MaxView

← Back to run

Log Summary

XPK Start: Sat Apr 18 22:17:57 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-18 22:18:22.204533: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303)
I0418 22:18:22.387958 138714412242752 max_utils.py:273] Attempting to initialize the jax distributed system...
I0418 22:18:31.429986 138714412242752 distributed.py:149] Starting JAX distributed service on [::]:8482
I0418 22:18:31.432466 138714412242752 distributed.py:172] Connecting to JAX distributed service on mt-01-base-w3y4l-slice-job-0-0.mt-01-base-w3y4l:8482
I0418 22:18:33.117606 138714412242752 max_utils.py:284] Jax distributed system initialized!
I0418 22:18:38.280078 138714412242752 max_utils.py:800] System Information: Jax Version: 0.9.2
I0418 22:18:38.280184 138714412242752 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0418 22:18:38.280224 138714412242752 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365
I0418 22:18:38.280257 138714412242752 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0418 22:18:38.971231 138714412242752 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0418 22:18:39.084744 138714412242752 checkpointing.py:688] Setting up checkpoint logger...
I0418 22:18:39.084884 138714412242752 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0418 22:18:39.084944 138714412242752 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0418 22:18:39.085160 138714412242752 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7e2856d1bfe0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0418 22:18:42.705353 138714412242752 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0418 22:18:42.705636 138714412242752 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e1498283380>}, handler_registry=None
I0418 22:18:42.705885 138714412242752 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e1498283380>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`.
I0418 22:18:42.705933 138714412242752 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e1498259dc0>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`.
I0418 22:18:42.705968 138714412242752 composite_checkpoint_handler.py:505] Initialized registry DefaultCheckpointHandlerRegistry({('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e1498283380>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e1498283380>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e1498259dc0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e1498259dc0>}).
I0418 22:18:42.706276 138714412242752 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0418 22:18:42.706343 138714412242752 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e1468425120> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0418 22:18:44.324842 138714412242752 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints
I0418 22:18:44.770123 138714412242752 checkpoint_manager.py:929] [process=0][thread=MainThread] CheckpointManager created,  primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e1498281250>
I0418 22:18:44.770291 138714412242752 checkpointing.py:302] Checkpoint manager created!
I0418 22:18:45.718836 138714412242752 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0418 22:18:45.718955 138714412242752 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0418 22:18:47.081616 138714412242752 maxtext_utils.py:1836]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0418 22:18:47.081732 138714412242752 maxtext_utils.py:1836]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0418 22:18:47.081783 138714412242752 maxtext_utils.py:1836]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0418 22:18:47.081839 138714412242752 maxtext_utils.py:1836]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0418 22:18:47.081880 138714412242752 maxtext_utils.py:1836]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0418 22:18:47.081915 138714412242752 maxtext_utils.py:1836]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0418 22:18:47.081950 138714412242752 maxtext_utils.py:1836]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0418 22:18:47.081986 138714412242752 maxtext_utils.py:1836]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0418 22:18:47.082020 138714412242752 maxtext_utils.py:1836]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0418 22:18:47.082053 138714412242752 maxtext_utils.py:1836]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0418 22:18:47.082085 138714412242752 maxtext_utils.py:1836]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0418 22:18:47.082115 138714412242752 maxtext_utils.py:1836]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0418 22:18:47.484982 138714412242752 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0418 22:18:47.485073 138714412242752 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 22:18:47.490775 138714412242752 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0418 22:18:47.490845 138714412242752 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 22:18:47.507219 138714412242752 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 22:18:47.507281 138714412242752 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 22:18:47.522826 138714412242752 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 22:18:47.522883 138714412242752 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 22:18:47.547987 138714412242752 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 22:18:47.548057 138714412242752 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 22:18:47.563828 138714412242752 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 22:18:47.563888 138714412242752 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 22:18:47.579617 138714412242752 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 22:18:47.579679 138714412242752 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 22:18:47.610101 138714412242752 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0418 22:18:47.610173 138714412242752 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 22:18:47.630011 138714412242752 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0418 22:18:47.630075 138714412242752 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0418 22:18:59.158669 138714412242752 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0418 22:18:59.168151 138714412242752 metric_logger.py:301] number parameters: 1.104 billion
I0418 22:19:10.772740 138714412242752 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0418 22:19:10.927720 138714412242752 checkpointing.py:798] Waited 0.15496420860290527 seconds for step 0 to finish before starting checkpointing.
