MaxView

← Back to run

Log Summary

XPK Start: Mon Apr 20 13:52:38 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-20 13:53:02.265180: 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)
I0420 13:53:02.445228 136818241464128 max_utils.py:273] Attempting to initialize the jax distributed system...
I0420 13:53:11.488345 136818241464128 distributed.py:149] Starting JAX distributed service on [::]:8482
I0420 13:53:11.490942 136818241464128 distributed.py:172] Connecting to JAX distributed service on mt-02-synthetic-4ekgx-slice-job-0-0.mt-02-synthetic-4ekgx:8482
I0420 13:53:12.577987 136818241464128 max_utils.py:284] Jax distributed system initialized!
I0420 13:53:17.650381 136818241464128 max_utils.py:800] System Information: Jax Version: 0.9.2
I0420 13:53:17.650483 136818241464128 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0420 13:53:17.650524 136818241464128 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
I0420 13:53:17.650562 136818241464128 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0420 13:53:18.697434 136818241464128 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 13:53:18.813180 136818241464128 checkpointing.py:688] Setting up checkpoint logger...
I0420 13:53:18.813292 136818241464128 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0420 13:53:18.813337 136818241464128 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0420 13:53:18.813532 136818241464128 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 0x7c6ef87ae600>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0420 13:53:21.521624 136818241464128 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0420 13:53:21.521810 136818241464128 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 0x7c6db81fa7b0>}, handler_registry=None
I0420 13:53:21.522043 136818241464128 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c6db81fa7b0>` 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`.
I0420 13:53:21.522101 136818241464128 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c6db736c170>` 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`.
I0420 13:53:21.522139 136818241464128 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 0x7c6db81fa7b0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c6db81fa7b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c6db736c170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c6db736c170>}).
I0420 13:53:21.522458 136818241464128 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0420 13:53:21.522528 136818241464128 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c5b345b05e0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0420 13:53:22.657101 136818241464128 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints
I0420 13:53:23.115304 136818241464128 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_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c6dac7c89e0>
I0420 13:53:23.115474 136818241464128 checkpointing.py:302] Checkpoint manager created!
I0420 13:53:24.064942 136818241464128 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0420 13:53:24.065054 136818241464128 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0420 13:53:25.435851 136818241464128 maxtext_utils.py:1836]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0420 13:53:25.435963 136818241464128 maxtext_utils.py:1836]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0420 13:53:25.436014 136818241464128 maxtext_utils.py:1836]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0420 13:53:25.436056 136818241464128 maxtext_utils.py:1836]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0420 13:53:25.436110 136818241464128 maxtext_utils.py:1836]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0420 13:53:25.436148 136818241464128 maxtext_utils.py:1836]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0420 13:53:25.436185 136818241464128 maxtext_utils.py:1836]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 13:53:25.436220 136818241464128 maxtext_utils.py:1836]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0420 13:53:25.436253 136818241464128 maxtext_utils.py:1836]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 13:53:25.436286 136818241464128 maxtext_utils.py:1836]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 13:53:25.436318 136818241464128 maxtext_utils.py:1836]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)

I0420 13:53:25.436349 136818241464128 maxtext_utils.py:1836]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')
I0420 13:53:25.838961 136818241464128 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 13:53:25.839053 136818241464128 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 13:53:25.845502 136818241464128 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 13:53:25.845559 136818241464128 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 13:53:25.862022 136818241464128 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 13:53:25.862107 136818241464128 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 13:53:25.877776 136818241464128 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 13:53:25.877840 136818241464128 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 13:53:25.901811 136818241464128 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 13:53:25.901883 136818241464128 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 13:53:25.917671 136818241464128 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 13:53:25.917733 136818241464128 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 13:53:25.933478 136818241464128 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 13:53:25.933539 136818241464128 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 13:53:25.963615 136818241464128 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0420 13:53:25.963687 136818241464128 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 13:53:25.983473 136818241464128 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0420 13:53:25.983535 136818241464128 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0420 13:53:37.102448 136818241464128 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.
I0420 13:53:37.111571 136818241464128 metric_logger.py:301] number parameters: 1.104 billion
I0420 13:53:48.976956 136818241464128 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0420 13:53:49.128952 136818241464128 checkpointing.py:798] Waited 0.1519789695739746 seconds for step 0 to finish before starting checkpointing.
