diff --git "a/attnserver.run_attnserver.slurm.sh.343226.out.log" "b/attnserver.run_attnserver.slurm.sh.343226.out.log" --- "a/attnserver.run_attnserver.slurm.sh.343226.out.log" +++ "b/attnserver.run_attnserver.slurm.sh.343226.out.log" @@ -725,3 +725,2676 @@ batch tensor after cp: labels torch.Size([2, 1024]) batch tensor after cp: loss_mask torch.Size([2, 1024]) batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) batch tensor after cp: position_ids torch.Size([2, 1024]) +Start exporting trace 0 +Done exporting trace 0 + [2025-06-21 21:34:03] iteration 1/ 10 | consumed samples: 1 | elapsed time per iteration (ms): 10014.7 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 4294967296.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +Number of parameters in transformer block in billions: 0.35 +Number of parameters in embedding layers in billions: 0.21 +Total number of parameters in billions: 0.56 +Number of parameters in most loaded shard in billions: 0.1400 +Theoretical memory footprints: weight and optimizer=2403.18 MB +[Rank 2] (after 1 iterations) memory (MB) | allocated: 1812.82470703125 | max allocated: 1812.82568359375 | reserved: 2304.0 | max reserved: 2304.0 +[Rank 1] (after 1 iterations) memory (MB) | allocated: 1812.82470703125 | max allocated: 1812.82568359375 | reserved: 2196.0 | max reserved: 2196.0[Rank 7] (after 1 iterations) memory (MB) | allocated: 1812.82470703125 | max allocated: 1812.82568359375 | reserved: 2304.0 | max reserved: 2304.0[Rank 5] (after 1 iterations) memory (MB) | allocated: 1812.82470703125 | max allocated: 1812.82568359375 | reserved: 2196.0 | max reserved: 2196.0[Rank 4] (after 1 iterations) memory (MB) | allocated: 1812.82470703125 | max allocated: 1812.82568359375 | reserved: 2304.0 | max reserved: 2304.0 + + + +[Rank 0] (after 1 iterations) memory (MB) | allocated: 1812.82470703125 | max allocated: 1812.82568359375 | reserved: 2304.0 | max reserved: 2304.0 +[Rank 6] (after 1 iterations) memory (MB) | allocated: 1812.82470703125 | max allocated: 1812.82568359375 | reserved: 2196.0 | max reserved: 2196.0 +[Rank 3] (after 1 iterations) memory (MB) | allocated: 1812.82470703125 | max allocated: 1812.82568359375 | reserved: 2304.0 | max reserved: 2304.0 +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels batch tensor after cp:torch.Size([2, 2048]) +tokensbatch tensor: loss_masktorch.Size([2, 1024]) +torch.Size([2, 2048]) +batch tensor after cp: batch tensor:labelsbatch tensor after cp: attention_masktorch.Size([2, 1024]) tokens +torch.Size([2, 1, 2048, 2048])batch tensor after cp: + torch.Size([2, 1024])loss_maskbatch tensor: + torch.Size([2, 1024])position_idsbatch tensor after cp: + torch.Size([2, 2048])batch tensor after cp:labels + attention_masktorch.Size([2, 1024]) +torch.Size([2, 1, 1024, 2048])batch tensor after cp: + loss_maskbatch tensor after cp: torch.Size([2, 1024])position_ids + torch.Size([2, 1024])batch tensor after cp:batch tensor: + attention_mask tokenstorch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024])batch tensor: +torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +tokensbatch tensor: batch tensor after cp:loss_mask torch.Size([2, 2048])torch.Size([2, 2048])tokens + + batch tensor: attention_masktorch.Size([2, 1024]) +batch tensor:torch.Size([2, 1, 2048, 2048])batch tensor after cp: + labelslabelsbatch tensor: torch.Size([2, 1024])torch.Size([2, 2048]) +position_ids +batch tensor after cp:batch tensor: loss_maskloss_mask torch.Size([2, 2048]) torch.Size([2, 2048]) +torch.Size([2, 1024]) + +batch tensor:batch tensor after cp: attention_maskattention_mask torch.Size([2, 1, 2048, 2048]) +torch.Size([2, 1, 1024, 2048])batch tensor: + position_idsbatch tensor after cp: torch.Size([2, 2048])position_ids + torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp:batch tensor: tokens torch.Size([2, 1024])tokens + batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp:torch.Size([2, 2048]) loss_mask +torch.Size([2, 1024]) +batch tensor:batch tensor after cp: attention_masklabels torch.Size([2, 1, 1024, 2048])torch.Size([2, 2048]) + +batch tensor after cp:batch tensor: position_idsloss_mask torch.Size([2, 1024])torch.Size([2, 2048]) + +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +Start exporting trace 1 +Done exporting trace 1 + [2025-06-21 21:34:03] iteration 2/ 10 | consumed samples: 2 | elapsed time per iteration (ms): 66.8 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 2147483648.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask batch tensor:torch.Size([2, 1, 2048, 2048]) +batch tensor: tokensposition_ids torch.Size([2, 2048]) +torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_idsbatch tensor: torch.Size([2, 1024]) + tokens batch tensor: torch.Size([2, 2048])tokens + batch tensor: labels torch.Size([2, 2048])torch.Size([2, 2048]) + +batch tensor: batch tensor: loss_masklabels torch.Size([2, 2048])torch.Size([2, 2048]) + +batch tensor:batch tensor: attention_maskloss_mask batch tensor:torch.Size([2, 1, 2048, 2048])torch.Size([2, 2048]) + +batch tensor:tokens batch tensor: position_ids attention_masktorch.Size([2, 2048]) torch.Size([2, 1, 2048, 2048]) + +torch.Size([2, 2048])batch tensor: + position_ids batch tensor:torch.Size([2, 2048]) +labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels batch tensor after cp:torch.Size([2, 1024]) +tokensbatch tensor after cp: torch.Size([2, 1024])loss_mask + batch tensor after cp:torch.Size([2, 1024]) +labelsbatch tensor after cp: torch.Size([2, 1024])attention_mask +batch tensor after cp: batch tensor after cp:torch.Size([2, 1, 1024, 2048]) +tokensloss_maskbatch tensor after cp: torch.Size([2, 1024])torch.Size([2, 1024])position_ids + +batch tensor after cp: torch.Size([2, 1024])batch tensor after cp: + labelsattention_mask torch.Size([2, 1024])torch.Size([2, 1, 1024, 2048]) + +batch tensor after cp:batch tensor after cp: position_idsloss_mask torch.Size([2, 1024])torch.Size([2, 1024]) + +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +Start exporting trace 2 +Done exporting trace 2 + [2025-06-21 21:34:03] iteration 3/ 10 | consumed samples: 3 | elapsed time per iteration (ms): 35.7 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 1073741824.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labelsbatch tensor: torch.Size([2, 2048]) + batch tensor:tokens loss_mask torch.Size([2, 2048]) +batch tensor: attention_masktorch.Size([2, 2048]) +torch.Size([2, 1, 2048, 2048]) +batch tensor:batch tensor: labelsposition_ids torch.Size([2, 2048])torch.Size([2, 2048]) + +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor:batch tensor after cp: tokens tokens torch.Size([2, 1024]) +batch tensor after cp: labelstorch.Size([2, 2048]) +torch.Size([2, 1024]) +batch tensor:batch tensor after cp: labelsloss_mask torch.Size([2, 2048])torch.Size([2, 1024]) + +batch tensor:batch tensor after cp: loss_maskattention_mask torch.Size([2, 2048])torch.Size([2, 1, 1024, 2048]) + +batch tensor:batch tensor after cp: attention_maskposition_ids torch.Size([2, 1, 2048, 2048])torch.Size([2, 1024]) + +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: batch tensor after cp:loss_mask tokenstorch.Size([2, 1024]) +batch tensor after cp:torch.Size([2, 1024]) +attention_mask batch tensor after cp:torch.Size([2, 1, 1024, 2048]) +labelsbatch tensor after cp: torch.Size([2, 1024])position_ids + batch tensor after cp:torch.Size([2, 1024]) +loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: batch tensor:tokens torch.Size([2, 1024]) +tokensbatch tensor after cp: labels torch.Size([2, 1024])torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) + +batch tensor after cp: attention_maskbatch tensor: torch.Size([2, 1, 1024, 2048])labels + batch tensor after cp:torch.Size([2, 2048]) +position_idsbatch tensor: torch.Size([2, 1024])loss_mask + torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +Start exporting trace 3 +Done exporting trace 3 + [2025-06-21 21:34:03] iteration 4/ 10 | consumed samples: 4 | elapsed time per iteration (ms): 35.3 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 536870912.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: batch tensor:loss_mask torch.Size([2, 2048]) +tokensbatch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +torch.Size([2, 2048])batch tensor: +position_ids torch.Size([2, 2048])batch tensor: + labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp:batch tensor after cp: batch tensor:tokenstokens torch.Size([2, 1024])tokenstorch.Size([2, 1024]) + +batch tensor after cp: batch tensor after cp:labels labelstorch.Size([2, 2048])torch.Size([2, 1024]) +torch.Size([2, 1024]) + +batch tensor after cp:batch tensor: batch tensor after cp:loss_mask labelstorch.Size([2, 1024])loss_mask + torch.Size([2, 2048])batch tensor after cp:torch.Size([2, 1024]) + +attention_maskbatch tensor:batch tensor after cp: torch.Size([2, 1, 1024, 2048])loss_maskattention_mask + batch tensor after cp:torch.Size([2, 1, 1024, 2048]) +position_idstorch.Size([2, 2048])batch tensor after cp: + torch.Size([2, 1024])position_ids +batch tensor: attention_masktorch.Size([2, 1024]) + torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +Start exporting trace 4 +Done exporting trace 4 + [2025-06-21 21:34:03] iteration 5/ 10 | consumed samples: 5 | elapsed time per iteration (ms): 32.3 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 268435456.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor:batch tensor: tokens torch.Size([2, 2048])tokens + batch tensor: labels torch.Size([2, 2048])torch.Size([2, 2048]) + +batch tensor: loss_mask torch.Size([2, 2048])batch tensor: + batch tensor:labels attention_masktorch.Size([2, 2048]) +torch.Size([2, 1, 2048, 2048])batch tensor: +batch tensor: position_ids torch.Size([2, 2048])loss_mask + torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +Start exporting trace 5 +Done exporting trace 5 + [2025-06-21 21:34:03] iteration 6/ 10 | consumed samples: 6 | elapsed time per iteration (ms): 32.2 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 134217728.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 2048]) +batch tensor:batch tensor: labels torch.Size([2, 2048])tokens +batch tensor:batch tensor: loss_masktokens torch.Size([2, 2048]) torch.Size([2, 2048]) + +batch tensor:batch tensor: torch.Size([2, 2048]) attention_masklabels + torch.Size([2, 2048])batch tensor:torch.Size([2, 1, 2048, 2048]) + +batch tensor:labelsbatch tensor: loss_maskposition_idstorch.Size([2, 2048]) +torch.Size([2, 2048])torch.Size([2, 2048]) + +batch tensor:batch tensor: loss_maskattention_mask torch.Size([2, 2048])torch.Size([2, 1, 2048, 2048]) + +batch tensor: batch tensor:attention_mask position_idstorch.Size([2, 1, 2048, 2048]) +torch.Size([2, 2048])batch tensor: + position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor:batch tensor: tokens tokens torch.Size([2, 2048]) +torch.Size([2, 2048])batch tensor: + labelsbatch tensor: torch.Size([2, 2048])labels + batch tensor:torch.Size([2, 2048]) +loss_mask batch tensor:torch.Size([2, 2048]) +loss_mask batch tensor:torch.Size([2, 2048]) +attention_mask batch tensor:torch.Size([2, 1, 2048, 2048]) +attention_mask batch tensor: torch.Size([2, 1, 2048, 2048])position_ids + batch tensor:torch.Size([2, 2048]) + position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024])batch tensor after cp: + batch tensor after cp:tokensbatch tensor after cp: tokensattention_mask torch.Size([2, 1024]) +torch.Size([2, 1024])torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: + batch tensor after cp:labelsbatch tensor after cp: labelstorch.Size([2, 1024])position_ids + batch tensor after cp:torch.Size([2, 1024])torch.Size([2, 1024]) +loss_mask +batch tensor after cp: torch.Size([2, 1024])loss_mask + torch.Size([2, 1024])batch tensor after cp: + attention_maskbatch tensor after cp: torch.Size([2, 1, 1024, 2048])attention_mask + batch tensor after cp:torch.Size([2, 1, 1024, 2048]) +position_idsbatch tensor after cp: torch.Size([2, 1024])position_ids + torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp:batch tensor after cp: tokenstokens torch.Size([2, 1024])torch.Size([2, 1024]) + +batch tensor after cp:batch tensor after cp: labels labelstorch.Size([2, 1024]) +torch.Size([2, 1024])batch tensor after cp: + loss_maskbatch tensor after cp: torch.Size([2, 1024])loss_mask + batch tensor after cp:torch.Size([2, 1024]) +attention_maskbatch tensor after cp: attention_masktorch.Size([2, 1, 1024, 2048]) +torch.Size([2, 1, 1024, 2048])batch tensor after cp: + position_idsbatch tensor after cp: torch.Size([2, 1024])position_ids + torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +Start exporting trace 6 +Done exporting trace 6 + [2025-06-21 21:34:03] iteration 7/ 10 | consumed samples: 7 | elapsed time per iteration (ms): 37.8 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 67108864.