XPK Start: Fri Apr 24 21:24:53 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
`rope_parameters`'s factor field must be a float >= 1, got 40
`rope_parameters`'s beta_fast field must be a float, got 32
`rope_parameters`'s beta_slow field must be a float, got 1
DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config.
2026-04-24 21:25:18.263525: 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)
I0424 21:25:18.476584 133270356543296 max_utils.py:273] Attempting to initialize the jax distributed system...
I0424 21:25:27.518981 133270356543296 distributed.py:149] Starting JAX distributed service on [::]:8482
I0424 21:25:27.521283 133270356543296 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-true-rs-j6llo-slice-job-0-0.mt-16-zarr3-true-rs-j6llo:8482
I0424 21:25:28.417994 133270356543296 max_utils.py:284] Jax distributed system initialized!
I0424 21:25:34.714994 133270356543296 max_utils.py:800] System Information: Jax Version: 0.9.2
I0424 21:25:34.715112 133270356543296 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0424 21:25:34.715153 133270356543296 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
I0424 21:25:34.715188 133270356543296 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0424 21:25:35.406289 133270356543296 maxtext_utils.py:1565] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0424 21:25:35.406577 133270356543296 checkpointing.py:677] Setting up checkpoint logger...
I0424 21:25:35.406633 133270356543296 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0424 21:25:35.406677 133270356543296 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0424 21:25:35.407018 133270356543296 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 0x7934ce5141d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0424 21:25:36.694725 133270356543296 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0424 21:25:36.694962 133270356543296 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79206809bf20>}, handler_registry=None
I0424 21:25:36.695216 133270356543296 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79206809bf20>` 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`.
I0424 21:25:36.695266 133270356543296 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79206809d7c0>` 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`.
I0424 21:25:36.695302 133270356543296 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 0x79206809bf20>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79206809bf20>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79206809d7c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79206809d7c0>}).
I0424 21:25:36.695633 133270356543296 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0424 21:25:36.695705 133270356543296 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79205066f2e0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0424 21:25:39.004760 133148667008768 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777065674915252727, 'commit_timestamp_nsecs': 1777065717072345062, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA
I0424 21:25:39.245682 133148658616064 checkpoint_manager.py:1871] Missing metrics for step 5
I0424 21:25:39.273819 133270356543296 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints
I0424 21:25:39.854148 133270356543296 checkpoint_manager.py:929] [process=6][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=5), 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_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79206809b590>
I0424 21:25:39.854325 133270356543296 checkpointing.py:301] Checkpoint manager created!
I0424 21:25:40.793987 133270356543296 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0424 21:25:40.794114 133270356543296 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 21:25:41.173640 133270356543296 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0424 21:25:41.173733 133270356543296 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 21:25:41.190090 133270356543296 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0424 21:25:41.190165 133270356543296 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 21:25:41.213599 133270356543296 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0424 21:25:41.213666 133270356543296 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 21:25:41.230124 133270356543296 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0424 21:25:41.230184 133270356543296 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 21:25:41.246874 133270356543296 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0424 21:25:41.246941 133270356543296 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 21:25:41.271790 133270356543296 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0424 21:25:41.271863 133270356543296 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 21:25:41.292885 133270356543296 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0424 21:25:41.292956 133270356543296 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0424 21:25:41.506879 133270356543296 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0424 21:25:41.506992 133270356543296 checkpointing.py:581] restoring from this run's directory step 9
I0424 21:25:41.746382 133270356543296 event_tracking.py:70] [process=6] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/9.
I0424 21:25:41.997120 133270356543296 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/9.
I0424 21:25:41.997564 133270356543296 event_tracking.py:125] [process=6] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/9.
I0424 21:25:43.424077 1458 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0424 21:25:50.092485 133270356543296 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 236.774 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.6741907596588135 s) (per-host)
I0424 21:25:50.098066 133270356543296 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 222.624 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.108834981918335 s) (per-host)
I0424 21:25:50.100273 133270356543296 event_tracking.py:138] [process=6] [sync] Finished load in 8.35 seconds @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/9
I0424 21:25:50.100378 133270356543296 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1777065941.7462811, 'checkpointer_duration_secs': 8.354068517684937, 'checkpoint_manager_start_time': 1777065941.5072494, 'checkpoint_manager_duration_secs': 8.59310245513916}
fsdp: 32
I0424 21:25:50.521526 133270356543296 maxtext_utils.py:1668] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0424 21:25:50.521651 133270356543296 maxtext_utils.py:1668] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0424 21:25:50.521703 133270356543296 maxtext_utils.py:1668] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0424 21:25:50.521761 133270356543296 maxtext_utils.py:1668] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0424 21:25:50.521812 133270356543296 maxtext_utils.py:1668] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0424 21:25:50.521850 133270356543296 maxtext_utils.py:1668] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0424 21:25:50.521900 133270356543296 maxtext_utils.py:1668] 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)
I0424 21:25:50.521950 133270356543296 maxtext_utils.py:1668] 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')
I0424 21:25:50.521988 133270356543296 maxtext_utils.py:1668] 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)
I0424 21:25:50.522022 133270356543296 maxtext_utils.py:1668] 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)
I0424 21:25:50.522068 133270356543296 maxtext_utils.py:1668] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0424 21:25:50.522128 133270356543296 maxtext_utils.py:1668] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0424 21:25:51.003387 133270356543296 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0424 21:25:51.003480 133270356543296 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0424 21:25:51.018946 133270356543296 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0424 21:25:51.019009 133270356543296 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0424 21:26:02.075995 133270356543296 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.
