Update tokenizer, configuration and training reports

reports/run/metrics.jsonl

@@ -197,3 +197,5 @@
 {"step": 3940, "tokens": 2065694720, "loss": 3.01178035736084, "tokens_per_second": 980920.9083517338, "lr": 0.0015, "gradient_norm": 0.09572461247444153, "benchmark": false}
 {"step": 3960, "tokens": 2076180480, "loss": 3.002391815185547, "tokens_per_second": 982260.303099604, "lr": 0.0015, "gradient_norm": 0.11502530425786972, "benchmark": false}
 {"step": 3980, "tokens": 2086666240, "loss": 2.9722618103027343, "tokens_per_second": 936205.2487208435, "lr": 0.0015, "gradient_norm": 0.13280029594898224, "benchmark": false}
+{"step": 4000, "tokens": 2097152000, "loss": 3.018246078491211, "tokens_per_second": 980606.9515894684, "lr": 0.0015, "gradient_norm": 0.1104731634259224, "benchmark": false}
+{"step": 4020, "tokens": 2107637760, "loss": 3.0328603744506837, "tokens_per_second": 817031.6501473963, "lr": 0.0015, "gradient_norm": 0.10828843712806702, "benchmark": false}

reports/run/validation.jsonl

@@ -1,3 +1,4 @@
 {"step": 1000, "source_validation_loss": {"fineweb_edu": 3.6357946395874023, "dclm_edu": 3.6318132877349854, "cosmopedia": 2.9174532890319824, "finemath": 2.940213680267334}}
 {"step": 2000, "source_validation_loss": {"fineweb_edu": 3.3628950119018555, "dclm_edu": 3.3812646865844727, "cosmopedia": 2.551811456680298, "finemath": 2.6517066955566406}}
 {"step": 3000, "source_validation_loss": {"fineweb_edu": 3.2158007621765137, "dclm_edu": 3.24115252494812, "cosmopedia": 2.362274169921875, "finemath": 2.494218349456787}}
+{"step": 4000, "source_validation_loss": {"fineweb_edu": 3.1462671756744385, "dclm_edu": 3.168818950653076, "cosmopedia": 2.2698538303375244, "finemath": 2.4097580909729004}}