoPipelineForText2Image.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16).to("cuda")
pipeline.load_lora_weights("TheLastBen/William_Eggleston_Style_SDXL", weight_name="wegg.safetensors")

# use by william eggleston in the prompt to trigger the LoRA
prompt = "a house by william eggleston, sunrays, beautiful, sunlight, sunrays, beautiful"
image = pipeline(prompt=prompt).images[0]
249
250
image
```
251

Steven Liu's avatar
Steven Liu committed
252
253
254
</hfoption>
</hfoptions>

Steven Liu's avatar
Steven Liu committed
255
## IP-Adapter
256

Steven Liu's avatar
Steven Liu committed
257
[IP-Adapter](https://ip-adapter.github.io/) is a lightweight adapter that enables image prompting for any diffusion model. This adapter works by decoupling the cross-attention layers of the image and text features. All the other model components are frozen and only the embedded image features in the UNet are trained. As a result, IP-Adapter files are typically only ~100MBs.
258

Steven Liu's avatar
Steven Liu committed
259
You can learn more about how to use IP-Adapter for different tasks and specific use cases in the [IP-Adapter](../using-diffusers/ip_adapter) guide.
260

Steven Liu's avatar
Steven Liu committed
261
262
263
> [!TIP]
> Diffusers currently only supports IP-Adapter for some of the most popular pipelines. Feel free to open a feature request if you have a cool use case and want to integrate IP-Adapter with an unsupported pipeline!
> Official IP-Adapter checkpoints are available from [h94/IP-Adapter](https://huggingface.co/h94/IP-Adapter).
264

Steven Liu's avatar
Steven Liu committed
265
To start, load a Stable Diffusion checkpoint.
266
267
268
269
270
271

```py
from diffusers import AutoPipelineForText2Image
import torch
from diffusers.utils import load_image

272
pipeline = AutoPipelineForText2Image.from_pretrained("stable-diffusion-v1-5/stable-diffusion-v1-5", torch_dtype=torch.float16).to("cuda")
273
274
```

Steven Liu's avatar
Steven Liu committed
275
Then load the IP-Adapter weights and add it to the pipeline with the [`~loaders.IPAdapterMixin.load_ip_adapter`] method.
276
277
278
279
280

```py
pipeline.load_ip_adapter("h94/IP-Adapter", subfolder="models", weight_name="ip-adapter_sd15.bin")
```

Steven Liu's avatar
Steven Liu committed
281
Once loaded, you can use the pipeline with an image and text prompt to guide the image generation process.
282
283
284
285
286

```py
image = load_image("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/load_neg_embed.png")
generator = torch.Generator(device="cpu").manual_seed(33)
images = pipeline(
M. Tolga Cangöz's avatar
M. Tolga Cangöz committed
287
    prompt='best quality, high quality, wearing sunglasses',
288
    ip_adapter_image=image,
M. Tolga Cangöz's avatar
M. Tolga Cangöz committed
289
    negative_prompt="monochrome, lowres, bad anatomy, worst quality, low quality",
290
291
    num_inference_steps=50,
    generator=generator,
Steven Liu's avatar
Steven Liu committed
292
293
).images[0]
images
294
295
296
297
298
299
```

<div class="flex justify-center">
    <img src="https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/ip-bear.png" />
</div>

Steven Liu's avatar
Steven Liu committed
300
### IP-Adapter Plus
301

M. Tolga Cangöz's avatar
M. Tolga Cangöz committed
302
IP-Adapter relies on an image encoder to generate image features. If the IP-Adapter repository contains an `image_encoder` subfolder, the image encoder is automatically loaded and registered to the pipeline. Otherwise, you'll need to explicitly load the image encoder with a [`~transformers.CLIPVisionModelWithProjection`] model and pass it to the pipeline.
303

Steven Liu's avatar
Steven Liu committed
304
This is the case for *IP-Adapter Plus* checkpoints which use the ViT-H image encoder.
305
306

```py
M. Tolga Cangöz's avatar
M. Tolga Cangöz committed
307
308
from transformers import CLIPVisionModelWithProjection

309
image_encoder = CLIPVisionModelWithProjection.from_pretrained(
Steven Liu's avatar
Steven Liu committed
310
    "h94/IP-Adapter",
311
    subfolder="models/image_encoder",
Steven Liu's avatar
Steven Liu committed
312
    torch_dtype=torch.float16
313
314
315
316
317
)

pipeline = AutoPipelineForText2Image.from_pretrained(
    "stabilityai/stable-diffusion-xl-base-1.0",
    image_encoder=image_encoder,
Steven Liu's avatar
Steven Liu committed
318
319
    torch_dtype=torch.float16
).to("cuda")
320

Steven Liu's avatar
Steven Liu committed
321
pipeline.load_ip_adapter("h94/IP-Adapter", subfolder="sdxl_models", weight_name="ip-adapter-plus_sdxl_vit-h.safetensors")
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

### IP-Adapter Face ID models

The IP-Adapter FaceID models are experimental IP Adapters that use image embeddings generated by `insightface` instead of CLIP image embeddings. Some of these models also use LoRA to improve ID consistency.
You need to install `insightface` and all its requirements to use these models.

<Tip warning={true}>
As InsightFace pretrained models are available for non-commercial research purposes, IP-Adapter-FaceID models are released exclusively for research purposes and are not intended for commercial use.
</Tip>

```py
pipeline = AutoPipelineForText2Image.from_pretrained(
    "stabilityai/stable-diffusion-xl-base-1.0",
    torch_dtype=torch.float16
).to("cuda")

pipeline.load_ip_adapter("h94/IP-Adapter-FaceID", subfolder=None, weight_name="ip-adapter-faceid_sdxl.bin", image_encoder_folder=None)
```

If you want to use one of the two IP-Adapter FaceID Plus models, you must also load the CLIP image encoder, as this models use both `insightface` and CLIP image embeddings to achieve better photorealism.

```py
from transformers import CLIPVisionModelWithProjection

image_encoder = CLIPVisionModelWithProjection.from_pretrained(
    "laion/CLIP-ViT-H-14-laion2B-s32B-b79K",
    torch_dtype=torch.float16,
)

pipeline = AutoPipelineForText2Image.from_pretrained(
353
    "stable-diffusion-v1-5/stable-diffusion-v1-5",
354
355
356
357
358
359
    image_encoder=image_encoder,
    torch_dtype=torch.float16
).to("cuda")

pipeline.load_ip_adapter("h94/IP-Adapter-FaceID", subfolder=None, weight_name="ip-adapter-faceid-plus_sd15.bin")
```