Update tokenizer, configuration and training reports

reports/run/metrics.jsonl

@@ -997,3 +997,4 @@
 {"step": 19940, "tokens": 10454302720, "loss": 2.7497489929199217, "tokens_per_second": 978481.6580073143, "lr": 0.0015, "gradient_norm": 0.1444910168647766, "benchmark": false}
 {"step": 19960, "tokens": 10464788480, "loss": 2.759548377990723, "tokens_per_second": 985718.732058168, "lr": 0.0015, "gradient_norm": 0.17241807281970978, "benchmark": false}
 {"step": 19980, "tokens": 10475274240, "loss": 2.765959548950195, "tokens_per_second": 984693.2569846577, "lr": 0.0015, "gradient_norm": 0.15740971267223358, "benchmark": false}
+{"step": 20000, "tokens": 10485760000, "loss": 2.7601818084716796, "tokens_per_second": 837521.6981495382, "lr": 0.0015, "gradient_norm": 0.140048548579216, "benchmark": false}

reports/run/validation.jsonl

@@ -16,3 +16,5 @@
 {"step": 16000, "source_validation_loss": {"fineweb_edu": 2.9361798763275146, "dclm_edu": 2.9739556312561035, "cosmopedia": 1.9942559003829956, "finemath": 2.171142578125}}
 {"step": 17000, "source_validation_loss": {"fineweb_edu": 2.931060791015625, "dclm_edu": 2.966684341430664, "cosmopedia": 1.9891211986541748, "finemath": 2.1641457080841064}}
 {"step": 18000, "source_validation_loss": {"fineweb_edu": 2.923886775970459, "dclm_edu": 2.9643118381500244, "cosmopedia": 1.983628511428833, "finemath": 2.1608521938323975}}
+{"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}}