Text-to-Image
Diffusers
Safetensors
MiniMax H3
Pruna AI
English
image-to-image
image-generation
image-editing
prompt-upsampling
modular-diffusers
Instructions to use PrunaAI/P-Image-OSS with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use PrunaAI/P-Image-OSS with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("PrunaAI/P-Image-OSS", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Pruna AI
How to use PrunaAI/P-Image-OSS with Pruna AI:
from pruna import PrunaModel pip install -U diffusers transformers accelerate
from pruna import PrunaModel import torch # switch to "mps" for apple devices pipe = PrunaModel.from_pretrained("PrunaAI/P-Image-OSS", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
update naming
Browse files- .gitattributes +4 -4
- NOTICE +3 -3
- README.md +19 -19
- assets/{P-H3-Image-latency.png → P-Image-OSS-latency.png} +0 -0
- assets/{P-H3-Image-memory.png → P-Image-OSS-memory.png} +0 -0
- assets/{P-H3-Image-weights.png → P-Image-OSS-weights.png} +0 -0
- assets/{p-h3-image.jpg → p-image-oss.jpg} +0 -0
- modular_model_index.json +10 -10
.gitattributes
CHANGED
|
@@ -39,10 +39,10 @@ tokenizer/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
|
| 39 |
prompt_upsampler_processor/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 40 |
assets/edits.jpg filter=lfs diff=lfs merge=lfs -text
|
| 41 |
assets/multi_image.jpg filter=lfs diff=lfs merge=lfs -text
|
| 42 |
-
assets/p-
|
| 43 |
assets/prompt.jpg filter=lfs diff=lfs merge=lfs -text
|
| 44 |
assets/resolution_ladder.jpg filter=lfs diff=lfs merge=lfs -text
|
| 45 |
assets/styles.jpg filter=lfs diff=lfs merge=lfs -text
|
| 46 |
-
assets/P-
|
| 47 |
-
assets/P-
|
| 48 |
-
assets/P-
|
|
|
|
| 39 |
prompt_upsampler_processor/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 40 |
assets/edits.jpg filter=lfs diff=lfs merge=lfs -text
|
| 41 |
assets/multi_image.jpg filter=lfs diff=lfs merge=lfs -text
|
| 42 |
+
assets/p-image-oss.jpg filter=lfs diff=lfs merge=lfs -text
|
| 43 |
assets/prompt.jpg filter=lfs diff=lfs merge=lfs -text
|
| 44 |
assets/resolution_ladder.jpg filter=lfs diff=lfs merge=lfs -text
|
| 45 |
assets/styles.jpg filter=lfs diff=lfs merge=lfs -text
|
| 46 |
+
assets/P-Image-OSS-latency.png filter=lfs diff=lfs merge=lfs -text
|
| 47 |
+
assets/P-Image-OSS-memory.png filter=lfs diff=lfs merge=lfs -text
|
| 48 |
+
assets/P-Image-OSS-weights.png filter=lfs diff=lfs merge=lfs -text
|
NOTICE
CHANGED
|
@@ -1,14 +1,14 @@
|
|
| 1 |
-
p-
|
| 2 |
Copyright 2026 Pruna AI
|
| 3 |
|
| 4 |
The files in this repository were modified by Pruna AI.
|
| 5 |
|
| 6 |
transformer/, vae/, text_connector/ and text_encoder/
|
| 7 |
-
p-
|
| 8 |
Community License Agreement, see LICENSE. The text encoder is also based on Qwen3-VL-4B by the Qwen team
|
| 9 |
(https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct); its Qwen3-VL-4B parts are under the Apache License 2.0, see
|
| 10 |
LICENSE-APACHE.
|
| 11 |
|
| 12 |
prompt_upsampler/
|
| 13 |
-
The p-
|
| 14 |
under the Apache License 2.0, see LICENSE-APACHE.
|
|
|
|
| 1 |
+
p-image-oss
|
| 2 |
Copyright 2026 Pruna AI
|
| 3 |
|
| 4 |
The files in this repository were modified by Pruna AI.
