Update tokenizer, configuration and training reports

reports/run/metrics.jsonl

@@ -990,3 +990,6 @@
 {"step": 19800, "tokens": 10380902400, "loss": 2.7468582153320313, "tokens_per_second": 982878.9220169957, "lr": 0.0015, "gradient_norm": 0.1454748809337616, "benchmark": false}
 {"step": 19820, "tokens": 10391388160, "loss": 2.77371826171875, "tokens_per_second": 988750.1280899943, "lr": 0.0015, "gradient_norm": 0.14740391075611115, "benchmark": false}
 {"step": 19840, "tokens": 10401873920, "loss": 2.7487648010253904, "tokens_per_second": 986397.7296181752, "lr": 0.0015, "gradient_norm": 0.15100659430027008, "benchmark": false}
+{"step": 19860, "tokens": 10412359680, "loss": 2.7202035903930666, "tokens_per_second": 990005.2628967476, "lr": 0.0015, "gradient_norm": 0.14142200350761414, "benchmark": false}
+{"step": 19880, "tokens": 10422845440, "loss": 2.743912124633789, "tokens_per_second": 980083.111584104, "lr": 0.0015, "gradient_norm": 0.1572934240102768, "benchmark": false}
+{"step": 19900, "tokens": 10433331200, "loss": 2.7287742614746096, "tokens_per_second": 975907.5155970105, "lr": 0.0015, "gradient_norm": 0.15567274391651154, "benchmark": false}