Text-to-Video
Diffusers
Safetensors
world-model
video-generation
image-to-video
interactive
distillation
Instructions to use AlayaLab/Evoke with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use AlayaLab/Evoke with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("AlayaLab/Evoke", 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
Document the weights layout at the repo root
Browse files- model_index.json +54 -0
model_index.json
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_comment": [
|
| 3 |
+
"This repository holds several EVOKE checkpoints. It is NOT a single loadable",
|
| 4 |
+
"diffusers pipeline root -- there is no top-level vae/ or transformer/, and",
|
| 5 |
+
"DiffusionPipeline.from_pretrained() on this repo id will not work.",
|
| 6 |
+
"",
|
| 7 |
+
"Every released model directory is the PARENT of a transformer/, and loads as",
|
| 8 |
+
" from_pretrained(<model_dir>, subfolder=\"transformer\")",
|
| 9 |
+
"with the shared components taken from evoke-base/.",
|
| 10 |
+
"",
|
| 11 |
+
"EVOKE also pins a development fork of diffusers that is not on PyPI. See",
|
| 12 |
+
"https://github.com/AlayaLab/Evoke for installation, weights layout and inference."
|
| 13 |
+
],
|
| 14 |
+
|
| 15 |
+
"_shipped_model": "evoke/stage3_post_distillation",
|
| 16 |
+
"_base_components": "evoke-base",
|
| 17 |
+
"_diffusers_fork_version": "0.37.0.dev0",
|
| 18 |
+
|
| 19 |
+
"_models": {
|
| 20 |
+
"evoke/stage1_camera_control": "multi-step camera-controllable model -- 50 steps, CFG 5.0",
|
| 21 |
+
"evoke/stage2_few_step_training": "few-step distillation -- 3-step pyramid",
|
| 22 |
+
"evoke/stage3_long_distillation": "30 s long-video distillation -- post-distill init",
|
| 23 |
+
"evoke/stage3_post_distillation": "the shipped model -- 3 steps, CFG-free",
|
| 24 |
+
"evoke/evoke_teacher": "dual-expert DMD teacher, {high,low}_noise -- training only"
|
| 25 |
+
},
|
| 26 |
+
|
| 27 |
+
"_components": {
|
| 28 |
+
"evoke/<model>/transformer": [
|
| 29 |
+
"evoke.modules.transformer_evoke",
|
| 30 |
+
"EvokeTransformer3DModel"
|
| 31 |
+
],
|
| 32 |
+
"evoke-base/vae": [
|
| 33 |
+
"diffusers",
|
| 34 |
+
"AutoencoderKLWan"
|
| 35 |
+
],
|
| 36 |
+
"evoke-base/text_encoder": [
|
| 37 |
+
"transformers",
|
| 38 |
+
"UMT5EncoderModel"
|
| 39 |
+
],
|
| 40 |
+
"evoke-base/tokenizer": [
|
| 41 |
+
"transformers",
|
| 42 |
+
"T5TokenizerFast"
|
| 43 |
+
],
|
| 44 |
+
"evoke-base/scheduler": [
|
| 45 |
+
"evoke.diffusers_version.scheduling_evoke_diffusers",
|
| 46 |
+
"EvokeScheduler"
|
| 47 |
+
]
|
| 48 |
+
},
|
| 49 |
+
|
| 50 |
+
"_external_weights": {
|
| 51 |
+
"ViGeo": "pkqbajng/ViGeo -- REQUIRED depth backend, downloaded separately",
|
| 52 |
+
"Depth-Anything-3": "depth-anything/da3-giant -- OPTIONAL, only for DEPTH_BACKEND=da3"
|
| 53 |
+
}
|
| 54 |
+
}
|