|
| 5 |
|
| 6 |
transformer/, vae/, text_connector/ and text_encoder/
|
| 7 |
+
p-image-oss, based on MiniMax-H3 by MiniMax (https://huggingface.co/MiniMaxAI/MiniMax-H3), under the MiniMax H3
|
| 8 |
Community License Agreement, see LICENSE. The text encoder is also based on Qwen3-VL-4B by the Qwen team
|
| 9 |
(https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct); its Qwen3-VL-4B parts are under the Apache License 2.0, see
|
| 10 |
LICENSE-APACHE.
|
| 11 |
|
| 12 |
prompt_upsampler/
|
| 13 |
+
The p-image-oss prompt upsampler, based on Qwen3.5-9B by the Qwen team (https://huggingface.co/Qwen/Qwen3.5-9B),
|
| 14 |
under the Apache License 2.0, see LICENSE-APACHE.
|
README.md
CHANGED
|
@@ -21,7 +21,7 @@ tags:
|
|
| 21 |
---
|
| 22 |
<!-- header start -->
|
| 23 |
<p align="center">
|
| 24 |
-
<img src="assets/p-
|
| 25 |
</p>
|
| 26 |
<!-- header end -->
|
| 27 |
|
|
@@ -48,7 +48,7 @@ tags:
|
|
| 48 |
</p>
|
| 49 |
<div align="center">
|
| 50 |
|
| 51 |
-
<h1 style="color: #9334E9;">P-
|
| 52 |
|
| 53 |
<h2>Image generation and editing based on MiniMax-H3</h2>
|
| 54 |
|
|
@@ -64,10 +64,10 @@ tags:
|
|
| 64 |
|
| 65 |
</div>
|
| 66 |
|
| 67 |
-
**P-
|
| 68 |
[MiniMax-H3](https://huggingface.co/MiniMaxAI/MiniMax-H3). It renders 8-megapixel images in 8 steps without
|
| 69 |
classifier-free guidance, edits with up to 8 input images, and comes with a prompt upsampler that turns a one-line
|
| 70 |
-
request into the full prompt P-
|
| 71 |
graphic styles with short quoted text.
|
| 72 |
|
| 73 |
This repository holds the [diffusers](https://github.com/huggingface/diffusers) pipeline. For ComfyUI, with int8 and
|
|
@@ -97,7 +97,7 @@ pip install -U transformers accelerate torch pillow
|
|
| 97 |
import torch
|
| 98 |
from diffusers import ModularPipeline
|
| 99 |
|
| 100 |
-
pipe = ModularPipeline.from_pretrained("PrunaAI/P-
|
| 101 |
pipe.load_components(dtype=torch.bfloat16)
|
| 102 |
pipe.to("cuda")
|
| 103 |
|
|
@@ -136,7 +136,7 @@ composed = pipe(
|
|
| 136 |
<img src="assets/resolutions.png" width="100%">
|
| 137 |
</p>
|
| 138 |
|
| 139 |
-
P-
|
| 140 |
in faces, hands and small details. For text-to-image, use one of the 8 MP sizes below. The default is 2912 × 2912.
|
| 141 |
|
| 142 |
Edits work best at 4 or 8 MP. Without `height` and `width`, the edit is generated at the size of Picture 1, scaled up to
|
|
@@ -178,7 +178,7 @@ resolutions because the noise changes with the image size.*
|
|
| 178 |
|
| 179 |
## Prompting
|
| 180 |
|
| 181 |
-
P-
|
| 182 |
therefore runs a prompt upsampler, based on Qwen3.5-9B, on every request. It writes a detailed prompt for the aspect
|
| 183 |
ratio of `height` and `width`, or, for an edit, a brief of what to keep and what to change. The prompt it wrote comes
|
| 184 |
back as the `prompt` output.
|
|
@@ -191,10 +191,10 @@ print(out["prompt"])
|
|
| 191 |
To render a prompt as written, remove the step from the blocks.
