format string | split string | ntokens int64 | ndocs int64 | eot_id int64 | vocab_size int64 | tokenizer string | doc_shuffled bool | shuffle_seed int64 | shuffle_buffer int64 | split_kind string | max_doc_tokens int64 | split_docs int64 | source_tokens dict | source_docs dict | source_weights dict |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
u16le | train | 5,993,742,277 | 3,621,602 | 0 | 50,176 | lwt-shared-bpe-50176.json | true | 42 | 300,000 | random | 16,384 | 34,208 | {
"open-web-math": 5581739826,
"cosmopedia-auto_math_text": 333471358,
"metamath": 23958947,
"mathinstruct": 25175887,
"orca-math": 35654565
} | {
"open-web-math": 2768455,
"cosmopedia-auto_math_text": 511588,
"metamath": 114326,
"mathinstruct": 116460,
"orca-math": 114434
} | {
"open-web-math": 0.78,
"cosmopedia-auto_math_text": 0.14,
"metamath": 0.03,
"mathinstruct": 0.03,
"orca-math": 0.02
} |
No dataset card yet
- Downloads last month
- 18