You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: TensorStack.Upscaler/README.md
+55-1Lines changed: 55 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -153,6 +153,7 @@ You can use the TensorStack.Audio package to restore audio from the source video
153
153
```
154
154
---
155
155
156
+
156
157
## Tiling Support
157
158
Tiling allows images and video frames to be processed in smaller sections (tiles) instead of all at once. This helps reduce memory usage and can improve performance when working with very large images or high-resolution videos.
158
159
@@ -182,4 +183,57 @@ Here is a list of some known and tested models compatible with `TensorStack.Upsc
TensorStack supports chaining multiple pipelines together using `IAsyncEnumerable` streams.
192
+
This allows complex video processing workflows—such as upscaling and frame interpolation—to be executed efficiently in a **single pass** without storing intermediate results in memory.
193
+
194
+
In the following example, the video is upscaled by **4×** and its frame rate is increased by **3×**:
195
+
```csharp
196
+
[nuget: TensorStack.Upscaler]
197
+
[nuget: TensorStack.Providers.DML]
198
+
[nuget: TensorStack.Video.Windows]
199
+
200
+
// Create Provider
201
+
varprovider=Provider.GetProvider();
202
+
203
+
// Upscaler Config
204
+
varupscaleConfig=newUpscalerConfig
205
+
{
206
+
ScaleFactor=4,
207
+
ExecutionProvider=provider,
208
+
Normalization=Normalization.ZeroToOne,
209
+
Path=@"M:\Models\RealESR-General-4x\model.onnx"
210
+
}
211
+
212
+
// Create Pipelines
213
+
using (varupscalePipeline=UpscalePipeline.Create(upscaleConfig))
214
+
using (varinterpolationPipeline=InterpolationPipeline.Create(provider))
0 commit comments