|
| 192 |
|
| 193 |
```python
|
| 194 |
-
blocks = ModularPipeline.from_pretrained("PrunaAI/P-
|
| 195 |
blocks.sub_blocks.pop("prompt_upsampler")
|
| 196 |
|
| 197 |
-
pipe = blocks.init_pipeline("PrunaAI/P-
|
| 198 |
pipe.load_components(dtype=torch.bfloat16)
|
| 199 |
pipe.to("cuda")
|
| 200 |
image = pipe(prompt=prompt, height=3360, width=2496, output="images")[0] # rendered as written
|
|
@@ -210,7 +210,7 @@ When you write prompts by hand, four things matter most.
|
|
| 210 |
3. **Spell out every word.** Text without its exact words turns into pseudo-text. Put each piece of text in quotes,
|
| 211 |
keep each string short, say where it sits, and end with "no other text". If you want no text, don't mention signs
|
| 212 |
or labels at all.
|
| 213 |
-
4. **Leave out hype words.** Stunning, epic, cinematic, vibrant, hyper-detailed, 8k or masterpiece push P-
|
| 214 |
toward a glossy, oversaturated, airbrushed look.
|
| 215 |
|
| 216 |
<p align="center">
|
|
@@ -233,7 +233,7 @@ text.
|
|
| 233 |
## Editing
|
| 234 |
|
| 235 |
Pass one image to edit it, or a list of up to 8. Picture 1 is the image to edit, and the others are references the
|
| 236 |
-
request can name as Picture 2, Picture 3 and so on. The upsampler writes the edit prompt in the form P-
|
| 237 |
best from, so a short request such as "make the pickup truck red" is enough.
|
| 238 |
|
| 239 |
<p align="center">
|
|
@@ -241,17 +241,17 @@ best from, so a short request such as "make the pickup truck red" is enough.
|
|
| 241 |
</p>
|
| 242 |
|
| 243 |
*The same short request given as typed (middle) and written as an edit prompt that names what to keep (right), the form
|
| 244 |
-
the upsampler writes. As typed, P-
|
| 245 |
changes.*
|
| 246 |
|
| 247 |
-
P-
|
| 248 |
-
picture and what to take from it, and P-
|
| 249 |
|
| 250 |
<p align="center">
|
| 251 |
<img src="assets/multi_image.jpg" width="100%">
|
| 252 |
</p>
|
| 253 |
|
| 254 |
-
*Top, eight public-domain portraits and the dinner party P-
|
| 255 |
Bottom, eight scientists at a poker table, seven presidents wading through the swamp with an astronaut, and characters
|
| 256 |
from our own renders in the same swamp.*
|
| 257 |
|
|
@@ -266,7 +266,7 @@ Edit <Picture 1>. <The change, described as the finished result>. Keep <everythi
|
|
| 266 |
- **Describe the result, not the action.** Say how the edited part looks in the finished image, and give new things a
|
| 267 |
size next to something already there, such as "about the size of the cat's head".
|
| 268 |
- **Say what stays.** List what must not change, such as faces, poses, the background and the light. Without this
|
| 269 |
-
list, P-
|
| 270 |
- **Quote new text exactly.** Put new text in quotes and name the text around it that must stay.
|
| 271 |
|
| 272 |
## Less GPU memory
|
|
@@ -281,7 +281,7 @@ from diffusers import ComponentsManager, ModularPipeline
|
|
| 281 |
manager = ComponentsManager()
|
| 282 |
manager.enable_auto_cpu_offload(device="cuda")
|
| 283 |
|
| 284 |
-
pipe = ModularPipeline.from_pretrained("PrunaAI/P-
|
| 285 |
pipe.load_components(dtype=torch.bfloat16)
|
| 286 |
image = pipe(prompt=prompt, height=3360, width=2496, output="images")[0]
|
| 287 |
```
|
|
@@ -308,11 +308,11 @@ The 8 steps run over the checkpoint's own sampling grid. Pass `sigmas` to use an
|
|
| 308 |
longer passages sometimes get single letters wrong. Short quoted strings render well.
