Update tokenizer, configuration and training reports

reports/run/metrics.jsonl

@@ -131,3 +131,5 @@
 {"step": 2620, "tokens": 1373634560, "loss": 3.120655632019043, "tokens_per_second": 988014.3404262743, "lr": 0.0015, "gradient_norm": 0.22039465606212616, "benchmark": false}
 {"step": 2640, "tokens": 1384120320, "loss": 3.119555854797363, "tokens_per_second": 989281.896508611, "lr": 0.0015, "gradient_norm": 0.22109156847000122, "benchmark": false}
 {"step": 2660, "tokens": 1394606080, "loss": 3.1065670013427735, "tokens_per_second": 960148.7382511444, "lr": 0.0015, "gradient_norm": 0.3518049716949463, "benchmark": false}
+{"step": 2680, "tokens": 1405091840, "loss": 3.0817359924316405, "tokens_per_second": 986287.0883485635, "lr": 0.0015, "gradient_norm": 0.23797304928302765, "benchmark": false}
+{"step": 2700, "tokens": 1415577600, "loss": 3.105073928833008, "tokens_per_second": 979995.0214306681, "lr": 0.0015, "gradient_norm": 0.24327975511550903, "benchmark": false}