I0418 22:19:10.930238 138714412242752 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 22:19:10.931881 138714412242752 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0418 22:19:10.933711 138714412242752 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0.
I0418 22:19:11.643029 138714412242752 signaling_client.py:364] Using JaxDistributedSignalingClient
I0418 22:19:11.662660 138714412242752 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0418 22:19:11.662731 138714412242752 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0418 22:19:11.922698 138587163236096 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0
I0418 22:19:11.936961 138714412242752 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.275739s
I0418 22:19:11.937787 138714412242752 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.380 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28674840927124023 s) (per-host)
I0418 22:19:11.937871 138714412242752 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.286851s (batch_requests_ready=0.003236s, total_serialization_initiated=0.282860s, others=0.000756s)
I0418 22:19:11.937954 138714412242752 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.290879s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.290863)
I0418 22:19:11.938755 138714412242752 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.01 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0.
I0418 22:19:11.939054 138587171628800 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 22:39:11.939018
I0418 22:19:11.947130 138714412242752 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 22:19:11.947404 138585118324480 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 22:19:11.947526 138714412242752 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776550750.9302206, 'wait_for_prev_duration_secs': 5.91278076171875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776550750.9319184, 'checkpointer_blocking_duration_secs': 1.007279634475708, 'get_old_steps_start_time': 1776550751.9392252, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1776550750.928152, 'checkpoint_manager_blocking_duration_secs': 1.019336223602295}
I0418 22:19:11.947644 138714412242752 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0418 22:19:11.947711 138714412242752 max_utils.py:750] 
Memstats: After params initialized:
I0418 22:19:11.947761 138714412242752 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0418 22:19:11.947793 138714412242752 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0418 22:19:11.947837 138714412242752 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0418 22:19:11.947862 138714412242752 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0418 22:19:12.313031 138714412242752 metric_logger.py:196] completed step: 0, seconds: 11.288, TFLOP/s/device: 1.204, Tokens/s/device: 181.435, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469
I0418 22:19:12.316128 138714412242752 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/tensorboard/'
I0418 22:19:12.734215 138714412242752 metric_logger.py:196] completed step: 1, seconds: 1.537, TFLOP/s/device: 8.843, Tokens/s/device: 1332.877, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469
I0418 22:19:12.876353 138714412242752 metric_logger.py:196] completed step: 2, seconds: 0.422, TFLOP/s/device: 32.185, Tokens/s/device: 4851.241, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133
I0418 22:19:13.023907 138714412242752 metric_logger.py:196] completed step: 3, seconds: 0.014, TFLOP/s/device: 951.013, Tokens/s/device: 143347.099, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16855.789
I0418 22:19:13.671522 138587146450688 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0/items
I0418 22:19:13.799914 138584048371456 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776550753414070078, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776550753414070078, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0/_CHECKPOINT_METADATA
I0418 22:19:14.786756    2818 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0418 22:19:16.957075 138585172866816 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0/items/array_metadatas/process_0
I0418 22:19:34.952592 138714412242752 metric_logger.py:196] completed step: 4, seconds: 0.142, TFLOP/s/device: 95.490, Tokens/s/device: 14393.242, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10661.594
I0418 22:19:34.965032 138714412242752 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.576, Tokens/s/device: 13953.995, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7245.978
I0418 22:19:35.108520 138714412242752 metric_logger.py:196] completed step: 6, seconds: 21.929, TFLOP/s/device: 0.620, Tokens/s/device: 93.393, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.156
I0418 22:19:35.255856 138714412242752 metric_logger.py:196] completed step: 7, seconds: 0.009, TFLOP/s/device: 1597.170, Tokens/s/device: 240742.918, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.465
I0418 22:19:35.403628 138714412242752 metric_logger.py:196] completed step: 8, seconds: 0.145, TFLOP/s/device: 93.846, Tokens/s/device: 14145.503, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.256
I0418 22:19:35.552050 138714412242752 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0418 22:19:35.552803 138714412242752 checkpointing.py:798] Waited 0.0007655620574951172 seconds for step 9 to finish before starting checkpointing.