|
| 309 |
- **Lower resolutions show more artifacts.** At 2 and 4 MP, faces, hands and small details break more often than at
|
| 310 |
8 MP.
|
| 311 |
-
- **Single images only.** P-
|
| 312 |
|
| 313 |
## License
|
| 314 |
|
| 315 |
-
P-
|
| 316 |
use. Among other terms, it excludes the European Union, the United Kingdom, the Republic of Korea and the United States
|
| 317 |
unless you have a separate license from MiniMax, and products built on it must show "Powered by MiniMax H3". The text
|
| 318 |
encoder's Qwen3-VL-4B parts and the prompt upsampler, based on Qwen3.5-9B, are under the
|
|
|
|
| 21 |
---
|
| 22 |
<!-- header start -->
|
| 23 |
<p align="center">
|
| 24 |
+
<img src="assets/p-image-oss.jpg" width="800">
|
| 25 |
</p>
|
| 26 |
<!-- header end -->
|
| 27 |
|
|
|
|
| 48 |
</p>
|
| 49 |
<div align="center">
|
| 50 |
|
| 51 |
+
<h1 style="color: #9334E9;">P-Image-OSS</h1>
|
| 52 |
|
| 53 |
<h2>Image generation and editing based on MiniMax-H3</h2>
|
| 54 |
|
|
|
|
| 64 |
|
| 65 |
</div>
|
| 66 |
|
| 67 |
+
**P-Image-OSS** is an image model for text-to-image generation and image editing, based on
|
| 68 |
[MiniMax-H3](https://huggingface.co/MiniMaxAI/MiniMax-H3). It renders 8-megapixel images in 8 steps without
|
| 69 |
classifier-free guidance, edits with up to 8 input images, and comes with a prompt upsampler that turns a one-line
|
| 70 |
+
request into the full prompt P-Image-OSS renders best. It is strongest at photorealism, film and editorial looks, and
|
| 71 |
graphic styles with short quoted text.
|
| 72 |
|
| 73 |
This repository holds the [diffusers](https://github.com/huggingface/diffusers) pipeline. For ComfyUI, with int8 and
|
|
|
|
| 97 |
import torch
|
| 98 |
from diffusers import ModularPipeline
|
| 99 |
|
| 100 |
+
pipe = ModularPipeline.from_pretrained("PrunaAI/P-Image-OSS")
|
| 101 |
pipe.load_components(dtype=torch.bfloat16)
|
| 102 |
pipe.to("cuda")
|
| 103 |
|
|
|
|
| 136 |
<img src="assets/resolutions.png" width="100%">
|
| 137 |
</p>
|
| 138 |
|
| 139 |
+
P-Image-OSS works best at **8 MP**. Images at 2 and 4 MP can be good too, but artifacts become more frequent, especially
|
| 140 |
in faces, hands and small details. For text-to-image, use one of the 8 MP sizes below. The default is 2912 × 2912.
|
| 141 |
|
| 142 |
Edits work best at 4 or 8 MP. Without `height` and `width`, the edit is generated at the size of Picture 1, scaled up to
|
|
|
|
| 178 |
|
| 179 |
## Prompting
|
| 180 |
|
| 181 |
+
P-Image-OSS is sensitive to prompting and renders full, specific prompts much better than short requests. The pipeline
|
| 182 |
therefore runs a prompt upsampler, based on Qwen3.5-9B, on every request. It writes a detailed prompt for the aspect
|
| 183 |
ratio of `height` and `width`, or, for an edit, a brief of what to keep and what to change. The prompt it wrote comes
|
| 184 |
back as the `prompt` output.
|
|
|
|
| 191 |
To render a prompt as written, remove the step from the blocks.
|
| 192 |
|
| 193 |
```python
|
| 194 |
+
blocks = ModularPipeline.from_pretrained("PrunaAI/P-Image-OSS").blocks
|
| 195 |
blocks.sub_blocks.pop("prompt_upsampler")
|
| 196 |
|
| 197 |
+
pipe = blocks.init_pipeline("PrunaAI/P-Image-OSS")
|
| 198 |
pipe.load_components(dtype=torch.bfloat16)
|
| 199 |
pipe.to("cuda")
|
| 200 |
image = pipe(prompt=prompt, height=3360, width=2496, output="images")[0] # rendered as written
|
|
|
|
| 210 |
3. **Spell out every word.** Text without its exact words turns into pseudo-text. Put each piece of text in quotes,
|
| 211 |
keep each string short, say where it sits, and end with "no other text". If you want no text, don't mention signs
|
| 212 |
or labels at all.