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp:batch tensor: position_ids torch.Size([2, 1024])tokens + torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor:batch tensor: position_ids torch.Size([2, 2048]) +tokens batch tensor after cp: tokens torch.Size([2, 2048])torch.Size([2, 1024]) + +batch tensor after cp:batch tensor: labelslabels torch.Size([2, 2048])torch.Size([2, 1024]) + +batch tensor:batch tensor after cp:batch tensor: loss_maskloss_mask tokens torch.Size([2, 2048])torch.Size([2, 1024]) + +torch.Size([2, 2048])batch tensor:batch tensor after cp: + attention_maskattention_maskbatch tensor: torch.Size([2, 1, 2048, 2048])torch.Size([2, 1, 1024, 2048])labels + + batch tensor:batch tensor:torch.Size([2, 2048])batch tensor after cp: position_ids +batch tensor: position_idsloss_mask tokenstorch.Size([2, 2048]) +batch tensor after cp: torch.Size([2, 1024]) + torch.Size([2, 2048])tokens + torch.Size([2, 2048])batch tensor: torch.Size([2, 1024]) +attention_mask + batch tensor:batch tensor after cp:torch.Size([2, 1, 2048, 2048]) +batch tensor:labelslabels position_idstorch.Size([2, 2048])torch.Size([2, 1024]) + +torch.Size([2, 2048])batch tensor after cp: +batch tensor: loss_maskloss_mask torch.Size([2, 1024])torch.Size([2, 2048]) + +batch tensor after cp:batch tensor: attention_maskattention_mask torch.Size([2, 1, 1024, 2048])torch.Size([2, 1, 2048, 2048]) + +batch tensor after cp: batch tensor:position_ids position_idstorch.Size([2, 1024]) +torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp:batch tensor: tokens tokenstorch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 2048])torch.Size([2, 1024]) + +batch tensor after cp: batch tensor:loss_mask torch.Size([2, 1024])labels + batch tensor after cp:torch.Size([2, 2048]) +attention_maskbatch tensor: torch.Size([2, 1, 1024, 2048])loss_mask +torch.Size([2, 2048]) +batch tensor after cp: batch tensor:position_ids attention_masktorch.Size([2, 1024]) +torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +Start exporting trace 7 +Done exporting trace 7 + [2025-06-21 21:34:03] iteration 8/ 10 | consumed samples: 8 | elapsed time per iteration (ms): 32.3 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 33554432.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: batch tensor:tokens tokenstorch.Size([2, 2048]) +batch tensor:torch.Size([2, 2048]) +labels batch tensor:torch.Size([2, 2048]) +labelsbatch tensor: torch.Size([2, 2048])loss_mask + batch tensor:torch.Size([2, 2048]) loss_mask + torch.Size([2, 2048])batch tensor: + attention_maskbatch tensor: attention_masktorch.Size([2, 1, 2048, 2048]) +torch.Size([2, 1, 2048, 2048])batch tensor: + batch tensor:position_ids position_idstorch.Size([2, 2048]) +torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask batch tensor:torch.Size([2, 2048]) +batch tensor: tokensattention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048])torch.Size([2, 2048]) + +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask batch tensor after cp:torch.Size([2, 2048]) + tokensbatch tensor: torch.Size([2, 1024]) +attention_mask batch tensor after cp: torch.Size([2, 1, 2048, 2048])labels + batch tensor:torch.Size([2, 1024]) +position_idsbatch tensor after cp: torch.Size([2, 2048])loss_mask + torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens batch tensor after cp:torch.Size([2, 1024]) +tokensbatch tensor after cp: labelstorch.Size([2, 1024]) +torch.Size([2, 1024]) +batch tensor after cp:batch tensor after cp: labelsloss_mask torch.Size([2, 1024])torch.Size([2, 1024]) + +batch tensor after cp: batch tensor after cp:loss_mask batch tensor:attention_masktorch.Size([2, 1024])batch tensor after cp: + tokenstorch.Size([2, 1, 1024, 2048])tokens batch tensor after cp: +torch.Size([2, 1024]) batch tensor after cp:torch.Size([2, 2048])attention_mask + + batch tensor after cp:batch tensor:torch.Size([2, 1, 1024, 2048])position_ids + labelslabelsbatch tensor after cp:torch.Size([2, 1024]) +torch.Size([2, 1024])torch.Size([2, 2048])position_ids + + batch tensor after cp:batch tensor:torch.Size([2, 1024]) +loss_maskloss_mask torch.Size([2, 1024])torch.Size([2, 2048]) + +batch tensor after cp:batch tensor after cp:batch tensor: attention_mask tokensattention_mask torch.Size([2, 1, 1024, 2048])torch.Size([2, 1024]) +torch.Size([2, 1, 2048, 2048]) +batch tensor after cp: +batch tensor after cp: batch tensor: position_idslabels position_idstorch.Size([2, 1024]) torch.Size([2, 1024]) + +torch.Size([2, 2048])batch tensor after cp: + loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +Start exporting trace 8 +Done exporting trace 8 + [2025-06-21 21:34:03] iteration 9/ 10 | consumed samples: 9 | elapsed time per iteration (ms): 33.3 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 16777216.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_maskbatch tensor after cp: torch.Size([2, 1, 2048, 2048])tokens + batch tensor: torch.Size([2, 1024])position_ids + batch tensor after cp:torch.Size([2, 2048]) +labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048])batch tensor: +batch tensor after cp: position_ids tokens torch.Size([2, 1024]) +torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +Start exporting trace 9 +Done exporting trace 9 + [2025-06-21 21:34:03] iteration 10/ 10 | consumed samples: 10 | elapsed time per iteration (ms): 34.2 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 8388608.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +[after training is done] datetime: 2025-06-21 21:34:03 +saving checkpoint at iteration 10 to gpt-checkpoint in torch_dist format +DEBUG:megatron.training.checkpointing:rank: 7, takes 0.03700566291809082 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 2, takes 0.037008047103881836 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 6, takes 0.03712654113769531 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 3, takes 0.037163496017456055 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 5, takes 0.037125587463378906 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 1, takes 0.0371546745300293 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 4, takes 0.0375676155090332 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 0, takes 0.2952592372894287 to prepare state dict for ckpt +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(417400832), 0), (np.int64(422576128), 1)] +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(417400832), 0), (np.int64(422576128), 1)] +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(417400832), 0), (np.int64(422576128), 1)] +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(417400832), 0), (np.int64(422576128), 1)] +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(417400832), 0), (np.int64(422576128), 1)] +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(417400832), 0), (np.int64(422576128), 1)] +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(430327808), 0), (np.int64(435159040), 1)] +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(430327808), 0), (np.int64(435159040), 1)] +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:parallel save sharding, time: 1.0959203243255615 +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:parallel save sharding, time: 1.0958845615386963 +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:parallel save sharding, time: 1.0959882736206055 +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:parallel save sharding, time: 1.0959982872009277 +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:parallel save sharding, time: 1.0959341526031494 +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:parallel save sharding, time: 1.0960769653320312 +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:parallel save sharding, time: 1.0970747470855713 +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:parallel save sharding, time: 0.008258819580078125 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 7, starting state dict save +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 6, starting state dict save +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 4, starting state dict save +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 5, starting state dict save +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:verifying reuse of global metadata +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:loaded global metadata reuse verification: no loaded plans passed +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:verifying reuse of global metadata +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:loaded global metadata reuse verification: no loaded plans passed +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:verifying reuse of global metadata +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:loaded global metadata reuse verification: no loaded plans passed +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:verifying reuse of global metadata +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:loaded global metadata reuse verification: no loaded plans passed +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 2, starting state dict save +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 1, starting state dict save +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:verifying reuse of global metadata +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:loaded global metadata reuse verification: no loaded plans passed +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:verifying reuse of global metadata +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:loaded global metadata reuse verification: no loaded plans passed +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 0, starting state dict save +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 3, starting state dict save +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:verifying reuse of global metadata +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:loaded global metadata reuse verification: no loaded plans passed +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:verifying reuse of global metadata +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:loaded global metadata reuse verification: no loaded plans passed +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 5, plan time: 0.00821685791015625 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 7, plan time: 0.008537769317626953 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 1, plan time: 0.0032274723052978516 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 4, plan time: 0.008394241333007812 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:thread_count: 2, time: 1750541646.0637105 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:thread_count: 