I0418 22:19:35.554738 138714412242752 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 22:19:53.030175 138585126717184 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.249782s (commit=36.304653s, array_metadata_write=1.945129s)
I0418 22:19:53.031424 138587171628800 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 38.173 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.38034749031067 s) (per-host)
I0418 22:19:53.031552 138587171628800 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.092385s.
I0418 22:19:53.695403 138587171628800 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776550753414070078, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0/_CHECKPOINT_METADATA
I0418 22:19:54.393796 138584048371456 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776550753414070078, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0/_CHECKPOINT_METADATA
I0418 22:19:54.648843 138587171628800 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 22:19:54.795013 138587171628800 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000433s.
I0418 22:19:55.735373 138587171628800 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.918485s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0/items
I0418 22:19:55.737081 138587171628800 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0/items
I0418 22:19:56.256799 138587171628800 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0
I0418 22:19:57.588304 138587171628800 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0`.
I0418 22:19:57.589036 138587171628800 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 46.66 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/0
I0418 22:19:57.590465 138587171628800 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 45.651297s.
I0418 22:19:57.590598 138585118324480 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 22:19:57.590662 138585118324480 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 22:19:57.590718 138585118324480 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0418 22:19:57.593266 138585118324480 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0418 22:19:57.593442 138714412242752 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0418 22:19:57.593580 138714412242752 checkpoint_manager.py:1452] Waiting for previous save to complete took 22.038841 seconds. If this number is high, consider checkpointing less frequently.
I0418 22:19:57.595437 138714412242752 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0418 22:19:57.597637 138714412242752 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/9.
I0418 22:19:58.300837 138714412242752 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0418 22:19:58.300930 138714412242752 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0418 22:19:58.327826 138714412242752 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028430s
I0418 22:19:58.328584 138714412242752 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 41.180 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03745913505554199 s) (per-host)
I0418 22:19:58.328649 138714412242752 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.037540s (batch_requests_ready=0.002877s, total_serialization_initiated=0.033954s, others=0.000709s)
I0418 22:19:58.328729 138714412242752 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.041522s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.041505)
I0418 22:19:58.329475 138714412242752 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.73 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/9.
I0418 22:19:58.329772 138585118324480 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 22:39:58.329736
I0418 22:19:58.343015 138714412242752 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 22:19:58.343317 138585656284928 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 22:19:58.343485 138714412242752 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776550775.5547063, 'wait_for_prev_duration_secs': 22.038841247558594, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776550797.5954752, 'checkpointer_blocking_duration_secs': 0.7344725131988525, 'get_old_steps_start_time': 1776550798.3299732, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1776550775.5530448, 'checkpoint_manager_blocking_duration_secs': 22.790404558181763}
I0418 22:19:58.343611 138714412242752 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0418 22:19:58.343656 138714412242752 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 22:19:58.536559 138587171628800 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/9
I0418 22:20:00.250041 138586238215936 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/9/items
I0418 22:20:03.507856 138585673070336 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/9/items/array_metadatas/process_0
I0418 22:20:43.415012 138585664677632 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.059826s (commit=40.543269s, array_metadata_write=1.516557s)
I0418 22:20:43.416169 138585118324480 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 35.005 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.125014781951904 s) (per-host)
I0418 22:20:43.416286 138585118324480 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.086370s.
I0418 22:20:45.447846 138585118324480 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 22:20:45.640503 138585118324480 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000440s.
I0418 22:20:46.391409 138585118324480 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.754194s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/9/items
I0418 22:20:46.393043 138585118324480 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/9/items
I0418 22:20:47.339536 138585118324480 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/9
I0418 22:20:48.706419 138585118324480 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/9`.
I0418 22:20:48.707303 138585118324480 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.11 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_01_base/checkpoints/9
I0418 22:20:48.708641 138585118324480 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.378721s.
I0418 22:20:48.708790 138585656284928 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 22:20:48.708888 138585656284928 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 22:20:48.708965 138585656284928 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0418 22:20:48.711794 138585656284928 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0418 22:20:48.711997 138714412242752 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0418 22:20:48.712161 138714412242752 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 22:20:48.713138 138714412242752 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.397, Tokens/s/device: 13927.046, total_weights: 65536, loss: 8.175, lm_loss: 8.175, perplexity: 3551.438
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Sat Apr 18 22:20:57 UTC 2026
EXIT_CODE=0