Upload folder using huggingface_hub

attnserver.run_attnserver.slurm.sh.343188.out.log

@@ -119674,3 +119674,645 @@ Theoretical memory footprints: weight and optimizer=1206.09 MB
 [Rank 59] (after 1 iterations) memory (MB) | allocated: 25522.66455078125 | max allocated: 25670.65087890625 | reserved: 29386.0 | max reserved: 29386.0[Rank 58] (after 1 iterations) memory (MB) | allocated: 25522.66455078125 | max allocated: 25670.65087890625 | reserved: 29386.0 | max reserved: 29386.0
 
 [Rank 37] (after 1 iterations) memory (MB) | allocated: 25522.66455078125 | max allocated: 25670.65087890625 | reserved: 29382.0 | max reserved: 29382.0
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 16384])
+batch tensor after cp: labels torch.Size([1, 16384])
+batch tensor after cp: loss_mask torch.Size([1, 16384])
+batch tensor after cp: attention_mask torch.Size([1, 1, 16384, 131072])
+batch tensor after cp: position_ids torch.Size([1, 16384])
+Start exporting trace 1
+Done exporting trace 1

attnserver.run_attnserver.slurm.sh.343195.out.log

@@ -63261,3 +63261,326 @@ batch tensor after cp: labels torch.Size([1, 24576])
 batch tensor after cp: loss_mask torch.Size([1, 24576])
 batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
 batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+Start exporting trace 5
+Done exporting trace 5
+ [2025-06-21 20:54:04] iteration        6/      10 | consumed samples:            6 | elapsed time per iteration (ms): 82380.2 | 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([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])
+batch tensor: tokens torch.Size([1, 98304])
+batch tensor: labels torch.Size([1, 98304])
+batch tensor: loss_mask torch.Size([1, 98304])
+batch tensor: attention_mask torch.Size([1, 1, 98304, 98304])
+batch tensor: position_ids torch.Size([1, 98304])
+batch tensor after cp: tokens torch.Size([1, 24576])
+batch tensor after cp: labels torch.Size([1, 24576])
+batch tensor after cp: loss_mask torch.Size([1, 24576])
+batch tensor after cp: attention_mask torch.Size([1, 1, 24576, 98304])
+batch tensor after cp: position_ids torch.Size([1, 24576])

attnserver.run_attnserver.slurm.sh.343196.err.log

@@ -184,3 +184,284 @@ W0621 20:53:01.122000 1480440 site-packages/torch/distributed/run.py:766]
 W0621 20:53:01.122000 1480440 site-packages/torch/distributed/run.py:766] *****************************************
 W0621 20:53:01.122000 1480440 site-packages/torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
 W0621 20:53:01.122000 1480440 site-packages/torch/distributed/run.py:766] *****************************************
+[rank16]:[W621 20:53:32.279096536 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 16]  using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank24]:[W621 20:53:32.363909187 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 24]  using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank0]:[W621 20:53:32.051519772 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 0]  using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank5]:[W621 20:53:32.436188333 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 5]  using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank29]:[W621 20:53:32.819664263 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 29]  using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank21]:[W621 20:53:32.743476734 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 21]  using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank13]:[W621 20:53:32.654256940 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 13]  using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank2]:[W621 20:53:32.449061221 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 2]  using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank22]:[W621 20:53:32.756792811 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 22]  using GPU 6 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank18]:[W621 20:53:32.756862251 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 18]  using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank6]:[W621 20:53:32.451596822 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 6]  using GPU 6 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank26]:[W621 20:53:32.837211917 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 26]  using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank14]:[W621 20:53:32.669149161 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 14]  using GPU 6 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank10]:[W621 20:53:32.669192753 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 10]  using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank30]:[W621 20:53:32.838954953 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 30]  using GPU 6 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank15]:[W621 20:53:32.675110418 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 15]  using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank7]:[W621 20:53:32.463604525 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 7]  using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank31]:[W621 20:53:32.846247510 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 31]  using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank23]:[W621 