Commit History
-
Update README.md
Verified signature1778147 -
Update README.md
Verified signatured504604 -
Create README.md
Verified signaturec7d1b89 -
Delete README.md
Verified signature235c22d -
Create README.md
Verified signatureThis 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 ### Metrics I've also calculated the balanced F-score (F1) on each labels. 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 train/epoch: 60 train/loss:0.6065855598449708 ### Licnece The weights are in the public domain.ab506c3 -
Training in progress, step 3000
Verified signature9603884 -
Training in progress, step 2500
Verified signature1e54664 -
Training in progress, step 2000
Verified signaturef605999 -
Training in progress, step 1500
Verified signature7718c74 -
Training in progress, step 1000
Verified signature3ef0411 -
Training in progress, step 500
Verified signature6fee0e4 -
Training in progress, step 2190
Verified signature202447b -
Training in progress, step 2000
Verified signature2086a57 -
Training in progress, step 1500
Verified signaturec1253cf -
Training in progress, step 1000
Verified signature8c19ecd -
Training in progress, step 500
Verified signature18cbd0e -
initial commit
Verified signature2e4284f