|
| 213 |
+
4. **Leave out hype words.** Stunning, epic, cinematic, vibrant, hyper-detailed, 8k or masterpiece push P-Image-OSS
|
| 214 |
toward a glossy, oversaturated, airbrushed look.
|
| 215 |
|
| 216 |
<p align="center">
|
|
|
|
| 233 |
## Editing
|
| 234 |
|
| 235 |
Pass one image to edit it, or a list of up to 8. Picture 1 is the image to edit, and the others are references the
|
| 236 |
+
request can name as Picture 2, Picture 3 and so on. The upsampler writes the edit prompt in the form P-Image-OSS edits
|
| 237 |
best from, so a short request such as "make the pickup truck red" is enough.
|
| 238 |
|
| 239 |
<p align="center">
|
|
|
|
| 241 |
</p>
|
| 242 |
|
| 243 |
*The same short request given as typed (middle) and written as an edit prompt that names what to keep (right), the form
|
| 244 |
+
the upsampler writes. As typed, P-Image-OSS often redraws the whole scene. As an edit prompt, only the requested part
|
| 245 |
changes.*
|
| 246 |
|
| 247 |
+
P-Image-OSS combines many references at once. Given up to 8 pictures and a one-line request, the upsampler names each
|
| 248 |
+
picture and what to take from it, and P-Image-OSS composes them into one scene.
|
| 249 |
|
| 250 |
<p align="center">
|
| 251 |
<img src="assets/multi_image.jpg" width="100%">
|
| 252 |
</p>
|
| 253 |
|
| 254 |
+
*Top, eight public-domain portraits and the dinner party P-Image-OSS composed from them, from a one-sentence request.
|
| 255 |
Bottom, eight scientists at a poker table, seven presidents wading through the swamp with an astronaut, and characters
|
| 256 |
from our own renders in the same swamp.*
|
| 257 |
|
|
|
|
| 266 |
- **Describe the result, not the action.** Say how the edited part looks in the finished image, and give new things a
|
| 267 |
size next to something already there, such as "about the size of the cat's head".
|
| 268 |
- **Say what stays.** List what must not change, such as faces, poses, the background and the light. Without this
|
| 269 |
+
list, P-Image-OSS tends to redraw the whole image.
|
| 270 |
- **Quote new text exactly.** Put new text in quotes and name the text around it that must stay.
|
| 271 |
|
| 272 |
## Less GPU memory
|
|
|
|
| 281 |
manager = ComponentsManager()
|
| 282 |
manager.enable_auto_cpu_offload(device="cuda")
|
| 283 |
|
| 284 |
+
pipe = ModularPipeline.from_pretrained("PrunaAI/P-Image-OSS", components_manager=manager)
|
| 285 |
pipe.load_components(dtype=torch.bfloat16)
|
| 286 |
image = pipe(prompt=prompt, height=3360, width=2496, output="images")[0]
|
| 287 |
```
|
|
|
|
| 308 |
longer passages sometimes get single letters wrong. Short quoted strings render well.
|
| 309 |
- **Lower resolutions show more artifacts.** At 2 and 4 MP, faces, hands and small details break more often than at
|
| 310 |
8 MP.
|
| 311 |
+
- **Single images only.** P-Image-OSS does not generate video or audio.
