Jul 19 00:27 Add Focus Crawl and BioCollection counts Add exact Marin-tokenizer counts for the Common Crawl Focus Crawl and both TheBioCollection streams registered in the Marin datakit source catalog. Jul 15 19:56 Add Nemotron Code v2 content token count 19:33 Add exact dolma4pdfs token count 07:14 Add Nemotron specialized v1.2 and legal token counts Jul 14 23:48 Fill missing dataset licenses 23:30 Add NuminaMath-TIR token estimate 22:25 Update SWE-ZERO-12M token estimate 21:59 Add dolma_code, dolma_code_prose, ghalogs/public, massive_function_calling dolma_code (1,221,318,442,892) and dolma_code_prose (65,538,632,427) are exact counts,
measured with marin-community/marin-tokenizer over the normalized data.
ghalogs/public and massive_function_calling carry the registry's estimates.
dolma4pdfs is not included: its normalize is still running. Jul 13 21:07 Add eai-taxonomy-code-w-dclm (591.9B tokens, marin tokenizer) May 19 22:01 Update token counts May 15 19:31 Update token counts 02:30 Update token counts May 1 02:31 Update token counts 02:28 Update token counts Apr 30 06:57 Update token counts 06:51 Update token counts Apr 10 00:13 Update token counts Apr 7 22:18 Update token counts 22:16 Update token counts Apr 5 21:44 Update token counts Apr 2 19:28 Update token counts 17:46 Update token counts Apr 1 08:40 Update token counts Mar 31 17:39 Update token counts 17:32 Update token counts 02:52 Update token counts 00:49 Update token counts 00:37 Update token counts 00:32 Update token counts Mar 30 18:28 Update token counts 18:18 Update token counts 17:43 Update token counts 17:27 Update token counts 