2, time: 1750541646.063714 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:thread_count: 2, time: 1750541646.063713 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:thread_count: 2, time: 1750541646.063717 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 3, plan time: 0.0021944046020507812 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 6, plan time: 0.008464574813842773 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:bucket_prep, time: 6.723403930664062e-05 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:thread_count: 2, time: 1750541646.0637758 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 2, plan time: 0.0034880638122558594 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:bucket_prep, time: 7.653236389160156e-05 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:bucket_prep, time: 8.368492126464844e-05 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:bucket_prep, time: 8.845329284667969e-05 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:thread_count: 2, time: 1750541646.0638037 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:thread_count: 2, time: 1750541646.0638173 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:bucket_prep, time: 0.00010132789611816406 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:bucket_prep, time: 8.225440979003906e-05 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:bucket_prep, time: 7.653236389160156e-05 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:rank: 0, plan time: 0.004988431930541992 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:thread_count: 2, time: 1750541646.0664604 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:bucket_prep, time: 0.00010251998901367188 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:D2H and push, time: 0.049828529357910156 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541646.114014 rank: 1, write(async) time: 0.05029797554016113 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:D2H and push, time: 0.05093526840209961 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541646.1151786 rank: 2, write(async) time: 0.05135941505432129 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:D2H and push, time: 0.05128598213195801 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541646.115486 rank: 3, write(async) time: 0.0517117977142334 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:D2H and push, time: 0.05167031288146973 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:D2H and push, time: 0.051851511001586914 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541646.1158936 rank: 6, write(async) time: 0.05208778381347656 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541646.1159801 rank: 7, write(async) time: 0.05226445198059082 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:D2H and push, time: 0.05207991600036621 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541646.116206 rank: 4, write(async) time: 0.05248737335205078 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:D2H and push, time: 0.05397963523864746 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541646.1182792 rank: 5, write(async) time: 0.05456900596618652 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:D2H and push, time: 0.053023338317871094 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541646.1200068 rank: 0, write(async) time: 0.0535433292388916 +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 6, takes 1.5974044799804688e-05 to finish D2H +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 7, takes 1.5974044799804688e-05 to finish D2H +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 4, takes 1.52587890625e-05 to finish D2H +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 2, takes 1.5735626220703125e-05 to finish D2H +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 3, takes 1.6689300537109375e-05 to finish D2H +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 6, takes 0.025777816772460938 to schedule async ckpt +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 7, takes 0.026104450225830078 to schedule async ckpt +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 4, takes 0.0262298583984375 to schedule async ckpt +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 2, takes 0.027044057846069336 to schedule async ckpt +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 3, takes 0.025261640548706055 to schedule async ckpt +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 5, takes 1.7881393432617188e-05 to finish D2H +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 started +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 started +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 1, takes 1.5497207641601562e-05 to finish D2H +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 started +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 started +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 5, takes 0.02705526351928711 to schedule async ckpt +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collecting worker results... +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 started +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collecting worker results... +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 started +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 started +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 1, takes 0.022681474685668945 to schedule async ckpt +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collecting worker results... +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 started +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collecting worker results... +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 started +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 started +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collecting worker results... +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 started +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 started +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collecting worker results... +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 started +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collecting worker results... +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 started +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 0, takes 2.1696090698242188e-05 to finish D2H +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 0, takes 0.034197092056274414 to schedule async ckpt +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 1, joining self.process +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 3, joining self.process +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 4, joining self.process +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 2, joining self.process +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 5, joining self.process +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 7, joining self.process +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 6, joining self.process +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:rank: 0, joining self.process +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 started +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collecting worker results... +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 started +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 consumed: 214769664, before: 1595891712, after: 1810661376 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 consumed: 217985024, before: 1595891712, after: 1813876736 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 consumed: 214929408, before: 1592475648, after: 1807405056 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 consumed: 217501696, before: 1592475648, after: 1809977344 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collected worker results successfully +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1750541646.9358253, rank: 6, write(sync,parallel): 0.5805976390838623 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 consumed: 226455552, before: 1594376192, after: 1820831744 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 consumed: 212402176, before: 1619238912, after: 1831641088 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 consumed: 212447232, before: 1593528320, after: 1805975552 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 consumed: 218718208, before: 1594376192, after: 1813094400 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 consumed: 214364160, before: 1619238912, after: 1833603072 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 consumed: 217817088, before: 1602723840, after: 1820540928 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 consumed: 214343680, before: 1593528320, after: 1807872000 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 consumed: 214847488, before: 1602658304, after: 1817505792 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 consumed: 212434944, before: 1586868224, after: 1799303168 +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:TemporalAsyncCaller: Async process join finished after 0.64s from forking +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collected worker results successfully +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collected worker results successfully +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collected worker results successfully +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1750541646.982197, rank: 4, write(sync,parallel): 0.6203939914703369 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1750541646.9834635, rank: 2, write(sync,parallel): 0.6179161071777344 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1750541646.986327, rank: 7, write(sync,parallel): 0.630373477935791 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 consumed: 214335488, before: 1586868224, after: 1801203712 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collected worker results successfully +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1750541646.9998002, rank: 1, write(sync,parallel): 0.5877251625061035 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collected worker results successfully +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1750541647.0139754, rank: 5, write(sync,parallel): 0.6128330230712891 +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:TemporalAsyncCaller: Async process join finished after 0.68s from forking +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:TemporalAsyncCaller: Async process join finished after 0.68s from forking +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:TemporalAsyncCaller: Async process join finished after 0.69s from forking +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:TemporalAsyncCaller: Async process join finished after 0.64s from forking +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collected worker results successfully +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1750541647.0323749, rank: 3, write(sync,parallel): 0.6617300510406494 +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:TemporalAsyncCaller: Async process join finished after 0.67s from forking +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:TemporalAsyncCaller: Async process join finished after 0.73s from forking +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:0 consumed: 220749824, before: 1879838720, after: 2100588544 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1 consumed: 219033600, before: 1879838720, after: 2098872320 +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:FileSystemWriterAsync: collected worker results successfully +DEBUG:megatron.core.dist_checkpointing.strategies.filesystem_async:1750541647.1939285, rank: 