Revise object classification warnings in README
Updated README to clarify object classification limitations and added warning about vibe-coding.
This commit is contained in:
@@ -11,14 +11,16 @@
|
|||||||
|
|
||||||
</div>
|
</div>
|
||||||
|
|
||||||
> [!WARNING]
|
|
||||||
> **Regarding Object Classification**
|
|
||||||
>
|
|
||||||
> Frigate **does not support** uploading custom images for object classification training via the UI or API.
|
|
||||||
> This tool currently prepares the dataset (crops and categorizes images) for training external models (like YOLO) manually.
|
|
||||||
|
|
||||||
---
|
---
|
||||||
|
|
||||||
|
> [!WARNING]
|
||||||
|
> This is 100% vibe-coded.
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
## ⚡ Why This Tool?
|
## ⚡ Why This Tool?
|
||||||
|
|
||||||
> **"Diversity matters far more than volume."** — *Frigate Developer Tips*
|
> **"Diversity matters far more than volume."** — *Frigate Developer Tips*
|
||||||
@@ -34,6 +36,12 @@ Training AI models on "bulk" data is often harmful. If you feed the model 50 ima
|
|||||||
|
|
||||||
Both use the same **Farthest Point Sampling (FPS)** algorithm that mathematically selects images until redundancy starts, ensuring optimal diversity whether that's 20 or 150 images.
|
Both use the same **Farthest Point Sampling (FPS)** algorithm that mathematically selects images until redundancy starts, ensuring optimal diversity whether that's 20 or 150 images.
|
||||||
|
|
||||||
|
> [!WARNING]
|
||||||
|
> **Regarding Object Classification**
|
||||||
|
>
|
||||||
|
> Frigate **does not support** uploading custom images for object classification training via the UI or API.
|
||||||
|
> This tool currently prepares the dataset (crops and categorizes images) for training external models (like YOLO) manually.
|
||||||
|
|
||||||
---
|
---
|
||||||
|
|
||||||
## ✨ Features
|
## ✨ Features
|
||||||
|
|||||||
Reference in New Issue
Block a user