MaxView

XPK Start: Tue Apr 21 17:37:58 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-21 17:38:22.895980: 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)
I0421 17:38:23.076570 139845849388864 max_utils.py:273] Attempting to initialize the jax distributed system...
I0421 17:38:32.118098 139845849388864 distributed.py:149] Starting JAX distributed service on [::]:8482
I0421 17:38:32.120401 139845849388864 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-false-jz5d8-slice-job-0-0.mt-10-shardy-false-jz5d8:8482
I0421 17:38:33.428409 139845849388864 max_utils.py:284] Jax distributed system initialized!
I0421 17:38:39.481329 139845849388864 max_utils.py:800] System Information: Jax Version: 0.9.2
I0421 17:38:39.481432 139845849388864 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0421 17:38:39.481473 139845849388864 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
I0421 17:38:39.481508 139845849388864 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0421 17:38:40.177293 139845849388864 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0421 17:38:40.177878 139845849388864 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0421 17:38:40.178068 139845849388864 checkpointing.py:688] Setting up checkpoint logger...
I0421 17:38:40.178120 139845849388864 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0421 17:38:40.178163 139845849388864 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0421 17:38:40.178504 139845849388864 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 0x7f2fd6b1d370>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0421 17:38:43.106143 139845849388864 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0421 17:38:43.106335 139845849388864 checkpoint_manager.py:708] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f1c4814f8f0>}, handler_registry=None
I0421 17:38:43.106577 139845849388864 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f1c4814f8f0>` 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`.
I0421 17:38:43.106625 139845849388864 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f2ea284efc0>` 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`.
I0421 17:38:43.106660 139845849388864 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 0x7f1c4814f8f0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f1c4814f8f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f2ea284efc0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f2ea284efc0>}).
I0421 17:38:43.106984 139845849388864 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0421 17:38:43.107055 139845849388864 async_checkpointer.py:192] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f1c302782c0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0421 17:38:44.245725 139845849388864 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900_10_shardy_false/checkpoints
I0421 17:38:44.686198 139845849388864 checkpoint_manager.py:929] [process=3][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/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f2ea2dd9310>
I0421 17:38:44.686369 139845849388864 checkpointing.py:302] Checkpoint manager created!
I0421 17:38:45.624137 139845849388864 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0421 17:38:45.624249 139845849388864 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 17:38:46.008084 139845849388864 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 17:38:46.008182 139845849388864 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 17:38:46.024470 139845849388864 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 17:38:46.024532 139845849388864 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 17:38:46.048117 139845849388864 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 17:38:46.048184 139845849388864 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 17:38:46.064405 139845849388864 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 17:38:46.064468 139845849388864 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 17:38:46.080668 139845849388864 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 17:38:46.080732 139845849388864 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 17:38:46.105158 139845849388864 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0421 17:38:46.105227 139845849388864 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 17:38:46.126066 139845849388864 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0421 17:38:46.126129 139845849388864 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0421 17:38:46.342340 139845849388864 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0421 17:38:46.342447 139845849388864 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0421 17:38:47.726382 139845849388864 maxtext_utils.py:1821]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0421 17:38:47.726516 139845849388864 maxtext_utils.py:1821]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0421 17:38:47.726571 139845849388864 maxtext_utils.py:1821]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0421 17:38:47.726629 139845849388864 maxtext_utils.py:1821]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   P('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0421 17:38:47.726684 139845849388864 maxtext_utils.py:1821]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0421 17:38:47.726724 139845849388864 maxtext_utils.py:1821]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0421 17:38:47.726776 139845849388864 maxtext_utils.py:1821]  params/params/decoder/layers/self_attention/key/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   P('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0421 17:38:47.726831 139845849388864 maxtext_utils.py:1821]  params/params/decoder/layers/self_attention/out/kernel
    Shape:     float32[16,16,128,2048]
    Logical:   P('heads', 'layers', 'kv', 'embed')
    Physical:  (None, None, None, 'fsdp')
I0421 17:38:47.726872 139845849388864 maxtext_utils.py:1821]  params/params/decoder/layers/self_attention/query/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   P('embed', 'layers', 'q_heads', 'kv')
    Physical:  ('fsdp', None, None, None)
I0421 17:38:47.726909 139845849388864 maxtext_utils.py:1821]  params/params/decoder/layers/self_attention/value/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   P('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0421 17:38:47.726970 139845849388864 maxtext_utils.py:1821]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)
I0421 17:38:47.727022 139845849388864 maxtext_utils.py:1821]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]

    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')
I0421 17:38:48.225167 139845849388864 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 17:38:48.225263 139845849388864 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 17:38:48.240689 139845849388864 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 17:38:48.240750 139845849388864 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 17:38:59.173727 139845849388864 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.
