Instructions to use PrunaAI/PrunaVAED with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use PrunaAI/PrunaVAED 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/PrunaVAED", torch_dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - LTX.io
How to use PrunaAI/PrunaVAED with LTX.io:
# Install the LTX-2 pipelines git clone https://github.com/Lightricks/LTX-2.git cd LTX-2 uv sync --frozen
# Download the weights from this repo, plus the Gemma text encoder hf download PrunaAI/PrunaVAED --local-dir models/PrunaVAED hf download google/gemma-3-12b-it-qat-q4_0-unquantized --local-dir models/gemma-3-12b
# Fast pipeline (distilled model, no distilled LoRA needed) uv run python -m ltx_pipelines.distilled \ --distilled-checkpoint-path models/PrunaVAED/<distilled-checkpoint>.safetensors \ --spatial-upsampler-path models/PrunaVAED/<spatial-upsampler>.safetensors \ --gemma-root models/gemma-3-12b \ --prompt "A beautiful sunset over the ocean" \ --output-path output.mp4 # For image-to-video, add: --image path/to/image.jpg 0 0.8# HQ pipeline (two-stage, higher quality) uv run python -m ltx_pipelines.ti2vid_two_stages_hq \ --checkpoint-path models/PrunaVAED/<checkpoint>.safetensors \ --distilled-lora models/PrunaVAED/<distilled-lora>.safetensors 0.8 \ --spatial-upsampler-path models/PrunaVAED/<spatial-upsampler>.safetensors \ --gemma-root models/gemma-3-12b \ --prompt "A beautiful sunset over the ocean" \ --output-path output.mp4 # For image-to-video, add: --image path/to/image.jpg 0 0.8 - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -127,19 +127,49 @@ pipeline and the
|
|
| 127 |
[`distilled`](https://github.com/Lightricks/LTX-2/blob/main/packages/ltx-pipelines/src/ltx_pipelines/distilled.py)
|
| 128 |
pipeline (recommended). (~4 s @ 24 fps)
|
| 129 |
|
| 130 |
-
|
| 131 |
-
|
| 132 |
-
|
| 133 |
-
|
| 134 |
-
|
| 135 |
-
|
| 136 |
-
|
| 137 |
-
|
| 138 |
-
|
| 139 |
-
|
| 140 |
-
|
| 141 |
-
|
| 142 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 143 |
|
| 144 |
**Δ-frame PSNR** is PSNR on consecutive-frame
|
| 145 |
differences (temporal consistency).
|
|
|
|
| 127 |
[`distilled`](https://github.com/Lightricks/LTX-2/blob/main/packages/ltx-pipelines/src/ltx_pipelines/distilled.py)
|
| 128 |
pipeline (recommended). (~4 s @ 24 fps)
|
| 129 |
|
| 130 |
+
<table style="width:100%">
|
| 131 |
+
<tr><th>Setting</th><th>Value</th></tr>
|
| 132 |
+
<tr><td>Model</td><td>PrunaVAED</td></tr>
|
| 133 |
+
<tr><td>Precision</td><td><code>bfloat16</code></td></tr>
|
| 134 |
+
<tr><td>Batch size</td><td>1</td></tr>
|
| 135 |
+
<tr><td>Device</td><td>NVIDIA H100 80GB</td></tr>
|
| 136 |
+
<tr><td>Decode</td><td>Full (no tiling / rolling / <code>torch.compile</code>)</td></tr>
|
| 137 |
+
</table>
|
| 138 |
+
|
| 139 |
+
<table style="width:100%">
|
| 140 |
+
<tr>
|
| 141 |
+
<th>Clip length</th><th>LTX-2.3 VAED</th><th>PrunaVAED</th><th>Speedup</th>
|
| 142 |
+
</tr>
|
| 143 |
+
<tr>
|
| 144 |
+
<td>5 s @ 720p</td><td align="right">841.3 ms</td><td align="right">500.2 ms</td><td align="right"><strong>1.68×</strong></td>
|
| 145 |
+
</tr>
|
| 146 |
+
<tr>
|
| 147 |
+
<td>10 s @ 720p</td><td align="right">1670.7 ms</td><td align="right">998.0 ms</td><td align="right"><strong>1.67×</strong></td>
|
| 148 |
+
</tr>
|
| 149 |
+
<tr>
|
| 150 |
+
<td>5 s @ 1080p</td><td align="right">1959.3 ms</td><td align="right">1152.1 ms</td><td align="right"><strong>1.7×</strong></td>
|
| 151 |
+
</tr>
|
| 152 |
+
<tr>
|
| 153 |
+
<td>10 s @ 1080p</td><td align="right">4903 ms*</td><td align="right">2356 ms</td><td align="right"><strong>2.08×</strong></td>
|
| 154 |
+
</tr>
|
| 155 |
+
</table>
|
| 156 |
+
|
| 157 |
+
\* Tiling required on H100; 14533ms otherwise
|
| 158 |
+
|
| 159 |
+
<table style="width:100%">
|
| 160 |
+
<tr>
|
| 161 |
+
<th>Latent set</th><th>Clips</th><th>PSNR ↑</th><th>LPIPS ↓</th><th>SSIM ↑</th><th>Δ-frame PSNR ↑</th>
|
| 162 |
+
</tr>
|
| 163 |
+
<tr>
|
| 164 |
+
<td>Distilled two-stage · 720p</td><td align="right">200</td><td align="right"><strong>37.87</strong></td><td align="right"><strong>0.0108</strong></td><td align="right"><strong>0.9898</strong></td><td align="right"><strong>37.28</strong></td>
|
| 165 |
+
</tr>
|
| 166 |
+
<tr>
|
| 167 |
+
<td>TI2Vid two-stage · 720p</td><td align="right">200</td><td align="right"><strong>38.86</strong></td><td align="right"><strong>0.0123</strong></td><td align="right"><strong>0.9803</strong></td><td align="right"><strong>38.98</strong></td>
|
| 168 |
+
</tr>
|
| 169 |
+
<tr>
|
| 170 |
+
<td>Distilled two-stage · 1080p</td><td align="right">200</td><td align="right"><strong>39.59</strong></td><td align="right"><strong>0.0067</strong></td><td align="right"><strong>0.9927</strong></td><td align="right"><strong>39.37</strong></td>
|
| 171 |
+
</tr>
|
| 172 |
+
</table>
|
| 173 |
|
| 174 |
**Δ-frame PSNR** is PSNR on consecutive-frame
|
| 175 |
differences (temporal consistency).
|