Update tokenizer, configuration and training reports

reports/run/metrics.jsonl

@@ -1096,3 +1096,5 @@
 {"step": 21920, "tokens": 11492392960, "loss": 2.7262691497802733, "tokens_per_second": 942358.7270998231, "lr": 0.0015, "gradient_norm": 0.1459842324256897, "benchmark": false}
 {"step": 21940, "tokens": 11502878720, "loss": 2.7390167236328127, "tokens_per_second": 979283.1353459321, "lr": 0.0015, "gradient_norm": 0.18871814012527466, "benchmark": false}
 {"step": 21960, "tokens": 11513364480, "loss": 2.7490068435668946, "tokens_per_second": 938317.9829802986, "lr": 0.0015, "gradient_norm": 0.1540951132774353, "benchmark": false}
+{"step": 21980, "tokens": 11523850240, "loss": 2.7099571228027344, "tokens_per_second": 972480.1144068335, "lr": 0.0015, "gradient_norm": 0.14107529819011688, "benchmark": false}
+{"step": 22000, "tokens": 11534336000, "loss": 2.7759721755981444, "tokens_per_second": 971667.9170136879, "lr": 0.0015, "gradient_norm": 0.17356452345848083, "benchmark": false}

reports/run/validation.jsonl

@@ -19,3 +19,4 @@
 {"step": 19000, "source_validation_loss": {"fineweb_edu": 2.920513153076172, "dclm_edu": 2.958641767501831, "cosmopedia": 1.972259521484375, "finemath": 2.146366834640503}}
 {"step": 20000, "source_validation_loss": {"fineweb_edu": 2.914907693862915, "dclm_edu": 2.9553182125091553, "cosmopedia": 1.9704746007919312, "finemath": 2.1503968238830566}}
 {"step": 21000, "source_validation_loss": {"fineweb_edu": 2.9123802185058594, "dclm_edu": 2.9512526988983154, "cosmopedia": 1.9636504650115967, "finemath": 2.145289421081543}}
+{"step": 22000, "source_validation_loss": {"fineweb_edu": 2.906125783920288, "dclm_edu": 2.946148633956909, "cosmopedia": 1.962751865386963, "finemath": 2.1453704833984375}}