|
| 312 |
|
| 313 |
## License
|
| 314 |
|
| 315 |
+
P-Image-OSS is based on MiniMax-H3 and released under the [MiniMax H3 Community License](LICENSE). Please read it before
|
| 316 |
use. Among other terms, it excludes the European Union, the United Kingdom, the Republic of Korea and the United States
|
| 317 |
unless you have a separate license from MiniMax, and products built on it must show "Powered by MiniMax H3". The text
|
| 318 |
encoder's Qwen3-VL-4B parts and the prompt upsampler, based on Qwen3.5-9B, are under the
|
assets/{P-H3-Image-latency.png → P-Image-OSS-latency.png}
RENAMED
|
File without changes
|
assets/{P-H3-Image-memory.png → P-Image-OSS-memory.png}
RENAMED
|
File without changes
|
assets/{P-H3-Image-weights.png → P-Image-OSS-weights.png}
RENAMED
|
File without changes
|
assets/{p-h3-image.jpg → p-image-oss.jpg}
RENAMED
|
File without changes
|
modular_model_index.json
CHANGED
|
@@ -25,7 +25,7 @@
|
|
| 25 |
"transformers",
|
| 26 |
"Qwen3VLForConditionalGeneration"
|
| 27 |
],
|
| 28 |
-
"pretrained_model_name_or_path": "PrunaAI/P-
|
| 29 |
"subfolder": "text_encoder",
|
| 30 |
"variant": null,
|
| 31 |
"revision": null
|
|
@@ -39,7 +39,7 @@
|
|
| 39 |
"diffusers",
|
| 40 |
"PH3ImageTextConnector"
|
| 41 |
],
|
| 42 |
-
"pretrained_model_name_or_path": "PrunaAI/P-
|
| 43 |
"subfolder": "text_connector",
|
| 44 |
"variant": null,
|
| 45 |
"revision": null
|
|
@@ -53,7 +53,7 @@
|
|
| 53 |
"transformers",
|
| 54 |
"Qwen2TokenizerFast"
|
| 55 |
],
|
| 56 |
-
"pretrained_model_name_or_path": "PrunaAI/P-
|
| 57 |
"subfolder": "tokenizer",
|
| 58 |
"variant": null,
|
| 59 |
"revision": null
|
|
@@ -67,7 +67,7 @@
|
|
| 67 |
"transformers",
|
| 68 |
"Qwen3VLProcessor"
|
| 69 |
],
|
| 70 |
-
"pretrained_model_name_or_path": "PrunaAI/P-
|
| 71 |
"subfolder": "processor",
|
| 72 |
"variant": null,
|
| 73 |
"revision": null
|
|
@@ -81,7 +81,7 @@
|
|
| 81 |
"diffusers",
|
| 82 |
"AutoencoderKLPH3Image"
|
| 83 |
],
|
| 84 |
-
"pretrained_model_name_or_path": "PrunaAI/P-
|
| 85 |
"subfolder": "vae",
|
| 86 |
"variant": null,
|
| 87 |
"revision": null
|
|
@@ -95,7 +95,7 @@
|
|
| 95 |
"diffusers",
|
| 96 |
"MiniMaxH3Transformer3DModel"
|
| 97 |
],
|
| 98 |
-
"pretrained_model_name_or_path": "PrunaAI/P-
|
| 99 |
"subfolder": "transformer",
|
| 100 |
"variant": null,
|
| 101 |
"revision": null
|
|
@@ -109,7 +109,7 @@
|
|
| 109 |
"diffusers",
|
| 110 |
"MiniMaxH3Scheduler"
|
| 111 |
],
|
| 112 |
-
"pretrained_model_name_or_path": "PrunaAI/P-
|
| 113 |
"subfolder": "scheduler",
|
| 114 |
"variant": null,
|
| 115 |
"revision": null
|
|
@@ -123,7 +123,7 @@
|
|
| 123 |
"diffusers",
|
| 124 |
"MiniMaxH3Scheduler"
|
| 125 |
],
|
| 126 |
-
"pretrained_model_name_or_path": "PrunaAI/P-
|
| 127 |
"subfolder": "audio_scheduler",
|
| 128 |
"variant": null,
|
| 129 |
"revision": null
|
|
@@ -133,7 +133,7 @@