20:53:32.770166045 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 23]  using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank11]:[W621 20:53:32.682046853 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 11]  using GPU 3 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank9]:[W621 20:53:32.682127974 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 9]  using GPU 1 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank27]:[W621 20:53:32.851364172 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 27]  using GPU 3 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank25]:[W621 20:53:32.851527033 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 25]  using GPU 1 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank1]:[W621 20:53:32.468677631 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 1]  using GPU 1 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank3]:[W621 20:53:32.468693698 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 3]  using GPU 3 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank17]:[W621 20:53:32.775797003 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 17]  using GPU 1 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank19]:[W621 20:53:32.775852770 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 19]  using GPU 3 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank8]:[W621 20:53:32.701045847 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 8]  using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank12]:[W621 20:53:32.708214824 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 12]  using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank28]:[W621 20:53:32.877280260 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 28]  using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank4]:[W621 20:53:32.499692504 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 4]  using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank20]:[W621 20:53:32.807377963 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 20]  using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+[rank7]:[W621 20:54:05.852469863 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank6]:[W621 20:54:05.957029135 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank5]:[W621 20:54:05.012904183 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank2]:[W621 20:54:05.023179044 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank3]:[W621 20:54:05.095949824 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank1]:[W621 20:54:05.156542055 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank4]:[W621 20:54:05.261689988 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank0]:[W621 20:54:05.381962495 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank28]:[W621 20:54:05.931318464 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank14]:[W621 20:54:05.779361497 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank31]:[W621 20:54:05.985568953 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank23]:[W621 20:54:05.928259412 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank10]:[W621 20:54:05.855658903 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank30]:[W621 20:54:05.059286001 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank25]:[W621 20:54:05.070427109 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank27]:[W621 20:54:05.071332041 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank15]:[W621 20:54:05.927669030 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank22]:[W621 20:54:05.022212298 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank11]:[W621 20:54:05.932691566 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank9]:[W621 20:54:05.941872807 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank13]:[W621 20:54:05.942988962 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank20]:[W621 20:54:05.035175231 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank29]:[W621 20:54:05.122233118 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank24]:[W621 20:54:05.132815528 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank12]:[W621 20:54:05.964062011 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank17]:[W621 20:54:05.073793095 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank26]:[W621 20:54:05.175702868 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank16]:[W621 20:54:06.105576035 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank18]:[W621 20:54:06.128356475 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank19]:[W621 20:54:06.134867285 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank8]:[W621 20:54:06.094994586 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank21]:[W621 20:54:06.316765263 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
++ set +x
++ set +x
++ set +x
++ set +x
++ for ctx_length in 1024 2048 4096 8192 12288 16384 24576 32768 40960 49152 65536 81920 98304 131072
++ export PROF_CTX_LENGTH=2048
++ PROF_CTX_LENGTH=2048
++ name='/mnt/sharefs/users/hao.zhang/junda/megatron-prof-data--unstable-v5/mytrace.L2048*tp8.cp4.bs2.json'
++ '[' -f '/mnt/sharefs/users/hao.zhang/junda/megatron-prof-data--unstable-v5/mytrace.L2048*tp8.cp4.bs2.json' ']'
++ echo 'Running ctx_length=2048, TP_SIZE=8, CP_SIZE=4, BATCH_SIZE=2'