I0424 21:26:02.076741 133270356543296 metric_logger.py:301] number parameters: 1.104 billion
I0424 21:26:13.605762 133270356543296 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0424 21:26:13.762184 133270356543296 checkpointing.py:776] Waited 0.15640783309936523 seconds for step 10 to finish before starting checkpointing.
I0424 21:26:13.764859 133270356543296 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0424 21:26:13.766712 133270356543296 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10
I0424 21:26:13.768199 133270356543296 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/10.
I0424 21:26:14.113303 133270356543296 signaling_client.py:364] Using JaxDistributedSignalingClient
I0424 21:26:14.114268 133270356543296 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0424 21:26:14.114326 133270356543296 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
I0424 21:26:14.386074 133270356543296 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.272849s
I0424 21:26:14.386259 133270356543296 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.575 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.276719331741333 s) (per-host)
I0424 21:26:14.386326 133270356543296 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.276792s (batch_requests_ready=0.002147s, total_serialization_initiated=0.274564s, others=0.000082s)
I0424 21:26:14.386433 133270356543296 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.280807s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.280789)
I0424 21:26:14.387174 133270356543296 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.62 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/10.
I0424 21:26:14.387517 133148063028992 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 21:46:14.387480
I0424 21:26:14.400911 133270356543296 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0424 21:26:14.401167 133144307214080 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0424 21:26:14.401290 133270356543296 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777065973.7648404, 'wait_for_prev_duration_secs': 6.318092346191406e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777065973.7667491, 'checkpointer_blocking_duration_secs': 0.6208720207214355, 'get_old_steps_start_time': 1777065974.387643, 'get_old_steps_duration_secs': 3.337860107421875e-05, 'checkpoint_manager_blocking_start_time': 1777065973.7626965, 'checkpoint_manager_blocking_duration_secs': 0.6385583877563477}
I0424 21:26:14.401396 133270356543296 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0424 21:26:14.401446 133270356543296 max_utils.py:750]
Memstats: After params initialized:
I0424 21:26:14.401493 133270356543296 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0424 21:26:14.401531 133270356543296 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0424 21:26:14.401558 133270356543296 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0424 21:26:14.401582 133270356543296 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0424 21:26:14.713275 133270356543296 metric_logger.py:196] completed step: 10, seconds: 11.529, TFLOP/s/device: 1.179, Tokens/s/device: 177.641, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0424 21:26:14.876304 133270356543296 metric_logger.py:196] completed step: 11, seconds: 1.106, TFLOP/s/device: 12.284, Tokens/s/device: 1851.649, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0424 21:26:15.024346 133270356543296 metric_logger.py:196] completed step: 12, seconds: 0.016, TFLOP/s/device: 859.238, Tokens/s/device: 129513.691, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0424 21:26:15.171758 133270356543296 metric_logger.py:196] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.717, Tokens/s/device: 13372.423, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0424 21:26:15.176136 133270356543296 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0424 21:26:15.465739 133270356543296 checkpointing.py:776] Waited 0.2895677089691162 seconds for step 15 to finish before starting checkpointing.
I0424 21:26:15.468811 133270356543296 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0424 21:26:19.419168 133144315606784 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/10/items/array_metadatas/process_6
I0424 21:26:55.525207 133148063028992 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.140 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.4156334400177 s) (per-host)
I0424 21:26:55.525316 133148063028992 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.137158s.
I0424 21:27:04.484281 133148063028992 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.096105s.
I0424 21:27:04.484499 133144307214080 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0424 21:27:04.484555 133144307214080 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0424 21:27:04.484601 133144307214080 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0424 21:27:04.486157 133144307214080 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0424 21:27:04.486330 133270356543296 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0424 21:27:04.486464 133270356543296 checkpoint_manager.py:1452] Waiting for previous save to complete took 49.017654 seconds. If this number is high, consider checkpointing less frequently.
I0424 21:27:04.488416 133270356543296 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15
I0424 21:27:04.491122 133270356543296 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/15.
I0424 21:27:04.787243 133270356543296 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0424 21:27:04.787336 133270356543296 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
I0424 21:27:04.814435 133270356543296 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028411s
I0424 21:27:04.814610 133270356543296 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 48.468 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03182649612426758 s) (per-host)
I0424 21:27:04.814663 133270356543296 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031890s (batch_requests_ready=0.001758s, total_serialization_initiated=0.030061s, others=0.000071s)
I0424 21:27:04.814756 133270356543296 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036745s (all_items=0.000013s, per_item={'items': '0.00001264'}, temp_paths=0.036733)
I0424 21:27:04.815455 133270356543296 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.33 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/15.