|
|
| 133 |
"transformers",
|
| 134 |
"Qwen3_5ForConditionalGeneration",
|
| 135 |
{
|
| 136 |
-
"pretrained_model_name_or_path": "PrunaAI/P-
|
| 137 |
"revision": null,
|
| 138 |
"subfolder": "prompt_upsampler",
|
| 139 |
"type_hint": [
|
|
@@ -147,7 +147,7 @@
|
|
| 147 |
"transformers",
|
| 148 |
"AutoProcessor",
|
| 149 |
{
|
| 150 |
-
"pretrained_model_name_or_path": "PrunaAI/P-
|
| 151 |
"revision": null,
|
| 152 |
"subfolder": "prompt_upsampler_processor",
|
| 153 |
"type_hint": [
|
|
|
|
| 25 |
"transformers",
|
| 26 |
"Qwen3VLForConditionalGeneration"
|
| 27 |
],
|
| 28 |
+
"pretrained_model_name_or_path": "PrunaAI/P-Image-OSS",
|
| 29 |
"subfolder": "text_encoder",
|
| 30 |
"variant": null,
|
| 31 |
"revision": null
|
|
|
|
| 39 |
"diffusers",
|
| 40 |
"PH3ImageTextConnector"
|
| 41 |
],
|
| 42 |
+
"pretrained_model_name_or_path": "PrunaAI/P-Image-OSS",
|
| 43 |
"subfolder": "text_connector",
|
| 44 |
"variant": null,
|
| 45 |
"revision": null
|
|
|
|
| 53 |
"transformers",
|
| 54 |
"Qwen2TokenizerFast"
|
| 55 |
],
|
| 56 |
+
"pretrained_model_name_or_path": "PrunaAI/P-Image-OSS",
|
| 57 |
"subfolder": "tokenizer",
|
| 58 |
"variant": null,
|
| 59 |
"revision": null
|
|
|
|
| 67 |
"transformers",
|
| 68 |
"Qwen3VLProcessor"
|
| 69 |
],
|
| 70 |
+
"pretrained_model_name_or_path": "PrunaAI/P-Image-OSS",
|
| 71 |
"subfolder": "processor",
|
| 72 |
"variant": null,
|
| 73 |
"revision": null
|
|
|
|
| 81 |
"diffusers",
|
| 82 |
"AutoencoderKLPH3Image"
|
| 83 |
],
|
| 84 |
+
"pretrained_model_name_or_path": "PrunaAI/P-Image-OSS",
|
| 85 |
"subfolder": "vae",
|
| 86 |
"variant": null,
|
| 87 |
"revision": null
|
|
|
|
| 95 |
"diffusers",
|
| 96 |
"MiniMaxH3Transformer3DModel"
|
| 97 |
],
|
| 98 |
+
"pretrained_model_name_or_path": "PrunaAI/P-Image-OSS",
|
| 99 |
"subfolder": "transformer",
|
| 100 |
"variant": null,
|
| 101 |
"revision": null
|
|
|
|
| 109 |
"diffusers",
|
| 110 |
"MiniMaxH3Scheduler"
|
| 111 |
],
|
| 112 |
+
"pretrained_model_name_or_path": "PrunaAI/P-Image-OSS",
|
| 113 |
"subfolder": "scheduler",
|
| 114 |
"variant": null,
|
| 115 |
"revision": null
|
|
|
|
| 123 |
"diffusers",
|
| 124 |
"MiniMaxH3Scheduler"
|
| 125 |
],
|
| 126 |
+
"pretrained_model_name_or_path": "PrunaAI/P-Image-OSS",
|
| 127 |
"subfolder": "audio_scheduler",
|
| 128 |
"variant": null,
|
| 129 |
"revision": null
|
|
|
|
| 133 |
"transformers",
|
| 134 |
"Qwen3_5ForConditionalGeneration",
|
| 135 |
{
|
| 136 |
+
"pretrained_model_name_or_path": "PrunaAI/P-Image-OSS",
|
| 137 |
"revision": null,
|
| 138 |
"subfolder": "prompt_upsampler",
|
| 139 |
"type_hint": [
|
|
|
|
| 147 |
"transformers",
|
| 148 |
"AutoProcessor",
|
| 149 |
{
|
| 150 |
+
"pretrained_model_name_or_path": "PrunaAI/P-Image-OSS",
|
| 151 |
"revision": null,
|
| 152 |
"subfolder": "prompt_upsampler_processor",
|
| 153 |
"type_hint": [
|