++ srun bash ./attnserver.sh
++ which python3
++ which python3
++ python3 -m torch.distributed.launch --nproc_per_node 8 --nnodes 4 --node_rank 1 --rdzv_id 343196 --rdzv_backend c10d --rdzv_endpoint fs-mbz-gpu-184:29500 ./pretrain_gpt_profile.py --tensor-model-parallel-size 8 --context-parallel-size 4 --num-layers 2 --hidden-size 4096 --num-attention-heads 64 --group-query-attention --num-query-groups 16 --seq-length 2048 --max-position-embeddings 2048 --micro-batch-size 1 --global-batch-size 1 --lr 0.0005 --train-iters 10 --lr-decay-iters 150000 --lr-decay-style cosine --lr-warmup-iters 2 --weight-decay .1 --adam-beta2 .999 --fp16 --log-interval 1 --save-interval 16 --eval-interval 16 --eval-iters 1 --vocab-file vocab.json --merge-file merges.txt --save gpt-checkpoint --load gpt-checkpoint --logging-level 0 --mock-data --tensorboard-dir tensorboard-logs/
++ python3 -m torch.distributed.launch --nproc_per_node 8 --nnodes 4 --node_rank 3 --rdzv_id 343196 --rdzv_backend c10d --rdzv_endpoint fs-mbz-gpu-184:29500 ./pretrain_gpt_profile.py --tensor-model-parallel-size 8 --context-parallel-size 4 --num-layers 2 --hidden-size 4096 --num-attention-heads 64 --group-query-attention --num-query-groups 16 --seq-length 2048 --max-position-embeddings 2048 --micro-batch-size 1 --global-batch-size 1 --lr 0.0005 --train-iters 10 --lr-decay-iters 150000 --lr-decay-style cosine --lr-warmup-iters 2 --weight-decay .1 --adam-beta2 .999 --fp16 --log-interval 1 --save-interval 16 --eval-interval 16 --eval-iters 1 --vocab-file vocab.json --merge-file merges.txt --save gpt-checkpoint --load gpt-checkpoint --logging-level 0 --mock-data --tensorboard-dir tensorboard-logs/
++ which python3
++ python3 -m torch.distributed.launch --nproc_per_node 8 --nnodes 4 --node_rank 2 --rdzv_id 343196 --rdzv_backend c10d --rdzv_endpoint fs-mbz-gpu-184:29500 ./pretrain_gpt_profile.py --tensor-model-parallel-size 8 --context-parallel-size 4 --num-layers 2 --hidden-size 4096 --num-attention-heads 64 --group-query-attention --num-query-groups 16 --seq-length 2048 --max-position-embeddings 2048 --micro-batch-size 1 --global-batch-size 1 --lr 0.0005 --train-iters 10 --lr-decay-iters 150000 --lr-decay-style cosine --lr-warmup-iters 2 --weight-decay .1 --adam-beta2 .999 --fp16 --log-interval 1 --save-interval 16 --eval-interval 16 --eval-iters 1 --vocab-file vocab.json --merge-file merges.txt --save gpt-checkpoint --load gpt-checkpoint --logging-level 0 --mock-data --tensorboard-dir tensorboard-logs/
++ which python3
++ python3 -m torch.distributed.launch --nproc_per_node 8 --nnodes 4 --node_rank 0 --rdzv_id 343196 --rdzv_backend c10d --rdzv_endpoint fs-mbz-gpu-184:29500 ./pretrain_gpt_profile.py --tensor-model-parallel-size 8 --context-parallel-size 4 --num-layers 2 --hidden-size 4096 --num-attention-heads 64 --group-query-attention --num-query-groups 16 --seq-length 2048 --max-position-embeddings 2048 --micro-batch-size 1 --global-batch-size 1 --lr 0.0005 --train-iters 10 --lr-decay-iters 150000 --lr-decay-style cosine --lr-warmup-iters 2 --weight-decay .1 --adam-beta2 .999 --fp16 --log-interval 1 --save-interval 16 --eval-interval 16 --eval-iters 1 --vocab-file vocab.json --merge-file merges.txt --save gpt-checkpoint --load gpt-checkpoint --logging-level 0 --mock-data --tensorboard-dir tensorboard-logs/
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/distributed/launch.py:207: FutureWarning: The module torch.distributed.launch is deprecated
+and will be removed in future. Use torchrun.
+Note that --use-env is set by default in torchrun.
+If your script expects `--local-rank` argument to be set, please
+change it to read from `os.environ['LOCAL_RANK']` instead. See 
+https://pytorch.org/docs/stable/distributed.html#launch-utility for 
+further instructions
+
+  main()
+W0621 20:54:12.173000 1956323 site-packages/torch/distributed/run.py:766] 
+W0621 20:54:12.173000 1956323 site-packages/torch/distributed/run.py:766] *****************************************
+W0621 20:54:12.173000 1956323 site-packages/torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
+W0621 20:54:12.173000 1956323 site-packages/torch/distributed/run.py:766] *****************************************
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/distributed/launch.py:207: FutureWarning: The module torch.distributed.launch is deprecated
+and will be removed in future. Use torchrun.
+Note that --use-env is set by default in torchrun.
+If your script expects `--local-rank` argument to be set, please
+change it to read from `os.environ['LOCAL_RANK']` instead. See 
+https://pytorch.org/docs/stable/distributed.html#launch-utility for 
+further instructions
+
+  main()
+W0621 20:54:12.274000 1139880 site-packages/torch/distributed/run.py:766] 
+W0621 20:54:12.274000 1139880 site-packages/torch/distributed/run.py:766] *****************************************
+W0621 20:54:12.274000 1139880 site-packages/torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
+W0621 20:54:12.274000 1139880 site-packages/torch/distributed/run.py:766] *****************************************
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/distributed/launch.py:207: FutureWarning: The module torch.distributed.launch is deprecated
+and will be removed in future. Use torchrun.