I0420 13:53:49.131260 136818241464128 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 13:53:49.133068 136818241464128 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0420 13:53:49.134485 136818241464128 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0.
I0420 13:53:49.453976 136818241464128 signaling_client.py:364] Using JaxDistributedSignalingClient
I0420 13:53:49.473297 136818241464128 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0420 13:53:49.473367 136818241464128 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
I0420 13:53:49.748423 136818241464128 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.276552s
I0420 13:53:49.749297 136818241464128 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.362 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2877023220062256 s) (per-host)
I0420 13:53:49.749366 136818241464128 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.287793s (batch_requests_ready=0.003113s, total_serialization_initiated=0.283880s, others=0.000800s)
I0420 13:53:49.749443 136818241464128 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.291849s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.291829)
I0420 13:53:49.750201 136818241464128 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.62 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0.
I0420 13:53:49.750481 136689710679808 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 14:13:49.750448
I0420 13:53:49.752283 136689702287104 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0
I0420 13:53:49.759338 136818241464128 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 13:53:49.759594 136685933754112 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0420 13:53:49.759716 136818241464128 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776693229.1312416, 'wait_for_prev_duration_secs': 6.151199340820312e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776693229.1331189, 'checkpointer_blocking_duration_secs': 0.6175062656402588, 'get_old_steps_start_time': 1776693229.75065, 'get_old_steps_duration_secs': 2.7894973754882812e-05, 'checkpoint_manager_blocking_start_time': 1776693229.129371, 'checkpoint_manager_blocking_duration_secs': 0.6303071975708008}
I0420 13:53:49.759834 136818241464128 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0420 13:53:49.759912 136818241464128 max_utils.py:750] 
Memstats: After params initialized:
I0420 13:53:49.759962 136818241464128 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0420 13:53:49.759996 136818241464128 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0420 13:53:49.760023 136818241464128 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0420 13:53:49.760048 136818241464128 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0420 13:53:50.111511 136818241464128 metric_logger.py:196] completed step: 0, seconds: 11.551, TFLOP/s/device: 1.176, Tokens/s/device: 177.297, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469
I0420 13:53:50.114706 136818241464128 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/tensorboard/'
I0420 13:53:50.549440 136818241464128 metric_logger.py:196] completed step: 1, seconds: 1.131, TFLOP/s/device: 12.016, Tokens/s/device: 1811.202, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469
I0420 13:53:50.691236 136818241464128 metric_logger.py:196] completed step: 2, seconds: 0.439, TFLOP/s/device: 30.951, Tokens/s/device: 4665.297, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133
I0420 13:53:50.838781 136818241464128 metric_logger.py:196] completed step: 3, seconds: 0.013, TFLOP/s/device: 1077.574, Tokens/s/device: 162423.666, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16855.789
I0420 13:53:52.133452 136689213241088 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/items
I0420 13:53:52.325215 136684866332416 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776693231862616715, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776693231862616715, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/_CHECKPOINT_METADATA
I0420 13:53:53.273414    2930 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0420 13:53:55.189724 136685950539520 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_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/items/array_metadatas/process_0
I0420 13:54:13.313364 136818241464128 metric_logger.py:196] completed step: 4, seconds: 0.143, TFLOP/s/device: 94.764, Tokens/s/device: 14283.821, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10661.594
I0420 13:54:13.323631 136818241464128 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.531, Tokens/s/device: 13947.343, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7245.978
I0420 13:54:13.468300 136818241464128 metric_logger.py:196] completed step: 6, seconds: 22.473, TFLOP/s/device: 0.605, Tokens/s/device: 91.132, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.156
I0420 13:54:13.615774 136818241464128 metric_logger.py:196] completed step: 7, seconds: 0.008, TFLOP/s/device: 1720.979, Tokens/s/device: 259404.687, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.465
I0420 13:54:13.763672 136818241464128 metric_logger.py:196] completed step: 8, seconds: 0.146, TFLOP/s/device: 92.992, Tokens/s/device: 14016.741, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.256
I0420 13:54:13.911412 136818241464128 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0420 13:54:13.912057 136818241464128 checkpointing.py:798] Waited 0.0006573200225830078 seconds for step 9 to finish before starting checkpointing.
