Update README.md

README.md

@@ -36,16 +36,22 @@ Series of models aiming at detecting and segmenting face accurately. Trained on
 | Model                                                                      | Target                | mAP 50                        | mAP 50-95                 |Classes        |Dataset size|Training Resolution|
 |----------------------------------------------------------------------------|-----------------------|--------------------------------|---------------------------|---------------|------------|-------------------|
 | [Anzhc Face -seg.pt](https://huggingface.co/Anzhc/Anzhcs_YOLOs/blob/main/Anzhc%20Face%20-seg.pt)         | Face: illustration, real   | LOST DATA               | LOST DATA             |2(male, female)|LOST DATA| 640|
-| [Anzhc Face seg 640 v2 y8n.pt](https://huggingface.co/Anzhc/Anzhcs_YOLOs/blob/main/Anzhc%20Face%20seg%20640%20v2%20y8n.pt)   | Face: illustration, real   | 0.872(box),0.872(mask)  | 0.835(box),0.752(mask)|1(face)        |~500| 640|
-| [Anzhc Face seg 768 v2 y8n.pt](https://huggingface.co/Anzhc/Anzhcs_YOLOs/blob/main/Anzhc%20Face%20seg%20768%20v2%20y8n.pt)   | Face: illustration, real   | 0.86(box),0.86(mask)    | 0.81(box),0.726(mask) |1(face)        |~500| 768|
-| [Anzhc Face seg 768MS v2 y8n.pt](https://huggingface.co/Anzhc/Anzhcs_YOLOs/blob/main/Anzhc%20Face%20seg%20768MS%20v2%20y8n.pt) | Face: illustration, real   | 0.866(box),0.866(mask)  | 0.816(box),0.72(mask) |1(face)        |~500| 768|(Multi-scale)|
-| [Anzhc Face seg 1024 v2 y8n.pt](https://huggingface.co/Anzhc/Anzhcs_YOLOs/blob/main/Anzhc%20Face%20seg%201024%20v2%20y8n.pt)  | Face: illustration, real   | 0.872(box),0.872(mask)  | 0.804(box),0.726(mask)|1(face)        |~500| 1024|
+| [Anzhc Face seg 640 v2 y8n.pt](https://huggingface.co/Anzhc/Anzhcs_YOLOs/blob/main/Anzhc%20Face%20seg%20640%20v2%20y8n.pt)   | Face: illustration, real   |0.791(box) 0.765(mask)  | 0.608(box) 0.445(mask)|1(face)        |~500| 640|
+| [Anzhc Face seg 768 v2 y8n.pt](https://huggingface.co/Anzhc/Anzhcs_YOLOs/blob/main/Anzhc%20Face%20seg%20768%20v2%20y8n.pt)   | Face: illustration, real   | 0.765(box) 0.748(mask)    | 0.572(box) 0.431(mask) |1(face)        |~500| 768|
+| [Anzhc Face seg 768MS v2 y8n.pt](https://huggingface.co/Anzhc/Anzhcs_YOLOs/blob/main/Anzhc%20Face%20seg%20768MS%20v2%20y8n.pt) | Face: illustration, real   | 0.807(box) 0.770(mask)  | 0.601(box) 0.432(mask) |1(face)        |~500| 768|(Multi-scale)|
+| [Anzhc Face seg 1024 v2 y8n.pt](https://huggingface.co/Anzhc/Anzhcs_YOLOs/blob/main/Anzhc%20Face%20seg%201024%20v2%20y8n.pt)  | Face: illustration, real   | 0.768(box) 0.740(mask)  | 0.557(box) 0.394(mask)|1(face)        |~500| 1024|
+| [Anzhc Face seg 640 v3 y11n.pt]()  | Face: illustration   | 0.882(box) 0.871(mask)  | 0.689(box) 0.570(mask)|1(face)        |~660| 640|
 
 
-Take those stats with a grain of salt, since im pretty sure i re-scrambled dataset partition after training those models ages ago.
+UPDATE: v3 model has a bit different face target compared to v2, so stats of v2 models suffer compared to v3 in newer benchmark, especially in mask, while box is +- same.
+Dataset for v3 and above is going to be targeting inclusion of eyebrows and full eyelashes, for better adetailer experience without large dillution parameter.
+
+Also starting from v3, im moving to yolo11 models, as they seem to be direct upgrade over v8. v12 did not show significant improvement while requiring 50% more time to train, even with installed Flash Attention, so it's unlikely i will switch to it anytime soon.
+
 Benchmark was performed in 640px.
 Difference in v2 models are only in their target resolution, so their performance spread is marginal.
-![image/png](https://cdn-uploads.huggingface.co/production/uploads/633b43d29fe04b13f46c8988/G1vywwrYQOPSK9t3MZTen.png)
+
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/633b43d29fe04b13f46c8988/mKsZdVdt8UeQ2l8GDLfMX.png)
 
 ![image/png](https://cdn-uploads.huggingface.co/production/uploads/633b43d29fe04b13f46c8988/jMTKRWVk5y0HhrqqePdp-.png)