+Note that --use-env is set by default in torchrun.
+If your script expects `--local-rank` argument to be set, please
+change it to read from `os.environ['LOCAL_RANK']` instead. See 
+https://pytorch.org/docs/stable/distributed.html#launch-utility for 
+further instructions
+
+  main()
+W0621 20:54:12.359000 1484017 site-packages/torch/distributed/run.py:766] 
+W0621 20:54:12.359000 1484017 site-packages/torch/distributed/run.py:766] *****************************************
+W0621 20:54:12.359000 1484017 site-packages/torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
+W0621 20:54:12.359000 1484017 site-packages/torch/distributed/run.py:766] *****************************************
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/distributed/launch.py:207: FutureWarning: The module torch.distributed.launch is deprecated
+and will be removed in future. Use torchrun.
+Note that --use-env is set by default in torchrun.
+If your script expects `--local-rank` argument to be set, please
+change it to read from `os.environ['LOCAL_RANK']` instead. See 
+https://pytorch.org/docs/stable/distributed.html#launch-utility for 
+further instructions
+
+  main()
+W0621 20:54:12.445000 1281491 site-packages/torch/distributed/run.py:766] 
+W0621 20:54:12.445000 1281491 site-packages/torch/distributed/run.py:766] *****************************************
+W0621 20:54:12.445000 1281491 site-packages/torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
+W0621 20:54:12.445000 1281491 site-packages/torch/distributed/run.py:766] *****************************************
+[rank8]:[W621 20:54:38.038379846 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 8]  using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank16]:[W621 20:54:38.135510632 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 16]  using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank0]:[W621 20:54:38.871244986 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 0]  using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank24]:[W621 20:54:38.537586267 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 24]  using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank5]:[W621 20:54:38.181728316 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 5]  using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank7]:[W621 20:54:38.181922829 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 7]  using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank2]:[W621 20:54:38.181931209 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 2]  using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank4]:[W621 20:54:38.182004179 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 4]  using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank18]:[W621 20:54:38.488096925 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 18]  using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank23]:[W621 20:54:38.488106584 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 23]  using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank21]:[W621 20:54:38.488118308 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 21]  using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank26]:[W621 20:54:38.565763990 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 26]  using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank20]:[W621 20:54:38.488286233 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 20]  using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank28]:[W621 20:54:38.565817999 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 28]  using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank6]:[W621 20:54:38.183551244 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 6]  using GPU 6 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank10]:[W621 20:54:38.397740736 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 10]  using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank15]:[W621 20:54:38.397740704 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 15]  using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank29]:[W621 20:54:38.567030553 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 29]  using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.

attnserver.run_attnserver.slurm.sh.343197.err.log

@@ -184,3 +184,266 @@ W0621 20:53:05.653000 3302834 site-packages/torch/distributed/run.py:766]
 W0621 20:53:05.653000 3302834 site-packages/torch/distributed/run.py:766] *****************************************
 W0621 20:53:05.653000 3302834 site-packages/torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
 W0621 20:53:05.653000 3302834 site-packages/torch/distributed/run.py:766] *****************************************
+[rank0]:[W621 20:53:36.476389910 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 0]  using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank8]:[W621 20:53:36.397410647 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 8]  using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank16]:[W621 20:53:36.909565270 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 16]  using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank7]:[W621 20:53:37.214991830 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 7]  using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank15]:[W621 20:53:37.103476459 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 15]  using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank23]:[W621 20:53:37.556635144 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 23]  using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank31]:[W621 