I0420 13:54:13.914136 136818241464128 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 13:54:31.382347 136685942146816 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.115432s (commit=36.454607s, array_metadata_write=1.660824s)
I0420 13:54:31.383387 136689710679808 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.680 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.92176342010498 s) (per-host)
I0420 13:54:31.383513 136689710679808 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.632921s.
I0420 13:54:32.133352 136689710679808 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776693231862616715, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/_CHECKPOINT_METADATA
I0420 13:54:32.958984 136684866332416 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776693231862616715, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/_CHECKPOINT_METADATA
I0420 13:54:33.139294 136689710679808 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000470s.
I0420 13:54:33.555282 136689710679808 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0420 13:54:34.495181 136689710679808 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.229832s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/items
I0420 13:54:34.496977 136689710679808 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/items
I0420 13:54:35.179314 136689710679808 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0
I0420 13:54:36.659458 136689710679808 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_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0`.
I0420 13:54:36.660295 136689710679808 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 47.53 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0
I0420 13:54:36.661610 136689710679808 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 46.911015s.
I0420 13:54:36.661839 136685933754112 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 13:54:36.661979 136685933754112 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0420 13:54:36.662043 136685933754112 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0420 13:54:36.664486 136685933754112 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0420 13:54:36.664618 136818241464128 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.
W0420 13:54:36.664710 136818241464128 checkpoint_manager.py:1452] Waiting for previous save to complete took 22.750577 seconds. If this number is high, consider checkpointing less frequently.
I0420 13:54:36.666651 136818241464128 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0420 13:54:36.668824 136818241464128 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9.
I0420 13:54:37.009588 136818241464128 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0420 13:54:37.009682 136818241464128 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
I0420 13:54:37.037626 136818241464128 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.029544s
I0420 13:54:37.038462 136818241464128 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 40.060 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03850698471069336 s) (per-host)
I0420 13:54:37.038528 136818241464128 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.038585s (batch_requests_ready=0.002739s, total_serialization_initiated=0.035110s, others=0.000736s)
I0420 13:54:37.038616 136818241464128 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.042582s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.042564)
I0420 13:54:37.039317 136818241464128 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.37 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9.
I0420 13:54:37.039588 136688005281536 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 14:14:37.039555
I0420 13:54:37.046241 136818241464128 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 13:54:37.046525 136685933754112 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0420 13:54:37.046700 136818241464128 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776693253.9141033, 'wait_for_prev_duration_secs': 22.750576734542847, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776693276.6666887, 'checkpointer_blocking_duration_secs': 0.3730010986328125, 'get_old_steps_start_time': 1776693277.0397139, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776693253.912291, 'checkpoint_manager_blocking_duration_secs': 23.13437247276306}
I0420 13:54:37.046832 136818241464128 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0420 13:54:37.046882 136818241464128 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 13:54:37.253247 136689710679808 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9
I0420 13:54:39.035509 136689194080000 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9/items
I0420 13:54:42.793065 136688676370176 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_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9/items/array_metadatas/process_0
I0420 13:55:23.189349 136688659584768 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.613254s (commit=41.364641s, array_metadata_write=1.248613s)
I0420 13:55:23.190426 136688005281536 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.198 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.19044327735901 s) (per-host)
I0420 13:55:23.190535 136688005281536 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.150870s.
I0420 13:55:24.632795 136688005281536 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000377s.
I0420 13:55:24.853427 136688005281536 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0420 13:55:25.945212 136688005281536 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.939945s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9/items
I0420 13:55:25.946825 136688005281536 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9/items
I0420 13:55:26.479457 136688005281536 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9
I0420 13:55:27.995175 136688005281536 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_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9`.
I0420 13:55:27.996037 136688005281536 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.33 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9
I0420 13:55:27.997712 136688005281536 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.958045s.
I0420 13:55:27.997920 136685933754112 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 13:55:27.998038 136685933754112 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0420 13:55:27.998114 136685933754112 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0420 13:55:28.000306 136685933754112 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0420 13:55:28.000495 136818241464128 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.
I0420 13:55:28.000667 136818241464128 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 13:55:28.001659 136818241464128 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 92.082, Tokens/s/device: 13879.664, 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: Mon Apr 20 13:55:38 UTC 2026
EXIT_CODE=0