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 | 5,786,249,803 | 3,787,345 | 0 | 50,176 | lwt-shared-bpe-50176.json | true | 42 | 300,000 | random | 16,384 | 30,628 | {
"gutenberg": 3601146989,
"cosmopedia-stories": 1698301469,
"ultrachat": 492772221
} | {
"gutenberg": 238329,
"cosmopedia-stories": 3089007,
"ultrachat": 463897
} | {
"gutenberg": 0.6,
"cosmopedia-stories": 0.28,
"ultrachat": 0.12
} |
No dataset card yet
- Downloads last month
- 77