20:53:37.626428569 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 31]  using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank24]:[W621 20:53:37.689221072 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 24]  using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank4]:[W621 20:53:37.287112172 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 4]  using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank2]:[W621 20:53:37.287117402 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 2]  using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank26]:[W621 20:53:37.696026079 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 26]  using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank20]:[W621 20:53:37.628797877 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 20]  using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank12]:[W621 20:53:37.176861429 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 12]  using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank18]:[W621 20:53:37.633084211 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 18]  using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank3]:[W621 20:53:37.294993219 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 3]  using GPU 3 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank10]:[W621 20:53:37.181751478 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 10]  using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank5]:[W621 20:53:37.295509111 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 5]  using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank29]:[W621 20:53:37.704865500 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 29]  using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank27]:[W621 20:53:37.704869504 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 27]  using GPU 3 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank28]:[W621 20:53:37.705096442 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 28]  using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank11]:[W621 20:53:37.191090393 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 11]  using GPU 3 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank19]:[W621 20:53:37.645800975 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 19]  using GPU 3 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank21]:[W621 20:53:37.646140458 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 21]  using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank1]:[W621 20:53:37.309853849 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 1]  using GPU 1 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank9]:[W621 20:53:37.197524871 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 9]  using GPU 1 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank17]:[W621 20:53:37.647656849 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 17]  using GPU 1 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank6]:[W621 20:53:37.317487619 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 6]  using GPU 6 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank13]:[W621 20:53:37.200277507 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 13]  using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank25]:[W621 20:53:37.720994991 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 25]  using GPU 1 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank22]:[W621 20:53:37.664677490 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 22]  using GPU 6 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank14]:[W621 20:53:37.208867094 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 14]  using GPU 6 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+[rank30]:[W621 20:53:37.725932637 ProcessGroupNCCL.cpp:4715] [PG ID 0 PG GUID 0 Rank 30]  using GPU 6 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can pecify device_id in init_process_group() to force use of a particular device.
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/junda/attnserver-megatron/megatron/core/models/gpt/gpt_layer_specs.py:94: UserWarning: The fp8 argument in "get_gpt_layer_with_transformer_engine_spec" has been deprecated and will be removed soon. Please update your code accordingly.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/transformer_engine/pytorch/cpu_offload.py:595: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
+  warnings.warn(
+[rank4]:[W621 20:54:13.424066979 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank5]:[W621 20:54:13.521024731 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank7]:[W621 20:54:13.539679935 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank6]:[W621 20:54:13.548186503 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank2]:[W621 20:54:13.552759866 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank3]:[W621 20:54:13.612504683 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank1]:[W621 20:54:13.686343546 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank0]:[W621 20:54:13.721086784 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank9]:[W621 20:54:13.791768962 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank25]:[W621 20:54:14.454751915 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank13]:[W621 20:54:14.967778716 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank8]:[W621 20:54:14.979701392 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank27]:[W621 20:54:14.507683165 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank31]:[W621 20:54:14.523350337 