I0421 17:38:59.174536 139845849388864 metric_logger.py:301] number parameters: 1.104 billion
I0421 17:39:10.368908 139845849388864 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0421 17:39:10.522541 139845849388864 checkpointing.py:798] Waited 0.15361332893371582 seconds for step 0 to finish before starting checkpointing.
I0421 17:39:10.524996 139845849388864 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 17:39:10.526666 139845849388864 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 0
I0421 17:39:10.528007 139845849388864 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900_10_shardy_false/checkpoints/0.
I0421 17:39:10.861776 139845849388864 signaling_client.py:364] Using JaxDistributedSignalingClient
I0421 17:39:10.862720 139845849388864 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 17:39:10.862778 139845849388864 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
I0421 17:39:11.133255 139845849388864 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.271552s
I0421 17:39:11.133449 139845849388864 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 5.558 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2775418758392334 s) (per-host)
I0421 17:39:11.133502 139845849388864 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.277605s (batch_requests_ready=0.002350s, total_serialization_initiated=0.275183s, others=0.000072s)
I0421 17:39:11.133594 139845849388864 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.281829s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.281811)
I0421 17:39:11.134471 139845849388864 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.61 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900_10_shardy_false/checkpoints/0.
I0421 17:39:11.134763 139720735573760 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 17:59:11.134729
I0421 17:39:11.151894 139845849388864 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 17:39:11.152160 139719154779904 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save.
I0421 17:39:11.152284 139845849388864 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776793150.5249753, 'wait_for_prev_duration_secs': 7.2479248046875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776793150.526713, 'checkpointer_blocking_duration_secs': 0.6081540584564209, 'get_old_steps_start_time': 1776793151.1348915, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776793150.5231102, 'checkpoint_manager_blocking_duration_secs': 0.6291356086730957}
I0421 17:39:11.152395 139845849388864 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0421 17:39:11.152446 139845849388864 max_utils.py:750] 
Memstats: After params initialized:
I0421 17:39:11.152495 139845849388864 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0))
I0421 17:39:11.152529 139845849388864 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0))
I0421 17:39:11.152553 139845849388864 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0))
I0421 17:39:11.152575 139845849388864 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0))
I0421 17:39:11.464272 139845849388864 metric_logger.py:196] completed step: 0, seconds: 11.194, TFLOP/s/device: 1.214, Tokens/s/device: 182.951, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0421 17:39:11.649915 139845849388864 metric_logger.py:196] completed step: 1, seconds: 1.094, TFLOP/s/device: 12.421, Tokens/s/device: 1872.288, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0421 17:39:12.078447 139845849388864 metric_logger.py:196] completed step: 2, seconds: 0.039, TFLOP/s/device: 348.433, Tokens/s/device: 52519.554, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129
I0421 17:39:12.225895 139845849388864 metric_logger.py:196] completed step: 3, seconds: 0.428, TFLOP/s/device: 31.713, Tokens/s/device: 4780.155, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.203
I0421 17:39:12.521350 139845849388864 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.897, Tokens/s/device: 13399.458, total_weights: 65536, loss: 9.277, lm_loss: 9.277, perplexity: 10694.614
I0421 17:39:12.527004 139845849388864 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.338, Tokens/s/device: 13918.149, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.717
I0421 17:39:14.068320    2750 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0421 17:39:16.533256 139719163172608 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900_10_shardy_false/checkpoints/0/items/array_metadatas/process_3
I0421 17:39:34.179647 139845849388864 metric_logger.py:196] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.927, Tokens/s/device: 6922.614, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.759
I0421 17:39:34.326999 139845849388864 metric_logger.py:196] completed step: 7, seconds: 21.506, TFLOP/s/device: 0.632, Tokens/s/device: 95.229, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4376.905
I0421 17:39:34.474349 139845849388864 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.389, Tokens/s/device: 13473.684, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3846.406
I0421 17:39:34.620855 139845849388864 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0421 17:39:34.621456 139845849388864 checkpointing.py:798] Waited 0.0006227493286132812 seconds for step 9 to finish before starting checkpointing.