0, write(sync,parallel): 0.5937509536743164 +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:TemporalAsyncCaller: Async process join finished after 0.67s from forking +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541647.2351289, 1, gather: 0.2049856185913086 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541647.2351787, 3, gather: 0.1633777618408203 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541647.2352345, 5, gather: 0.19353938102722168 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541647.2353487, 6, gather: 0.26531386375427246 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541647.2353427, 4, gather: 0.21900391578674316 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541647.2353475, 2, gather: 0.21796727180480957 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541647.2354062, 7, gather: 0.21528196334838867 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541647.2366471, 0, gather: 0.0037147998809814453 +DEBUG:megatron.core.dist_checkpointing.strategies.state_dict_saver:1750541647.2449548, metadata_write: 0.008182287216186523 +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:finalize took 0.1754s +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:finalize took 0.2767s +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:finalize took 0.2305s +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:finalize took 0.2051s +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:finalize took 0.0143s +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:finalize took 0.2269s +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:finalize took 0.2296s +DEBUG:megatron.core.dist_checkpointing.strategies.async_utils:finalize took 0.2166s + successfully saved checkpoint from iteration 10 to gpt-checkpoint [ t 1/4, p 1/1 ] +DEBUG:megatron.training.checkpointing:rank: 4, takes 0.002124309539794922 to finalize ckpt save +DEBUG:megatron.training.checkpointing:rank: 3, takes 0.00205230712890625 to finalize ckpt save +DEBUG:megatron.training.checkpointing:rank: 7, takes 0.0021219253540039062 to finalize ckpt save +DEBUG:megatron.training.checkpointing:rank: 6, takes 0.0021369457244873047 to finalize ckpt save +DEBUG:megatron.training.checkpointing:rank: 2, takes 0.002108335494995117 to finalize ckpt save +DEBUG:megatron.training.checkpointing:rank: 1, takes 0.0020492076873779297 to finalize ckpt save +DEBUG:megatron.training.checkpointing:rank: 5, takes 0.0021142959594726562 to finalize ckpt save +DEBUG:megatron.training.checkpointing:rank: 0, takes 0.0020835399627685547 to finalize ckpt save +WARNING:megatron.core.rerun_state_machine:Setting RerunStateMachine mode RerunMode.DISABLED +Evaluating on 1 samples +Evaluating iter 1/1 +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: batch tensor:loss_mask torch.Size([2, 2048]) + tokens batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor:torch.Size([2, 2048]) +position_ids torch.Size([2, 2048])batch tensor: + labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor:batch tensor: attention_mask tokenstorch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_idsbatch tensor after cp: torch.Size([2, 1024])tokens + torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: batch tensor:position_ids torch.Size([2, 1024]) + tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048])batch tensor after cp: + batch tensor:tokens position_ids torch.Size([2, 2048])torch.Size([2, 1024]) + +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp:batch tensor after cp: labelstokens torch.Size([2, 1024])torch.Size([2, 1024]) + +batch tensor after cp:batch tensor after cp: loss_masklabels torch.Size([2, 1024])torch.Size([2, 1024]) + +batch tensor after cp:batch tensor after cp: attention_maskloss_mask torch.Size([2, 1, 1024, 2048])torch.Size([2, 1024]) + +batch tensor after cp:batch tensor after cp: position_idsattention_mask torch.Size([2, 1024]) +torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +Start exporting trace 10 +Done exporting trace 10 +(min, max) time across ranks (ms): + evaluate .......................................: (2421.47, 2422.59) +---------------------------------------------------------------------------------------------------------------- + validation loss at iteration 10 on validation set | lm loss value: 1.276627E+01 | lm loss PPL: 3.502022E+05 | +---------------------------------------------------------------------------------------------------------------- +WARNING:megatron.core.rerun_state_machine:Setting RerunStateMachine mode RerunMode.DISABLED +WARNING:megatron.core.rerun_state_machine:Setting RerunStateMachine mode RerunMode.DISABLED +WARNING:megatron.core.rerun_state_machine:Setting RerunStateMachine mode RerunMode.DISABLED +Evaluating on 1 samples +Evaluating iter 1/1 +batch tensor: tokens torch.Size([2, 2048]) +batch tensor:batch tensor: labels torch.Size([2, 2048])tokens + batch tensor: loss_mask torch.Size([2, 2048]) +torch.Size([2, 2048])batch tensor: + attention_mask batch tensor: torch.Size([2, 1, 2048, 2048])labels + batch tensor:torch.Size([2, 2048]) +position_ids torch.Size([2, 2048])batch tensor: + loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor:batch tensor: loss_mask torch.Size([2, 2048]) +tokens batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +torch.Size([2, 2048])batch tensor: +position_ids torch.Size([2, 2048])batch tensor: + labels torch.Size([2, 2048]) +batch tensor: loss_mask batch tensor after cp:torch.Size([2, 2048]) + tokens batch tensor: torch.Size([2, 1024])attention_maskbatch tensor: batch tensor after cp: +batch tensor:torch.Size([2, 1, 2048, 2048]) +batch tensor after cp:tokens batch tensor: tokenstokens labels torch.Size([2, 1024])torch.Size([2, 1024]) +torch.Size([2, 2048])position_ids + +batch tensor after cp: torch.Size([2, 2048])batch tensor:batch tensor after cp: torch.Size([2, 2048]) + + labelsloss_masklabels batch tensor:torch.Size([2, 1024]) torch.Size([2, 1024]) + +torch.Size([2, 2048])labelsbatch tensor after cp:batch tensor after cp: + torch.Size([2, 2048]) loss_mask +batch tensor:attention_mask torch.Size([2, 1024])batch tensor:loss_masktorch.Size([2, 1, 1024, 2048]) + + loss_maskbatch tensor after cp:torch.Size([2, 2048])batch tensor after cp: +position_idstorch.Size([2, 2048])attention_maskbatch tensor: + torch.Size([2, 1024])attention_maskbatch tensor: torch.Size([2, 1, 1024, 2048]) torch.Size([2, 1, 2048, 2048]) + +attention_mask +batch tensor after cp: batch tensor:torch.Size([2, 1, 2048, 2048])position_ids + position_idstorch.Size([2, 1024])batch tensor: + position_idstorch.Size([2, 2048]) +torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask batch tensor after cp:torch.Size([2, 1024]) +tokensbatch tensor after cp: attention_masktorch.Size([2, 1024]) +torch.Size([2, 1, 1024, 2048])batch tensor after cp: + labelsbatch tensor after cp: torch.Size([2, 1024])position_ids + batch tensor after cp:torch.Size([2, 1024]) +loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask batch tensor:torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: tokensposition_ids torch.Size([2, 1024]) +torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 2048]) +batch tensor: loss_mask torch.Size([2, 2048]) +batch tensor: attention_mask torch.Size([2, 1, 2048, 2048]) +batch tensor: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +batch tensor after cp: tokens torch.Size([2, 1024]) +batch tensor after cp: labels torch.Size([2, 1024]) +batch tensor after cp: loss_mask torch.Size([2, 1024]) +batch tensor after cp: attention_mask torch.Size([2, 1, 1024, 2048]) +batch tensor after cp: position_ids torch.Size([2, 1024]) +Start exporting trace 11 +Done exporting trace 11 +(min, max) time across ranks (ms): + evaluate .......................................: (16.83, 22.94) +---------------------------------------------------------------------------------------------------------- + validation loss at iteration 10 on test set | lm loss value: 1.276627E+01 | lm loss PPL: 3.502022E+05 | +---------------------------------------------------------------------------------------------------------- +WARNING:megatron.core.rerun_state_machine:Setting RerunStateMachine mode RerunMode.DISABLED +WARNING:megatron.core.rerun_state_machine:Setting RerunStateMachine mode RerunMode.DISABLED +Running ctx_length=2048, TP_SIZE=4, CP_SIZE=2, BATCH_SIZE=2 +Cleaning up checkpoint directory: gpt-checkpoint +-------------------------------- +CTX_LENGTH: 2048 +TP_SIZE: 4 +CP_SIZE: 2 +CHECKPOINT_PATH: gpt-checkpoint +PWD: /mnt/weka/home/hao.zhang/junda/attnserver-megatron +-------------------------------- +/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/python3 +using world size: 8, data-parallel size: 1, context-parallel size: 2, hierarchical context-parallel sizes: Nonetensor-model-parallel size: 4, encoder-tensor-model-parallel size: 0, pipeline-model-parallel size: 1, encoder-pipeline-model-parallel size: 0 +Number of virtual stages per pipeline stage: None +WARNING: Setting args.check_for_nan_in_loss_and_grad to False since dynamic loss scaling is being used +using torch.float16 for parameters ... +------------------------ arguments ------------------------ + account_for_embedding_in_pipeline_split ......... False + account_for_loss_in_pipeline_split .............. False + accumulate_allreduce_grads_in_fp32 .............. False + adam_beta1 ...................................... 0.9 + adam_beta2 ...................................... 0.999 + adam_eps ........................................ 1e-08 + add_bias_linear ................................. True + add_position_embedding .......................... True + add_qkv_bias .................................... True + adlr_autoresume ................................. False + adlr_autoresume_interval ........................ 1000 + align_grad_reduce ............................... True + align_param_gather .............................. False + app_tag_run_name ................................ None + app_tag_run_version ............................. 0.0.0 + apply_layernorm_1p .............................. False + apply_query_key_layer_scaling ................... False + apply_residual_connection_post_layernorm ........ False + apply_rope_fusion ............................... False + async_save ...................................... None + async_tensor_model_parallel_allreduce ........... True + attention_backend ............................... AttnBackend.auto + attention_dropout ............................... 0.1 + attention_softmax_in_fp32 ....................... False + auto_detect_ckpt_format ......................... False + barrier_with_L1_time ............................ True + bert_binary_head ................................ True + bert_embedder_type .............................. megatron + bert_load ....................................... None + bf16 ............................................ False + bias_dropout_fusion ............................. True + bias_gelu_fusion ................................ True + bias_swiglu_fusion .............................. True + biencoder_projection_dim ........................ 0 + biencoder_shared_query_context_model ............ False + block_data_path ................................. None + calc_ft_timeouts ................................ False + calculate_per_token_loss ........................ False + check_for_large_grads ........................... False + check_for_nan_in_loss_and_grad .................. False + check_for_spiky_loss ............................ False + check_weight_hash_across_dp_replicas_interval ... None + ckpt_assume_constant_structure .................. False + ckpt_convert_format ............................. None + ckpt_convert_save ............................... None + ckpt_convert_update_legacy_dist_opt_format ...... False + ckpt_format ..................................... torch_dist + ckpt_fully_parallel_load ........................ False + ckpt_fully_parallel_save ........................ True + ckpt_fully_parallel_save_deprecated ............. False + ckpt_step ....................................... None + classes_fraction ................................ 1.0 + clip_grad ....................................... 1.0 + clone_scatter_output_in_embedding ............... True + config_logger_dir ............................... + consumed_train_samples .......................... 0 + consumed_valid_samples .......................... 0 + context_parallel_size ........................... 2 + cp_comm_type .................................... ['p2p'] + create_attention_mask_in_dataloader ............. True + cross_entropy_fusion_impl ....................... native + cross_entropy_loss_fusion ....................... False + cuda_graph_scope ................................ full + cuda_graph_warmup_steps ......................... 3 + data_args_path .................................. None + data_cache_path ................................. None + data_parallel_random_init ....................... False + data_parallel_sharding_strategy ................. no_shard + data_parallel_size .............................. 1 + data_path ....................................... None + data_per_class_fraction ......................... 1.0 + data_sharding ................................... True + dataloader_type ................................. single + ddp_average_in_collective ....................... False + ddp_bucket_size ................................. None + ddp_num_buckets ................................. None + ddp_pad_buckets_for_high_nccl_busbw ............. False + decoder_first_pipeline_num_layers ............... None + decoder_last_pipeline_num_layers ................ None + decoder_num_layers .............................. None + decoder_seq_length .............................. None + decoupled_lr .................................... None + decoupled_min_lr ................................ None + decrease_batch_size_if_needed ................... False + defer_embedding_wgrad_compute ................... False + deprecated_use_mcore_models ..................... False + deterministic_mode .............................. False + dino_bottleneck_size ............................ 256 + dino_freeze_last_layer .......................... 1 + dino_head_hidden_size ........................... 2048 + dino_local_crops_number ......................... 10 + dino_local_img_size ............................. 96 + dino_norm_last_layer ............................ False + dino_teacher_temp ............................... 0.07 + dino_warmup_teacher_temp ........................ 0.04 + dino_warmup_teacher_temp_epochs ................. 30 + disable_bf16_reduced_precision_matmul ........... False + disable_mamba_mem_eff_path ...................... False + disable_straggler_on_startup .................... False + dist_ckpt_format_deprecated ..................... None + dist_ckpt_strictness ............................ assume_ok_unexpected + distribute_saved_activations .................... False + distributed_backend ............................. nccl + distributed_timeout_minutes ..................... 10 + embedding_path .................................. None + empty_unused_memory_level ....................... 0 + enable_cuda_graph ............................... False + enable_ft_package ............................... False + enable_gloo_process_groups ...................... True + enable_msc ...................................... True + enable_one_logger ............................... True + encoder_num_layers .............................. 2 + encoder_pipeline_model_parallel_size ............ 0 + encoder_seq_length .............................. 2048 + encoder_tensor_model_parallel_size .............. 0 + end_weight_decay ................................ 0.1 + eod_mask_loss ................................... False + error_injection_rate ............................ 0 + error_injection_type ............................ transient_error + eval_interval ................................... 16 + eval_iters ...................................... 1 + evidence_data_path .............................. None + exit_duration_in_mins ........................... None + exit_interval ................................... None + exit_on_missing_checkpoint ...................... False + exit_signal_handler ............................. False + exp_avg_dtype ................................... torch.float32 + exp_avg_sq_dtype ................................ torch.float32 + expert_model_parallel_size ...................... 1 + expert_tensor_parallel_size ..................... 4 + external_cuda_graph ............................. False + ffn_hidden_size ................................. 16384 + finetune ........................................ False + first_last_layers_bf16 .......................... False + flash_decode .................................... False + fp16 ............................................ True + fp16_lm_cross_entropy ........................... False + fp32_residual_connection ........................ False + fp8 ............................................. None + fp8_amax_compute_algo ........................... most_recent + fp8_amax_history_len ............................ 1 + fp8_interval .................................... 1 + fp8_margin ...................................... 0 + fp8_param_gather ................................ False + fp8_recipe ...................................... delayed + fp8_wgrad ....................................... True + fsdp_double_buffer .............................. False + global_batch_size ............................... 1 + grad_reduce_in_bf16 ............................. False + gradient_accumulation_fusion .................... True + gradient_reduce_div_fusion ...................... True + group_query_attention ........................... True + head_lr_mult .................................... 1.0 + heterogeneous_layers_config_encoded_json ........ None + heterogeneous_layers_config_path ................ None + hidden_dropout .................................. 0.1 + hidden_size ..................................... 4096 + hierarchical_context_parallel_sizes ............. None + high_priority_stream_groups ..................... [] + hybrid_attention_ratio .......................... 0.0 + hybrid_mlp_ratio ................................ 0.0 + hybrid_override_pattern ......................... None + hysteresis ...................................... 2 + ict_head_size ................................... None + ict_load ........................................ None + img_h ........................................... 224 + img_w ........................................... 224 + indexer_batch_size .............................. 128 + indexer_log_interval ............................ 1000 + inference_batch_times_seqlen_threshold .......... -1 + inference_dynamic_batching ...................... False + inference_dynamic_batching_buffer_guaranteed_fraction 0.2 + inference_dynamic_batching_buffer_overflow_factor None + inference_dynamic_batching_buffer_size_gb ....... 40.0 + inference_dynamic_batching_chunk_size ........... 256 + inference_dynamic_batching_max_requests_override None + inference_dynamic_batching_max_tokens_override .. None + inference_max_batch_size ........................ 8 + inference_max_seq_length ........................ 2560 + inference_rng_tracker ........................... False + init_method_std ................................. 0.02 + init_method_xavier_uniform ...................... False + init_model_with_meta_device ..................... False + initial_loss_scale .............................. 4294967296 + inprocess_active_world_size ..................... 8 + inprocess_barrier_timeout ....................... 120 + inprocess_completion_timeout .................... 120 + inprocess_empty_cuda_cache ...................... False + inprocess_granularity ........................... node + inprocess_hard_timeout .......................... 90 + inprocess_heartbeat_interval .................... 30 + inprocess_heartbeat_timeout ..................... 60 + inprocess_last_call_wait ........................ 1 + inprocess_max_iterations ........................ None + inprocess_monitor_process_interval .............. 1.0 + inprocess_monitor_thread_interval ............... 1.0 + inprocess_progress_watchdog_interval ............ 1.0 + inprocess_restart ............................... False + inprocess_soft_timeout .......................... 60 + inprocess_termination_grace_time ................ 1 + is_hybrid_model ................................. False + iter_per_epoch .................................. 1250 + iterations_to_skip .............................. [] + keep_fp8_transpose_cache_when_using_custom_fsdp . False + kv_channels ..................................... 64 + kv_lora_rank .................................... 32 + lazy_mpu_init ................................... None + load ............................................ gpt-checkpoint + load_model_opt_format ........................... False + local_rank ...................................... 0 + log_interval .................................... 1 + log_loss_scale_to_tensorboard ................... True + log_memory_to_tensorboard ....................... False + log_num_zeros_in_grad ........................... False + log_params_norm ................................. False + log_progress .................................... False + log_straggler ................................... False + log_throughput .................................. False + log_timers_to_tensorboard ....................... False + log_validation_ppl_to_tensorboard ............... False + log_world_size_to_tensorboard ................... False + logging_level ................................... 0 + loss_scale ...................................... None + loss_scale_window ............................... 1000 + lr .............................................. 0.0005 + lr_decay_iters .................................. 150000 + lr_decay_samples ................................ None + lr_decay_style .................................. cosine + lr_warmup_fraction .............................. None + lr_warmup_init .................................. 0.0 + lr_warmup_iters ................................. 2 + lr_warmup_samples ............................... 0 + lr_wsd_decay_iters .............................. None + lr_wsd_decay_samples ............................ None + lr_wsd_decay_style .............................. exponential + main_grads_dtype ................................ torch.float32 + main_params_dtype ............................... torch.float32 + make_vocab_size_divisible_by .................... 128 + mamba_head_dim .................................. 64 + mamba_num_groups ................................ 8 + mamba_num_heads ................................. None + mamba_state_dim ................................. 128 + manual_gc ....................................... False + manual_gc_eval .................................. True + manual_gc_interval .............................. 0 + mask_factor ..................................... 1.0 + mask_prob ....................................... 0.15 + mask_type ....................................... random + masked_softmax_fusion ........................... True + max_position_embeddings ......................... 2048 + max_tokens_to_oom ............................... 12000 + memory_snapshot_path ............................ snapshot.pickle + merge_file ...................................... merges.txt + micro_batch_size ................................ 1 + microbatch_group_size_per_vp_stage .............. None + mid_level_dataset_surplus ....................... 0.005 + min_loss_scale .................................. 1.0 + min_lr .......................................... 0.0 + mlp_chunks_for_prefill .......................... 1 + mmap_bin_files .................................. True + mock_data ....................................... True + moe_apply_probs_on_input ........................ False + moe_aux_loss_coeff .............................. 0.0 + moe_enable_deepep ............................... False + moe_expert_capacity_factor ...................... None + moe_extended_tp ................................. False + moe_ffn_hidden_size ............................. None + moe_grouped_gemm ................................ False + moe_input_jitter_eps ............................ None + moe_layer_freq .................................. 1 + moe_layer_recompute ............................. False + moe_pad_expert_input_to_capacity ................ False + moe_per_layer_logging ........................... False + moe_permute_fusion .............................. False + moe_router_bias_update_rate ..................... 0.001 + moe_router_dtype ................................ None + moe_router_enable_expert_bias ................... False + moe_router_force_load_balancing ................. False + moe_router_group_topk ........................... None + moe_router_load_balancing_type .................. aux_loss + moe_router_num_groups ........................... None + moe_router_padding_for_fp8 ...................... False + moe_router_pre_softmax .......................... False + moe_router_score_function ....................... softmax + moe_router_topk ................................. 2 + moe_router_topk_scaling_factor .................. None + moe_shared_expert_intermediate_size ............. None + moe_shared_expert_overlap ....................... False + moe_token_dispatcher_type ....................... allgather + moe_token_drop_policy ........................... probs + moe_use_legacy_grouped_gemm ..................... False + moe_use_upcycling ............................... False + moe_z_loss_coeff ................................ None + mrope_section ................................... None + mscale .......................................... 1.0 + mscale_all_dim .................................. 1.0 + mtp_loss_scaling_factor ......................... 0.1 + mtp_num_layers .................................. None + multi_latent_attention .......................... False + nccl_all_reduce_for_prefill ..................... False + nccl_communicator_config_path ................... None + nccl_ub ......................................... False + no_load_optim ................................... None + no_load_rng ..................................... None + no_persist_layer_norm ........................... False + no_rope_freq .................................... None + no_save_optim ................................... None + no_save_rng ..................................... None + non_persistent_ckpt_type ........................ None + non_persistent_global_ckpt_dir .................. None + non_persistent_local_ckpt_algo .................. fully_parallel + non_persistent_local_ckpt_dir ................... None + non_persistent_save_interval .................... None + norm_epsilon .................................... 1e-05 + normalization ................................... LayerNorm + num_attention_heads ............................. 64 + num_channels .................................... 3 + num_classes ..................................... 1000 + num_dataset_builder_threads ..................... 1 + num_distributed_optimizer_instances ............. 1 + num_experts ..................................... None + num_layers ...................................... 2 + num_layers_at_end_in_bf16 ....................... 1 + num_layers_at_start_in_bf16 ..................... 1 + num_layers_per_virtual_pipeline_stage ........... None + num_query_groups ................................ 16 + num_virtual_stages_per_pipeline_rank ............ None + num_workers ..................................... 2 + object_storage_cache_path ....................... None + one_logger_async ................................ False + one_logger_project .............................. megatron-lm + one_logger_run_name ............................. None + onnx_safe ....................................... None + openai_gelu ..................................... False + optimizer ....................................... adam + optimizer_cpu_offload ........................... False + optimizer_offload_fraction ...................... 1.0 + output_bert_embeddings .......................... False + overlap_cpu_optimizer_d2h_h2d ................... False + overlap_grad_reduce ............................. False + overlap_p2p_comm ................................ False + overlap_p2p_comm_warmup_flush ................... False + overlap_param_gather ............................ False + overlap_param_gather_with_optimizer_step ........ False + override_opt_param_scheduler .................... False + params_dtype .................................... torch.float16 + patch_dim ....................................... 16 + per_split_data_args_path ........................ None + perform_initialization .......................... True + pin_cpu_grads ................................... True + pin_cpu_params .................................. True + pipeline_model_parallel_comm_backend ............ None + pipeline_model_parallel_size .................... 1 + pipeline_model_parallel_split_rank .............. None + position_embedding_type ......................... learned_absolute + pretrained_checkpoint ........................... None + profile ......................................... False + profile_ranks ................................... [0] + profile_step_end ................................ 12 + profile_step_start .............................. 10 + q_lora_rank ..................................... None + qk_head_dim ..................................... 128 + qk_l2_norm ...................................... False + qk_layernorm .................................... False + qk_pos_emb_head_dim ............................. 64 + query_in_block_prob ............................. 0.1 + rampup_batch_size ............................... None + rank ............................................ 0 + recompute_granularity ........................... None + recompute_method ................................ None + recompute_modules ............................... None + recompute_num_layers ............................ None + record_memory_history ........................... False + relative_attention_max_distance ................. 128 + relative_attention_num_buckets .................. 32 + replication ..................................... False + replication_factor .............................. 2 + replication_jump ................................ None + rerun_mode ...................................... disabled + reset_attention_mask ............................ False + reset_position_ids .............................. False + result_rejected_tracker_filename ................ None + retriever_report_topk_accuracies ................ [] + retriever_score_scaling ......................... False + retriever_seq_length ............................ 256 + retro_add_retriever ............................. False + retro_attention_gate ............................ 1 + retro_cyclic_train_iters ........................ None + retro_encoder_attention_dropout ................. 0.1 + retro_encoder_hidden_dropout .................... 0.1 + retro_encoder_layers ............................ 2 + retro_num_neighbors ............................. 2 + retro_num_retrieved_chunks ...................... 2 + retro_project_dir ............................... None + retro_verify_neighbor_count ..................... True + rope_scaling_factor ............................. 8.0 + rotary_base ..................................... 10000 + rotary_interleaved .............................. False + rotary_percent .................................. 1.0 + rotary_scaling_factor ........................... 1.0 + rotary_seq_len_interpolation_factor ............. None + run_workload_inspector_server ................... False + sample_rate ..................................... 1.0 + save ............................................ gpt-checkpoint + save_interval ................................... 16 + scatter_gather_tensors_in_pipeline .............. True + seed ............................................ 1234 + seq_length ...................................... 2048 + sequence_parallel ............................... False + sgd_momentum .................................... 0.9 + short_seq_prob .................................. 0.1 + skip_train ...................................... False + skipped_train_samples ........................... 0 + spec ............................................ None + split ........................................... None + squared_relu .................................... False + start_weight_decay .............................. 0.1 + straggler_ctrlr_port ............................ 65535 + straggler_minmax_count .......................... 1 + suggested_communication_unit_size ............... None + swiglu .......................................... False + swin_backbone_type .............................. tiny + symmetric_ar_type ............................... None + te_rng_tracker .................................. False + tensor_model_parallel_size ...................... 4 + tensorboard_dir ................................. tensorboard-logs/ + tensorboard_log_interval ........................ 1 + tensorboard_queue_size .......................... 1000 + test_data_path .................................. None + test_mode ....................................... False + tiktoken_num_special_tokens ..................... 1000 + tiktoken_pattern ................................ None + tiktoken_special_tokens ......................... None + timing_log_level ................................ 0 + timing_log_option ............................... minmax + titles_data_path ................................ None + tokenizer_model ................................. None + tokenizer_type .................................. GPT2BPETokenizer + torch_fsdp2_reshard_after_forward ............... True + tp_comm_bootstrap_backend ....................... nccl + tp_comm_bulk_dgrad .............................. True + tp_comm_bulk_wgrad .............................. True + tp_comm_overlap ................................. False + tp_comm_overlap_ag .............................. True + tp_comm_overlap_cfg ............................. None + tp_comm_overlap_rs .............................. True + tp_comm_overlap_rs_dgrad ........................ False + tp_comm_split_ag ................................ True + tp_comm_split_rs ................................ True + train_data_path ................................. None + train_iters ..................................... 10 + train_samples ................................... None + train_sync_interval ............................. None + transformer_impl ................................ transformer_engine + transformer_pipeline_model_parallel_size ........ 