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank26]:[W621 20:54:14.526671273 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank28]:[W621 20:54:14.528885182 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank23]:[W621 20:54:14.475122137 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank20]:[W621 20:54:14.510825805 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank24]:[W621 20:54:14.632187007 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank16]:[W621 20:54:14.573884922 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank17]:[W621 20:54:14.628522551 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank30]:[W621 20:54:14.716619743 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank29]:[W621 20:54:14.721612508 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank15]:[W621 20:54:14.213641707 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank19]:[W621 20:54:14.671145596 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank21]:[W621 20:54:14.683378062 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank18]:[W621 20:54:14.710130671 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank10]:[W621 20:54:14.267877352 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank22]:[W621 20:54:14.749166904 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank12]:[W621 20:54:14.299484416 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank14]:[W621 20:54:14.306472709 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+[rank11]:[W621 20:54:14.308930759 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
++ set +x
++ set +x
++ set +x
++ set +x
++ for ctx_length in 1024 2048 4096 8192 12288 16384 24576 32768 40960 49152 65536 81920 98304 131072
++ export PROF_CTX_LENGTH=2048
++ PROF_CTX_LENGTH=2048
++ name='/mnt/sharefs/users/hao.zhang/junda/megatron-prof-data--unstable-v5/mytrace.L2048*tp8.cp4.bs4.json'
++ '[' -f '/mnt/sharefs/users/hao.zhang/junda/megatron-prof-data--unstable-v5/mytrace.L2048*tp8.cp4.bs4.json' ']'
++ echo 'Running ctx_length=2048, TP_SIZE=8, CP_SIZE=4, BATCH_SIZE=4'
++ srun bash ./attnserver.sh
++ which python3
++ python3 -m torch.distributed.launch --nproc_per_node 8 --nnodes 4 --node_rank 0 --rdzv_id 343197 --rdzv_backend c10d --rdzv_endpoint fs-mbz-gpu-852:29500 ./pretrain_gpt_profile.py --tensor-model-parallel-size 8 --context-parallel-size 4 --num-layers 2 --hidden-size 4096 --num-attention-heads 64 --group-query-attention --num-query-groups 16 --seq-length 2048 --max-position-embeddings 2048 --micro-batch-size 1 --global-batch-size 1 --lr 0.0005 --train-iters 10 --lr-decay-iters 150000 --lr-decay-style cosine --lr-warmup-iters 2 --weight-decay .1 --adam-beta2 .999 --fp16 --log-interval 1 --save-interval 16 --eval-interval 16 --eval-iters 1 --vocab-file vocab.json --merge-file merges.txt --save gpt-checkpoint --load gpt-checkpoint --logging-level 0 --mock-data --tensorboard-dir tensorboard-logs/
++ which python3
++ python3 -m torch.distributed.launch --nproc_per_node 8 --nnodes 4 --node_rank 2 --rdzv_id 343197 --rdzv_backend c10d --rdzv_endpoint fs-mbz-gpu-852:29500 ./pretrain_gpt_profile.py --tensor-model-parallel-size 8 --context-parallel-size 4 --num-layers 2 --hidden-size 4096 --num-attention-heads 64 --group-query-attention --num-query-groups 16 --seq-length 2048 --max-position-embeddings 2048 --micro-batch-size 1 --global-batch-size 1 --lr 0.0005 --train-iters 10 --lr-decay-iters 150000 --lr-decay-style cosine --lr-warmup-iters 2 --weight-decay .1 --adam-beta2 .999 --fp16 --log-interval 1 --save-interval 16 --eval-interval 16 --eval-iters 1 --vocab-file vocab.json --merge-file merges.txt --save gpt-checkpoint --load gpt-checkpoint --logging-level 0 --mock-data --tensorboard-dir tensorboard-logs/
++ which python3
++ python3 -m torch.distributed.launch --nproc_per_node 8 --nnodes 4 --node_rank 3 --rdzv_id 343197 --rdzv_backend c10d --rdzv_endpoint fs-mbz-gpu-852:29500 ./pretrain_gpt_profile.py --tensor-model-parallel-size 8 --context-parallel-size 4 --num-layers 2 --hidden-size 4096 --num-attention-heads 64 --group-query-attention --num-query-groups 16 --seq-length 2048 --max-position-embeddings 2048 --micro-batch-size 1 --global-batch-size 1 --lr 0.0005 --train-iters 10 --lr-decay-iters 150000 --lr-decay-style cosine --lr-warmup-iters 2 --weight-decay .1 --adam-beta2 .999 --fp16 --log-interval 1 --save-interval 16 --eval-interval 16 --eval-iters 1 --vocab-file vocab.json --merge-file merges.txt --save gpt-checkpoint --load gpt-checkpoint --logging-level 0 --mock-data --tensorboard-dir tensorboard-logs/
++ which python3
++ python3 -m torch.distributed.launch --nproc_per_node 8 --nnodes 4 --node_rank 1 --rdzv_id 343197 --rdzv_backend c10d --rdzv_endpoint fs-mbz-gpu-852:29500 ./pretrain_gpt_profile.py --tensor-model-parallel-size 8 --context-parallel-size 4 --num-layers 2 --hidden-size 4096 --num-attention-heads 64 --group-query-attention --num-query-groups 16 --seq-length 2048 --max-position-embeddings 2048 --micro-batch-size 1 --global-batch-size 1 --lr 0.0005 --train-iters 10 --lr-decay-iters 150000 --lr-decay-style cosine --lr-warmup-iters 2 --weight-decay .1 --adam-beta2 .999 --fp16 --log-interval 1 --save-interval 16 --eval-interval 16 --eval-iters 1 --vocab-file vocab.json --merge-file merges.txt --save gpt-checkpoint --load gpt-checkpoint --logging-level 0 --mock-data --tensorboard-dir tensorboard-logs/
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/distributed/launch.py:207: FutureWarning: The module torch.distributed.launch is deprecated
+and will be removed in future. Use torchrun.
+Note that --use-env is set by default in torchrun.