I0421 17:39:34.623647 139845849388864 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 17:39:47.804444 139720735573760 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 42.751 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.94849967956543 s) (per-host)
I0421 17:39:47.804551 139720735573760 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.669713s.
I0421 17:39:57.014697 139720735573760 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 45.879843s.
I0421 17:39:57.014973 139719154779904 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 17:39:57.015044 139719154779904 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save.
I0421 17:39:57.015095 139719154779904 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0421 17:39:57.017775 139719154779904 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0421 17:39:57.017956 139845849388864 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0421 17:39:57.018112 139845849388864 checkpoint_manager.py:1452] Waiting for previous save to complete took 22.394462 seconds. If this number is high, consider checkpointing less frequently.
I0421 17:39:57.019952 139845849388864 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 9
I0421 17:39:57.022075 139845849388864 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900_10_shardy_false/checkpoints/9.
I0421 17:39:57.347414 139845849388864 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 17:39:57.347509 139845849388864 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
I0421 17:39:57.383607 139845849388864 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037384s
I0421 17:39:57.383790 139845849388864 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 37.703 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04091334342956543 s) (per-host)
I0421 17:39:57.383843 139845849388864 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.040983s (batch_requests_ready=0.001885s, total_serialization_initiated=0.039020s, others=0.000078s)
I0421 17:39:57.383952 139845849388864 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.045121s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.045105)
I0421 17:39:57.392160 139845849388864 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.37 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900_10_shardy_false/checkpoints/9.
I0421 17:39:57.392467 139719154779904 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 17:59:57.392433
I0421 17:39:57.394392 139845849388864 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 17:39:57.394666 139692670449408 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save.
I0421 17:39:57.394805 139845849388864 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776793174.623616, 'wait_for_prev_duration_secs': 22.39446210861206, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776793197.0199957, 'checkpointer_blocking_duration_secs': 0.372617244720459, 'get_old_steps_start_time': 1776793197.3926358, 'get_old_steps_duration_secs': 2.7894973754882812e-05, 'checkpoint_manager_blocking_start_time': 1776793174.6216984, 'checkpoint_manager_blocking_duration_secs': 22.773070335388184}
I0421 17:39:57.394985 139845849388864 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0421 17:39:57.395034 139845849388864 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 17:40:03.657580 139719163172608 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_161900_10_shardy_false/checkpoints/9/items/array_metadatas/process_3
I0421 17:40:39.538713 139719154779904 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 37.435 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.19581151008606 s) (per-host)
I0421 17:40:39.538839 139719154779904 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.146257s.
I0421 17:40:48.271615 139719154779904 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 50.879019s.
I0421 17:40:48.271921 139692670449408 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 17:40:48.272079 139692670449408 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save.
I0421 17:40:48.272140 139692670449408 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0421 17:40:48.274838 139692670449408 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0421 17:40:48.275020 139845849388864 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0421 17:40:48.275169 139845849388864 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 17:40:48.276007 139845849388864 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 92.098, Tokens/s/device: 13882.016, total_weights: 65536, loss: 8.179, lm_loss: 8.179, perplexity: 3564.635
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Tue Apr 21 17:40:59 UTC 2026
EXIT_CODE=0