1 + untie_embeddings_and_output_weights ............. False + use_checkpoint_args ............................. False + use_checkpoint_opt_param_scheduler .............. False + use_cpu_initialization .......................... None + use_custom_fsdp ................................. False + use_dist_ckpt ................................... True + use_dist_ckpt_deprecated ........................ False + use_distributed_optimizer ....................... False + use_flash_attn .................................. False + use_legacy_models ............................... False + use_mp_args_from_checkpoint_args ................ False + use_one_sent_docs ............................... False + use_persistent_ckpt_worker ...................... False + use_precision_aware_optimizer ................... False + use_pytorch_profiler ............................ False + use_ring_exchange_p2p ........................... False + use_rope_scaling ................................ False + use_rotary_position_embeddings .................. False + use_sharp ....................................... False + use_tokenizer_model_from_checkpoint_args ........ True + use_torch_fsdp2 ................................. False + use_torch_optimizer_for_cpu_offload ............. False + use_tp_pp_dp_mapping ............................ False + v_head_dim ...................................... 128 + valid_data_path ................................. None + variable_seq_lengths ............................ False + virtual_pipeline_model_parallel_size ............ None + vision_backbone_type ............................ vit + vision_pretraining .............................. False + vision_pretraining_type ......................... classify + vocab_extra_ids ................................. 0 + vocab_file ...................................... vocab.json + vocab_size ...................................... None + wandb_exp_name .................................. + wandb_project ................................... + wandb_save_dir .................................. + weight_decay .................................... 0.1 + weight_decay_incr_style ......................... constant + wgrad_deferral_limit ............................ 0 + world_size ...................................... 8 + yaml_cfg ........................................ None +-------------------- end of arguments --------------------- +INFO:megatron.core.num_microbatches_calculator:setting number of microbatches to constant 1 +> building GPT2BPETokenizer tokenizer ... + > padded vocab (size: 50257) with 431 dummy tokens (new size: 50688) +INFO:megatron.training.initialize:Setting logging level to 0 +WARNING:megatron.core.rerun_state_machine:RerunStateMachine initialized in mode RerunMode.DISABLED +> initializing torch distributed ... +INFO:megatron.training.initialize:Setting logging level to 0 +INFO:megatron.training.initialize:Setting logging level to 0 +INFO:megatron.training.initialize:Setting logging level to 0 +INFO:megatron.training.initialize:Setting logging level to 0 +INFO:megatron.training.initialize:Setting logging level to 0 +INFO:megatron.training.initialize:Setting logging level to 0 +WARNING: TensorBoard writing requested but is not available (are you using PyTorch 1.1.0 or later?), no TensorBoard logs will be written. +WARNING: one_logger package is required to enable e2e metrics tracking. please go to https://confluence.nvidia.com/display/MLWFO/Package+Repositories for details to install it +INFO:megatron.training.initialize:Setting logging level to 0 +> initialized tensor model parallel with size 4 +> initialized pipeline model parallel with size 1 +> setting random seeds to 1234 ... +> compiling dataset index builder ... +make: Entering directory '/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/datasets' +make: Nothing to be done for 'default'. +make: Leaving directory '/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/datasets' +>>> done with dataset index builder. Compilation time: 0.058 seconds +> compiling and loading fused kernels ... +>>> done with compiling and loading fused kernels. Compilation time: 2.448 seconds +time to initialize megatron (seconds): 8.457 +[after megatron is initialized] datetime: 2025-06-21 21:34:49 +building GPT model ... +>>> embedding>>> embedding + +>>> decoder>>> decoder + +>>> output_layer>>> output_layer + +>>> embedding +>>> decoder +>>> output_layer +>>> embedding>>> embedding + +>>> decoder>>> decoder + +>>> output_layer>>> output_layer + +>>> embedding +>>> decoder +>>> output_layer +>>> embedding +>>> decoder +>>> output_layer + > number of parameters on (tensor, pipeline) model parallel rank (0, 0): 148442112 + > number of parameters on (tensor, pipeline) model parallel rank (1, 0): 148442112 +>>> embedding +>>> decoder +>>> output_layer + > number of parameters on (tensor, pipeline) model parallel rank (3, 0): 148442112 + > number of parameters on (tensor, pipeline) model parallel rank (2, 0): 148442112 > number of parameters on (tensor, pipeline) model parallel rank (0, 0): 148442112 + + > number of parameters on (tensor, pipeline) model parallel rank (2, 0): 148442112 + > number of parameters on (tensor, pipeline) model parallel rank (1, 0): 148442112 + > number of parameters on (tensor, pipeline) model parallel rank (3, 0): 148442112 +INFO:megatron.core.distributed.distributed_data_parallel:Setting up DistributedDataParallel with config DistributedDataParallelConfig(grad_reduce_in_fp32=False, overlap_grad_reduce=False, overlap_param_gather=False, align_param_gather=False, use_distributed_optimizer=False, num_distributed_optimizer_instances=1, check_for_nan_in_grad=False, check_for_large_grads=False, bucket_size=None, pad_buckets_for_high_nccl_busbw=False, average_in_collective=False, fp8_param_gather=False, use_custom_fsdp=False, data_parallel_sharding_strategy='no_shard', gradient_reduce_div_fusion=True, suggested_communication_unit_size=None, preserve_fp32_weights=True, keep_fp8_transpose_cache_when_using_custom_fsdp=False, nccl_ub=False, fsdp_double_buffer=False) +INFO:megatron.core.distributed.param_and_grad_buffer:Number of buckets for gradient all-reduce / reduce-scatter: 1 +Params for bucket 1 (148442112 elements, 148442112 padded size): + module.decoder.layers.1.self_attention.linear_qkv.layer_norm_bias + module.decoder.layers.0.self_attention.linear_qkv.layer_norm_bias + module.decoder.layers.0.self_attention.linear_qkv.layer_norm_weight + module.embedding.position_embeddings.weight + module.decoder.layers.1.mlp.linear_fc1.bias + module.decoder.layers.0.mlp.linear_fc1.bias + module.decoder.layers.1.self_attention.linear_qkv.weight + module.decoder.layers.1.self_attention.linear_proj.weight + module.decoder.layers.0.self_attention.linear_qkv.weight + module.embedding.word_embeddings.weight + module.decoder.layers.1.mlp.linear_fc2.weight + module.decoder.layers.1.self_attention.linear_proj.bias + module.decoder.layers.1.mlp.linear_fc1.layer_norm_bias + module.decoder.layers.0.mlp.linear_fc2.weight + module.decoder.layers.0.mlp.linear_fc1.layer_norm_bias + module.decoder.layers.0.self_attention.linear_proj.weight + module.decoder.final_layernorm.weight + module.decoder.layers.1.mlp.linear_fc1.layer_norm_weight + module.decoder.layers.1.self_attention.linear_qkv.bias + module.decoder.layers.0.mlp.linear_fc2.bias + module.decoder.layers.0.mlp.linear_fc1.layer_norm_weight + module.decoder.layers.0.self_attention.linear_qkv.bias + module.decoder.layers.0.self_attention.linear_proj.bias + module.decoder.layers.1.mlp.linear_fc1.weight + module.decoder.layers.0.mlp.linear_fc1.weight + module.decoder.final_layernorm.bias + module.decoder.layers.1.mlp.linear_fc2.bias + module.decoder.layers.1.self_attention.linear_qkv.layer_norm_weight +INFO:megatron.core.optimizer:Setting up optimizer with config OptimizerConfig(optimizer='adam', lr=0.0005, min_lr=0.0, decoupled_lr=None, decoupled_min_lr=None, weight_decay=0.1, fp16=True, bf16=False, params_dtype=torch.float16, use_precision_aware_optimizer=False, store_param_remainders=True, main_grads_dtype=torch.float32, main_params_dtype=torch.float32, exp_avg_dtype=torch.float32, exp_avg_sq_dtype=torch.float32, loss_scale=None, initial_loss_scale=4294967296, min_loss_scale=1.0, loss_scale_window=1000, hysteresis=2, adam_beta1=0.9, adam_beta2=0.999, adam_eps=1e-08, sgd_momentum=0.9, use_distributed_optimizer=False, overlap_param_gather_with_optimizer_step=False, optimizer_cpu_offload=False, optimizer_offload_fraction=1.0, use_torch_optimizer_for_cpu_offload=False, overlap_cpu_optimizer_d2h_h2d=False, pin_cpu_grads=True, pin_cpu_params=True, clip_grad=1.0, log_num_zeros_in_grad=False, barrier_with_L1_time=True, timers=, config_logger_dir='') +INFO:megatron.core.optimizer_param_scheduler:> learning rate decay style: cosine +WARNING: could not find the metadata file gpt-checkpoint/latest_checkpointed_iteration.txt + will not load any checkpoints and will start from random +(min, max) time across ranks (ms): + load-checkpoint ................................: (4.03, 4.06) +[after model, optimizer, and learning rate scheduler are built] datetime: 2025-06-21 21:34:49 +> building train, validation, and test datasets ... + > datasets target sizes (minimum size): + train: 10 + validation: 1 + test: 1 +INFO:megatron.core.datasets.blended_megatron_dataset_config:Let mock = True, as both blend and blend_per_split are None +INFO:megatron.core.datasets.blended_megatron_dataset_config:Let split = 1,1,1, an arbitrarily even split, as mock is True +INFO:megatron.core.datasets.blended_megatron_dataset_config:Let split_matrix = [(0, 0.3333333333333333), (0.3333333333333333, 0.6666666666666666), (0.6666666666666666, 1.0)] +> building train, validation, and test datasets for GPT ... +INFO:megatron.core.datasets.blended_megatron_dataset_builder:Building MockGPTDataset splits with sizes=(10, 1, 1) and config=GPTDatasetConfig(random_seed=1234, sequence_length=2048, blend=None, blend_per_split=None, split='1,1,1', split_matrix=[(0, 0.3333333333333333), (0.3333333333333333, 0.6666666666666666), (0.6666666666666666, 1.0)], num_dataset_builder_threads=1, path_to_cache=None, mmap_bin_files=True, mock=True, tokenizer=, mid_level_dataset_surplus=0.005, reset_position_ids=False, reset_attention_mask=False, eod_mask_loss=False, create_attention_mask=True, drop_last_partial_validation_sequence=True, add_extra_token_to_sequence=True, object_storage_cache_path=None) +INFO:megatron.core.datasets.gpt_dataset:Build and save the MockGPTDataset train indices +DEBUG:megatron.core.datasets.gpt_dataset:> separate_final_epoch: False +WARNING:megatron.core.datasets.gpt_dataset:Unable to save MockGPTDataset indexes because path_to_cache is None +DEBUG:megatron.core.datasets.gpt_dataset: > time elapsed: 0.005389 seconds +INFO:megatron.core.datasets.gpt_dataset:> total number of samples: 33296 +INFO:megatron.core.datasets.gpt_dataset:> total number of epochs: 1 +INFO:megatron.core.datasets.gpt_dataset:Build and save the MockGPTDataset valid indices +DEBUG:megatron.core.datasets.gpt_dataset:> separate_final_epoch: False +WARNING:megatron.core.datasets.gpt_dataset:Unable to save MockGPTDataset indexes because path_to_cache is None +DEBUG:megatron.core.datasets.gpt_dataset: > time elapsed: 0.002493 seconds +INFO:megatron.core.datasets.gpt_dataset:> total number of samples: 33281 +INFO:megatron.core.datasets.gpt_dataset:> total number of epochs: 1 +INFO:megatron.core.datasets.gpt_dataset:Build and save the MockGPTDataset test indices +DEBUG:megatron.core.datasets.gpt_dataset:> separate_final_epoch: False +WARNING:megatron.core.datasets.gpt_dataset:Unable to save MockGPTDataset indexes because path_to_cache is None +DEBUG:megatron.core.datasets.gpt_dataset: > time elapsed: 0.002502 seconds +INFO:megatron.core.datasets.gpt_dataset:> total number of samples: 33343 +INFO:megatron.core.datasets.gpt_dataset:> total number of epochs: 1 +> finished creating GPT datasets ... +[after dataloaders are built] datetime: 2025-06-21 21:34:49 +done with setup ... +(min, max) time across ranks (ms): + model-and-optimizer-setup ......................: (334.26, 339.99) + train/valid/test-data-iterators-setup ..........: (23.35, 118.62) +training ... +Setting rerun_state_machine.current_iteration to 0... +[before the start of training step] datetime: 2025-06-21 21:34:49 +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096])batch tensor: +batch tensor: attention_mask tokenstorch.Size([2, 1, 4096, 4096]) + batch tensor: position_ids torch.Size([2, 4096]) +torch.Size([2, 4096]) +batch tensor:batch tensor: labelstokens torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +torch.Size([2, 4096])batch tensor: + attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: batch tensor:labels position_ids torch.Size([2, 4096])torch.Size([2, 4096]) + +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +Start exporting trace 0 +Done exporting trace 0 + [2025-06-21 21:35:00] iteration 1/ 10 | consumed samples: 1 | elapsed time per iteration (ms): 10767.0 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 4294967296.