Adding dataset stats
c682b14 - configs Model at 210k steps, mlm acc 0.6509
- mc4 Model at 210k steps, mlm acc 0.6509
- outputs Step... (220001/250000 | Loss: 1.7591936588287354, Acc: 0.6520245671272278): 88%|βββββββββββββββββββ | 220518/250000 [106:16:53<14:56:57, 1.83s/it]
- 823 Bytes Model at 210k steps, mlm acc 0.6509
- 38 Bytes Model at 210k steps, mlm acc 0.6509
- 618 Bytes Model at 210k steps, mlm acc 0.6509
- 876 Bytes Model at 210k steps, mlm acc 0.6509
- 250 MB Step... (220001/250000 | Loss: 1.7591936588287354, Acc: 0.6520245671272278): 88%|βββββββββββββββββββ | 220518/250000 [106:16:53<14:56:57, 1.83s/it]
- 551 MB Adding dataset stats
- 514 kB Model at 210k steps, mlm acc 0.6509
- 499 MB Step... (220001/250000 | Loss: 1.7591936588287354, Acc: 0.6520245671272278): 88%|βββββββββββββββββββ | 220518/250000 [106:16:53<14:56:57, 1.83s/it]
- 30.8 kB Model at 210k steps, mlm acc 0.6509
- 930 Bytes Model at 210k steps, mlm acc 0.6509
- 239 Bytes Model at 210k steps, mlm acc 0.6509
- 1.47 MB Model at 210k steps, mlm acc 0.6509
- 292 Bytes Model at 210k steps, mlm acc 0.6509
- 855 kB Model at 210k steps, mlm acc 0.6509