tokenizer string | vocab_size int64 | num_tokens int64 | source string | split string | dtype string | gen_params dict |
|---|---|---|---|---|---|---|
nca_patch | 10,002 | 6,545,880,000 | nca | train | uint16 | {
"variant": "share",
"num_rules": 16000,
"num_sims": 16000,
"num_epochs": 200,
"filter_steps": 10,
"grid_size": 12,
"num_colors": 10,
"patch_size": 2,
"seq_len": 2048,
"tokens_per_grid": 38,
"grids_per_trajectory": 54,
"tokens_per_trajectory": 2052,
"start_token": 10000,
"end_token": 10001,... |
No dataset card yet
- Downloads last month
- 53