| tokens 640009970 | |
| shards 7 | |
| batches 39058 | |
| window 16384 | |
| grad_accum 8 | |
| usable_steps 4882 | |
| mix b3_16384_32768 target 0.5000 actual 0.5000 tokens 320011377 | |
| mix b2_8192_16384 target 0.3000 actual 0.3000 tokens 192001713 | |
| mix b1_4096_8192 target 0.2000 actual 0.2000 tokens 127996880 | |