I0424 21:27:04.815720 133144307214080 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 21:47:04.815695
I0424 21:27:04.825741 133270356543296 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0424 21:27:04.825989 133144315606784 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0424 21:27:04.826161 133270356543296 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777065975.4687808, 'wait_for_prev_duration_secs': 49.01765418052673, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777066024.488455, 'checkpointer_blocking_duration_secs': 0.32735705375671387, 'get_old_steps_start_time': 1777066024.815831, 'get_old_steps_duration_secs': 3.147125244140625e-05, 'checkpoint_manager_blocking_start_time': 1777065975.4663603, 'checkpoint_manager_blocking_duration_secs': 49.3597674369812}
I0424 21:27:04.826269 133270356543296 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0424 21:27:04.827125 133270356543296 metric_logger.py:196] completed step: 14, seconds: 0.148, TFLOP/s/device: 91.878, Tokens/s/device: 13848.786, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0424 21:27:04.833164 133270356543296 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.204, Tokens/s/device: 13897.937, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0424 21:27:09.918348 133142694319872 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/15/items/array_metadatas/process_6
I0424 21:27:25.757219 133270356543296 metric_logger.py:196] completed step: 16, seconds: 49.656, TFLOP/s/device: 0.274, Tokens/s/device: 41.244, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0424 21:27:25.904831 133270356543296 metric_logger.py:196] completed step: 17, seconds: 20.778, TFLOP/s/device: 0.654, Tokens/s/device: 98.564, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0424 21:27:26.051792 133270356543296 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.450, Tokens/s/device: 13482.909, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0424 21:27:26.198483 133270356543296 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0424 21:27:26.199196 133270356543296 checkpointing.py:776] Waited 0.0007348060607910156 seconds for step 19 to finish before starting checkpointing.
I0424 21:27:26.201260 133270356543296 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0424 21:27:47.214593 133144307214080 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.227 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.43178057670593 s) (per-host)
I0424 21:27:47.214699 133144307214080 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.398908s.
I0424 21:27:53.979628 133144307214080 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.163821s.
I0424 21:27:53.979854 133144315606784 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0424 21:27:53.979910 133144315606784 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0424 21:27:53.979954 133144315606784 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0424 21:27:53.981574 133144315606784 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0424 21:27:53.981695 133270356543296 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0424 21:27:53.981756 133270356543296 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.780505 seconds. If this number is high, consider checkpointing less frequently.
I0424 21:27:53.983377 133270356543296 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19
I0424 21:27:53.985662 133270356543296 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/19.
I0424 21:27:54.285901 133270356543296 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0424 21:27:54.285991 133270356543296 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
I0424 21:27:54.311383 133270356543296 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026326s
I0424 21:27:54.311556 133270356543296 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 51.814 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.02977156639099121 s) (per-host)
I0424 21:27:54.311613 133270356543296 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.029839s (batch_requests_ready=0.001641s, total_serialization_initiated=0.028126s, others=0.000073s)
I0424 21:27:54.311709 133270356543296 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034149s (all_items=0.000010s, per_item={'items': '0.00001049'}, temp_paths=0.034139)
I0424 21:27:54.312340 133270356543296 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.33 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/19.
I0424 21:27:54.312657 133141582362368 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 21:47:54.312616
I0424 21:27:54.316787 133270356543296 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0424 21:27:54.317033 133144315606784 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0424 21:27:54.317151 133270356543296 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777066046.2012303, 'wait_for_prev_duration_secs': 27.78050470352173, 'time_between_consecutive_saves_sec': 21.71503210067749, 'checkpointer_blocking_start_time': 1777066073.983415, 'checkpointer_blocking_duration_secs': 0.32939600944519043, 'get_old_steps_start_time': 1777066074.3128245, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1777066046.1994257, 'checkpoint_manager_blocking_duration_secs': 28.117695808410645}
I0424 21:27:54.317307 133270356543296 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0424 21:27:54.317350 133270356543296 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0424 21:27:59.421738 133142694319872 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260424_200834/linen_xpk_test_pipeline_scan_nnx_20260424_200834_16_zarr3_true/checkpoints/19/items/array_metadatas/process_6
I0424 21:28:35.634249 133141582362368 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.198 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.35242438316345 s) (per-host)
I0424 21:28:35.634369 133141582362368 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.321589s.
I0424 21:28:41.807535 133141582362368 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 47.494736s.
I0424 21:28:41.807795 133144315606784 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0424 21:28:41.807916 133144315606784 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0424 21:28:41.807958 133144315606784 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0424 21:28:41.810198 133144315606784 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0424 21:28:41.810392 133270356543296 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0424 21:28:41.810570 133270356543296 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0424 21:28:41.811468 133270356543296 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.126, Tokens/s/device: 13886.252, total_weights: 65536, loss: 5.391, lm_loss: 5.391, perplexity: 219.349
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Fri Apr 24 21:28:53 UTC 2026
EXIT_CODE=0