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +Number of parameters in transformer block in billions: 0.35 +Number of parameters in embedding layers in billions: 0.21 +Total number of parameters in billions: 0.56 +Number of parameters in most loaded shard in billions: 0.1400 +Theoretical memory footprints: weight and optimizer=2403.18 MB +[Rank 1] (after 1 iterations) memory (MB) | allocated: 1896.98876953125 | max allocated: 2113.56494140625 | reserved: 2554.0 | max reserved: 2554.0 +[Rank 3] (after 1 iterations) memory (MB) | allocated: 1896.98876953125 | max allocated: 2113.56494140625 | reserved: 2554.0 | max reserved: 2554.0[Rank 5] (after 1 iterations) memory (MB) | allocated: 1896.98876953125 | max allocated: 2113.56494140625 | reserved: 2554.0 | max reserved: 2554.0 + +[Rank 6] (after 1 iterations) memory (MB) | allocated: 1896.98876953125 | max allocated: 2113.56494140625 | reserved: 2554.0 | max reserved: 2554.0 +[Rank 2] (after 1 iterations) memory (MB) | allocated: 1896.98876953125 | max allocated: 2113.56494140625 | reserved: 2570.0 | max reserved: 2570.0 +[Rank 7] (after 1 iterations) memory (MB) | allocated: 1896.98876953125 | max allocated: 2113.56494140625 | reserved: 2554.0 | max reserved: 2554.0 +[Rank 0] (after 1 iterations) memory (MB) | allocated: 1896.98876953125 | max allocated: 2113.56494140625 | reserved: 2554.0 | max reserved: 2554.0[Rank 4] (after 1 iterations) memory (MB) | allocated: 1896.98876953125 | max allocated: 2113.56494140625 | reserved: 2554.0 | max reserved: 2554.0 + +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +Start exporting trace 1 +Done exporting trace 1 + [2025-06-21 21:35:00] iteration 2/ 10 | consumed samples: 2 | elapsed time per iteration (ms): 77.7 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 2147483648.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens batch tensor:torch.Size([2, 2048]) +batch tensor after cp: labels tokenstorch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp:torch.Size([2, 4096]) position_ids + torch.Size([2, 2048]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids batch tensor after cp: torch.Size([2, 4096])tokens + torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096])batch tensor: +batch tensor: position_idstokens torch.Size([2, 4096]) +torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens batch tensor after cp: torch.Size([2, 4096])tokens + batch tensor: torch.Size([2, 2048])labels + batch tensor after cp:torch.Size([2, 4096]) +labelsbatch tensor: torch.Size([2, 2048])loss_mask + batch tensor after cp:torch.Size([2, 4096]) +loss_mask batch tensor:torch.Size([2, 2048]) +attention_maskbatch tensor after cp: torch.Size([2, 1, 4096, 4096])attention_mask + batch tensor:torch.Size([2, 1, 2048, 4096]) +position_ids batch tensor after cp:torch.Size([2, 4096]) +position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask batch tensor after cp:torch.Size([2, 4096]) +tokens batch tensor: torch.Size([2, 2048])attention_mask + batch tensor after cp:torch.Size([2, 1, 4096, 4096]) +labels batch tensor:torch.Size([2, 2048]) +batch tensor after cp:position_ids loss_masktorch.Size([2, 4096]) +torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +Start exporting trace 2 +Done exporting trace 2 + [2025-06-21 21:35:00] iteration 3/ 10 | consumed samples: 3 | elapsed time per iteration (ms): 50.0 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 1073741824.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor:batch tensor after cp: loss_masktokens torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 4096])torch.Size([2, 1, 2048, 4096]) + +batch tensor:batch tensor after cp: labelsposition_ids torch.Size([2, 2048])torch.Size([2, 4096]) + +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +Start exporting trace 3 +Done exporting trace 3 + [2025-06-21 21:35:00] iteration 4/ 10 | consumed samples: 4 | elapsed time per iteration (ms): 43.1 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 536870912.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor:batch tensor: tokens tokens torch.Size([2, 4096])torch.Size([2, 4096]) + +batch tensor:batch tensor: labels labelstorch.Size([2, 4096]) +torch.Size([2, 4096])batch tensor: + batch tensor:loss_mask loss_masktorch.Size([2, 4096]) +torch.Size([2, 4096]) +batch tensor: batch tensor:attention_mask attention_mask torch.Size([2, 1, 4096, 4096]) +torch.Size([2, 1, 4096, 4096]) +batch tensor: batch tensor:position_ids torch.Size([2, 4096])position_ids + torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp:batch tensor after cp: labelstokens torch.Size([2, 2048]) +torch.Size([2, 2048])batch tensor after cp: + batch tensor after cp:loss_mask labelstorch.Size([2, 2048]) +torch.Size([2, 2048])batch tensor after cp: + batch tensor after cp:attention_mask loss_masktorch.Size([2, 1, 2048, 4096]) +torch.Size([2, 2048])batch tensor after cp: +position_idsbatch tensor after cp: torch.Size([2, 2048])attention_mask + torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +Start exporting trace 4 +Done exporting trace 4 + [2025-06-21 21:35:00] iteration 5/ 10 | consumed samples: 5 | elapsed time per iteration (ms): 42.5 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 268435456.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor after cp:batch tensor: position_idstokens torch.Size([2, 4096])torch.Size([2, 2048]) + +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor:batch tensor after cp: tokenstokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 4096])torch.Size([2, 2048]) + +batch tensor after cp: batch tensor:loss_mask labelstorch.Size([2, 2048]) +torch.Size([2, 4096])batch tensor after cp: + batch tensor:attention_mask loss_masktorch.Size([2, 1, 2048, 4096]) +torch.Size([2, 4096])batch tensor after cp: + position_idsbatch tensor: torch.Size([2, 2048])attention_mask + torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +Start exporting trace 5 +Done exporting trace 5 + [2025-06-21 21:35:00] iteration 6/ 10 | consumed samples: 6 | elapsed time per iteration (ms): 40.3 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 134217728.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp:batch tensor: tokens torch.Size([2, 2048])tokens + batch tensor after cp: labels torch.Size([2, 2048])torch.Size([2, 4096]) +batch tensor after cp: + loss_maskbatch tensor: torch.Size([2, 2048])labels +torch.Size([2, 4096])batch tensor after cp: + attention_maskbatch tensor: torch.Size([2, 1, 2048, 4096])loss_mask + batch tensor after cp:torch.Size([2, 4096]) +position_ids batch tensor:torch.Size([2, 2048]) +attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp:batch tensor: position_ids torch.Size([2, 2048])tokens + torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +Start exporting trace 6 +Done exporting trace 6 + [2025-06-21 21:35:00] iteration 7/ 10 | consumed samples: 7 | elapsed time per iteration (ms): 40.9 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 67108864.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +Start exporting trace 7 +Done exporting trace 7 + [2025-06-21 21:35:00] iteration 8/ 10 | consumed samples: 8 | elapsed time per iteration (ms): 40.8 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 33554432.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor:batch tensor after cp: position_ids tokenstorch.Size([2, 2048]) +torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels batch tensor after cp:torch.Size([2, 4096]) + batch tensor:tokens loss_masktorch.Size([2, 2048]) +torch.Size([2, 4096]) +batch tensor after cp: batch tensor:labels attention_mask torch.Size([2, 2048]) +torch.Size([2, 1, 4096, 4096])batch tensor after cp: + loss_maskbatch tensor: torch.Size([2, 2048])position_ids + batch tensor after cp:torch.Size([2, 4096]) +attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +Start exporting trace 8 +Done exporting trace 8 + [2025-06-21 21:35:00] iteration 9/ 10 | consumed samples: 9 | elapsed time per iteration (ms): 43.1 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 16777216.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +batch tensor: tokens torch.Size([2, 4096]) +batch tensor: labels torch.Size([2, 4096]) +batch tensor: loss_mask torch.Size([2, 4096]) +batch tensor: attention_mask torch.Size([2, 1, 4096, 4096]) +batch tensor: position_ids torch.Size([2, 4096]) +batch tensor after cp: tokens torch.Size([2, 2048]) +batch tensor after cp: labels torch.Size([2, 2048]) +batch tensor after cp: loss_mask torch.Size([2, 2048]) +batch tensor after cp: attention_mask torch.Size([2, 1, 2048, 4096]) +batch tensor after cp: position_ids torch.Size([2, 2048]) +Start exporting trace 9 +Done exporting trace 9 + [2025-06-21 21:35:00] iteration 10/ 10 | consumed samples: 10 | elapsed time per iteration (ms): 40.8 | learning rate: 0.000000E+00 | global batch size: 1 | loss scale: 8388608.0 | number of skipped iterations: 1 | number of nan iterations: 0 | +[after training is done] datetime: 2025-06-21 21:35:00 +saving checkpoint at iteration 10 to gpt-checkpoint in torch_dist format +DEBUG:megatron.training.checkpointing:rank: 1, takes 0.05642819404602051 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 2, takes 0.05646109580993652 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 7, takes 0.05647540092468262 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 5, takes 0.056499481201171875 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 6, takes 0.05650687217712402 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 3, takes 0.05653095245361328 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 0, takes 0.056817054748535156 to prepare state dict for ckpt +DEBUG:megatron.training.checkpointing:rank: 4, takes 0.05696582794189453 to prepare state dict for ckpt +WARNING:megatron.core.dist_checkpointing.serialization:Overwriting old incomplete / corrupted checkpoint... +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.strategies.fully_parallel:Apply save parallelization +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(417400832), 0), (np.int64(422576128), 1)] +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(417400832), 0), (np.int64(422576128), 1)] +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(417400832), 0), (np.int64(422576128), 1)] +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(417400832), 0), (np.int64(422576128), 1)] +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(417400832), 0), (np.int64(422576128), 1)] +DEBUG:megatron.core.dist_checkpointing.exchange_utils:distribute_shards_to_ranks distribution: [(np.int64(417400832), 0), (np.int64(422576128), 1)]