File size: 276 Bytes
36461cb
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
tokens 900068420
shards 10
batches 13728
window 65536
grad_accum 8
usable_steps 1716
mix b5_65536_131072 target 0.5000 actual 0.5001 tokens 450083582
mix b3_16384_32768 target 0.3000 actual 0.3000 tokens 269995569
mix b1_4096_8192 target 0.2000 actual 0.2000 tokens 179989269