Update README.md

README.md

@@ -4,27 +4,31 @@ datasets:
 - lucrbrtv/doom-e1-gameplay
 library_name: transformers
 ---
-This is the best checkpoint of the Inverse Dynamic Model I've trained on DOOM (1993) gameplay. It's good enought to labelize data for a world model; it can still have a lot of improvements
+
+This is the best checkpoint of the Inverse Dynamic Model I've trained on DOOM (1993) gameplay. It's good enough to label data for a world model, though it can still be improved.
 
 ### Metrics
 
-I've also calculated the balanced F-score (F1) on each labels.
+I've calculated the balanced F-score (F1) for each label:
 
-eval/accuracy: 0.8335202932357788 (83,35%)
-eval/f1_ATTACK: 0.5636902451515198
-eval/f1_BCK: 0.5812191367149353
-eval/f1_FWD: 0.8092130422592163
-eval/f1_LEFT: 0.586035966873169
-eval/f1_RIGHT: 0.6066449284553528
-eval/f1_SPEED: 0 (I did not used this action in the training data)
-eval/f1_TURN_L: 0.5413390398025513
-eval/f1_TURN_R: 0.5131934285163879
-eval/f1_USE: 0.4528301954269409
-eval/loss: 0.37457531690597534
+| Action | F1 Score | Note |
+|--------|----------|------|
+| ATTACK | 0.564 | |
+| BCK | 0.581 | |
+| FWD | 0.809 ||
+| LEFT | 0.586 | |
+| RIGHT | 0.607 | |
+| SPEED | 0.000 | Not used in training |
+| TURN_L | 0.541 | |
+| TURN_R | 0.513 | |
+| USE | 0.453 | Lowest score |
+| Accuracy | 0.834| 83.35% |
+| Loss (eval) | 0.375 | |
 
-train/epoch: 60
-train/loss:0.6065855598449708
+Training:
+- Epoch: 60
+- Loss: 0.607
 
-### Licnece
+### License
 
 The weights are in the public domain.