+If your script expects `--local-rank` argument to be set, please
+change it to read from `os.environ['LOCAL_RANK']` instead. See 
+https://pytorch.org/docs/stable/distributed.html#launch-utility for 
+further instructions
+
+  main()
+W0621 20:54:20.099000 3375010 site-packages/torch/distributed/run.py:766] 
+W0621 20:54:20.099000 3375010 site-packages/torch/distributed/run.py:766] *****************************************
+W0621 20:54:20.099000 3375010 site-packages/torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
+W0621 20:54:20.099000 3375010 site-packages/torch/distributed/run.py:766] *****************************************
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/distributed/launch.py:207: FutureWarning: The module torch.distributed.launch is deprecated
+and will be removed in future. Use torchrun.
+Note that --use-env is set by default in torchrun.
+If your script expects `--local-rank` argument to be set, please
+change it to read from `os.environ['LOCAL_RANK']` instead. See 
+https://pytorch.org/docs/stable/distributed.html#launch-utility for 
+further instructions
+
+  main()
+W0621 20:54:20.165000 3306626 site-packages/torch/distributed/run.py:766] 
+W0621 20:54:20.165000 3306626 site-packages/torch/distributed/run.py:766] *****************************************
+W0621 20:54:20.165000 3306626 site-packages/torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
+W0621 20:54:20.165000 3306626 site-packages/torch/distributed/run.py:766] *****************************************
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/distributed/launch.py:207: FutureWarning: The module torch.distributed.launch is deprecated
+and will be removed in future. Use torchrun.
+Note that --use-env is set by default in torchrun.
+If your script expects `--local-rank` argument to be set, please
+change it to read from `os.environ['LOCAL_RANK']` instead. See 
+https://pytorch.org/docs/stable/distributed.html#launch-utility for 
+further instructions
+
+  main()
+W0621 20:54:20.250000 87332 site-packages/torch/distributed/run.py:766] 
+W0621 20:54:20.250000 87332 site-packages/torch/distributed/run.py:766] *****************************************
+W0621 20:54:20.250000 87332 site-packages/torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
+W0621 20:54:20.250000 87332 site-packages/torch/distributed/run.py:766] *****************************************
+/mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/distributed/launch.py:207: FutureWarning: The module torch.distributed.launch is deprecated
+and will be removed in future. Use torchrun.
+Note that --use-env is set by default in torchrun.
+If your script expects `--local-rank` argument to be set, please
+change it to read from `os.environ['LOCAL_RANK']` instead. See 
+https://pytorch.org/docs/stable/distributed.html#launch-utility for 
+further instructions
+
+  main()
+W0621 20:54:20.309000 2008025 site-packages/torch/distributed/run.py:766] 
+W0621 20:54:20.309000 2008025 site-packages/torch/distributed/run.py:766] *****************************************
+W0621 20:54:20.309000 2008025 site-packages/torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
+W0621 20:54:20.309000 2008025 site-packages/torch/distributed/run.py:766] *****************************************

attnserver.run_attnserver.slurm.sh.343201.out.log

@@ -37843,3 +37843,189 @@ batch tensor after cp: labels torch.Size([1, 65536])
 batch tensor after cp: loss_mask torch.Size([1, 65536])
 batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
 batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+Start exporting trace 7
+Done exporting trace 7
+ [2025-06-21 20:53:31] iteration        8/      10 | consumed samples:            8 | elapsed time per iteration (ms): 40195.5 | 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([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+batch tensor: tokens torch.Size([1, 131072])
+batch tensor: labels torch.Size([1, 131072])
+batch tensor: loss_mask torch.Size([1, 131072])
+batch tensor: attention_mask torch.Size([1, 1, 131072, 131072])
+batch tensor: position_ids torch.Size([1, 131072])
+batch tensor after cp: tokens torch.Size([1, 65536])
+batch tensor after cp: labels torch.Size([1, 65536])
+batch tensor after cp: loss_mask torch.Size([1, 65536])
+batch tensor after cp: attention_mask torch.Size([1, 1, 65536, 131072])
+batch tensor after cp: position_ids torch.Size([1, 65536])
+Start exporting trace 8
+Done exporting trace 8
+ [2025-06-21 20:54:11] iteration        9/      10 | consumed samples:            9 | elapsed time per iteration (ms): 39801.9 | learning rate: 0.000000E+00 | global batch size:     1 | loss scale: 16777216.0 | number of skipped iterations:   1 | number of nan iterations:   0 |