Update tokenizer, configuration and training reports

reports/run/metrics.jsonl

@@ -1055,3 +1055,5 @@
 {"step": 21100, "tokens": 11062476800, "loss": 2.7537540435791015, "tokens_per_second": 974425.6193885907, "lr": 0.0015, "gradient_norm": 0.17343269288539886, "benchmark": false}
 {"step": 21120, "tokens": 11072962560, "loss": 2.738983917236328, "tokens_per_second": 985374.9987583428, "lr": 0.0015, "gradient_norm": 0.158656507730484, "benchmark": false}
 {"step": 21140, "tokens": 11083448320, "loss": 2.7355117797851562, "tokens_per_second": 977298.3800180755, "lr": 0.0015, "gradient_norm": 0.1741272658109665, "benchmark": false}
+{"step": 21160, "tokens": 11093934080, "loss": 2.696071243286133, "tokens_per_second": 979845.1336802829, "lr": 0.0015, "gradient_norm": 0.1505553275346756, "benchmark": false}
+{"step": 21180, "tokens": 11104419840, "loss": 2.7300960540771486, "tokens_per_second": 975829.9134785949, "lr": 0.0015, "gradient_norm": 0.1768353432416916, "benchmark": false}