Merge pull request #16 from Eve-ning/deploy

Bump model to use 10K Dataset

README.md

@@ -6,15 +6,19 @@ opal is an accuracy-prediction model.
 It uses Neural Collaborative Filtering to learn associations between user and maps, then using those associations to
 predict new scores never before seen.
 
+**Performance Error Graph**
+![Performance Graph](models/V2_2023_01/error.png)
+
 ## Project Status
 
 Currently, it's in its early access, that means, it'll have many problems!
 
 However, we're working on it to minimize these issues o wo)b
 
+
 ## Dataset Used
 
-I used the top 1K mania users data from https://data.ppy.sh.
+I used the top 10K mania users data from https://data.ppy.sh.
 After preprocessing, we use
 - 1m scores for training
 - 120k scores for validation and testing each

models/README.md

@@ -2,15 +2,16 @@
 
 **See below on how to load the model.**
 
-| Model      | MAE   | RMSE  | Error Distribution             |
-|------------|-------|-------|--------------------------------|
-| V1_2022_11 | 0.96% | 1.59% | ![Error](V1_2022_11/error.png) |
+| Model      | R2     | MAE   | RMSE  | Error Distribution             |
+|------------|--------|-------|-------|--------------------------------|
+| V2_2023_01 | 77.77% | 1.23% | 1.90% | ![Error](V2_2023_01/error.png) |
 
+**Note: V1 is deprecated due to incompatibility with new model signature.**
 ## Limitations
 
 The model cannot ...
-- predict maps not played by at least 50 players within the top 1k
-- predict players not in the top 1k
+- predict maps not played by at least 50 players within the top 10k
+- predict players not in the top 10k
 - predict players who have not played at least 50 unique ranked maps in that year.
 
 The predictive power (i.e. accuracy) is dependent on the number of players associated with each map.