XPK Start: Tue Apr 21 15:41:49 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. 2026-04-21 15:42:14.877940: 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 15:42:15.090787 135937944586048 max_utils.py:273] Attempting to initialize the jax distributed system... I0421 15:42:24.132560 135937944586048 distributed.py:149] Starting JAX distributed service on [::]:8482 I0421 15:42:24.134943 135937944586048 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-6lq8l-slice-job-0-0.mt-15-ocdbt-true-sv-6lq8l:8482 I0421 15:42:25.164176 135937944586048 max_utils.py:284] Jax distributed system initialized! I0421 15:42:31.383394 135937944586048 max_utils.py:800] System Information: Jax Version: 0.9.2 I0421 15:42:31.383503 135937944586048 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0421 15:42:31.383546 135937944586048 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 15:42:31.383582 135937944586048 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0421 15:42:32.079398 135937944586048 maxtext_utils.py:1565] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 15:42:32.079699 135937944586048 checkpointing.py:677] Setting up checkpoint logger... I0421 15:42:32.079758 135937944586048 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0421 15:42:32.079804 135937944586048 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0421 15:42:32.080137 135937944586048 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 0x7ba1e3f18230>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0421 15:42:35.337501 135937944586048 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0421 15:42:35.337758 135937944586048 checkpoint_manager.py:708] [process=4][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b8d9c5a7b60>}, handler_registry=None I0421 15:42:35.337996 135937944586048 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b8d9c5a7b60>` 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 15:42:35.338046 135937944586048 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b8d9c5aac00>` 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 15:42:35.338080 135937944586048 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 0x7b8d9c5a7b60>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b8d9c5a7b60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b8d9c5aac00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b8d9c5aac00>}). I0421 15:42:35.338411 135937944586048 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0421 15:42:35.338480 135937944586048 async_checkpointer.py:192] [process=4][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7b8cbc689580> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0421 15:42:36.135041 135937944586048 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints I0421 15:42:36.594069 135937944586048 checkpoint_manager.py:929] [process=4][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_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b8d9c5a97c0> I0421 15:42:36.594242 135937944586048 checkpointing.py:301] Checkpoint manager created! I0421 15:42:37.520430 135937944586048 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0421 15:42:37.520554 135937944586048 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 15:42:37.905599 135937944586048 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 15:42:37.905703 135937944586048 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 15:42:37.922336 135937944586048 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 15:42:37.922397 135937944586048 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 15:42:37.946674 135937944586048 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 15:42:37.946743 135937944586048 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 15:42:37.963457 135937944586048 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 15:42:37.963517 135937944586048 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 15:42:37.980195 135937944586048 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 15:42:37.980258 135937944586048 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 15:42:38.005219 135937944586048 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0421 15:42:38.005291 135937944586048 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 15:42:38.026968 135937944586048 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0421 15:42:38.027040 135937944586048 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0421 15:42:38.243734 135937944586048 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0421 15:42:38.243841 135937944586048 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0421 15:42:39.698130 135937944586048 maxtext_utils.py:1668] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0421 15:42:39.698255 135937944586048 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) I0421 15:42:39.698312 135937944586048 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) I0421 15:42:39.698372 135937944586048 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') I0421 15:42:39.698425 135937944586048 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) I0421 15:42:39.698464 135937944586048 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) I0421 15:42:39.698517 135937944586048 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) I0421 15:42:39.698575 135937944586048 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') I0421 15:42:39.698617 135937944586048 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) I0421 15:42:39.698666 135937944586048 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) I0421 15:42:39.698716 135937944586048 maxtext_utils.py:1668] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0421 15:42:39.699036 135937944586048 maxtext_utils.py:1668] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0421 15:42:40.199463 135937944586048 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 15:42:40.199563 135937944586048 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0421 15:42:40.215329 135937944586048 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 15:42:40.215388 135937944586048 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0421 15:42:50.896063 135937944586048 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 15:42:50.896898 135937944586048 metric_logger.py:301] number parameters: 1.104 billion I0421 15:43:02.120127 135937944586048 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0421 15:43:02.684007 135937944586048 checkpointing.py:776] Waited 0.5638601779937744 seconds for step 0 to finish before starting checkpointing. I0421 15:43:02.686309 135937944586048 checkpoint_manager.py:2009] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 15:43:02.687943 135937944586048 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 0 I0421 15:43:02.689328 135937944586048 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints/0. I0421 15:43:03.017729 135937944586048 signaling_client.py:364] Using JaxDistributedSignalingClient I0421 15:43:03.018751 135937944586048 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 15:43:03.018807 135937944586048 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 15:43:03.297767 135937944586048 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.280125s I0421 15:43:03.297941 135937944586048 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 5.395 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28594326972961426 s) (per-host) I0421 15:43:03.297994 135937944586048 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.286006s (batch_requests_ready=0.002534s, total_serialization_initiated=0.283401s, others=0.000070s) I0421 15:43:03.298081 135937944586048 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.290162s (all_items=0.000018s, per_item={'items': '0.00001812'}, temp_paths=0.290144) I0421 15:43:03.298924 135937944586048 event_tracking.py:125] [process=4] [async] Finished blocking save in 0.61 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints/0. I0421 15:43:03.299199 135809074378496 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 16:03:03.299170 I0421 15:43:03.305242 135937944586048 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0421 15:43:03.305484 135807461484288 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0421 15:43:03.305608 135937944586048 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776786182.686291, 'wait_for_prev_duration_secs': 6.580352783203125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776786182.687983, 'checkpointer_blocking_duration_secs': 0.6113591194152832, 'get_old_steps_start_time': 1776786183.2993674, 'get_old_steps_duration_secs': 3.075599670410156e-05, 'checkpoint_manager_blocking_start_time': 1776786182.684428, 'checkpoint_manager_blocking_duration_secs': 0.6211428642272949} I0421 15:43:03.305722 135937944586048 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0421 15:43:03.305771 135937944586048 max_utils.py:750] Memstats: After params initialized: I0421 15:43:03.305820 135937944586048 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_16(process=4,(0,4,0,0)) I0421 15:43:03.305852 135937944586048 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_17(process=4,(1,4,0,0)) I0421 15:43:03.305880 135937944586048 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_20(process=4,(0,5,0,0)) I0421 15:43:03.305906 135937944586048 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_21(process=4,(1,5,0,0)) I0421 15:43:03.625537 135937944586048 metric_logger.py:196] completed step: 0, seconds: 11.223, TFLOP/s/device: 1.211, Tokens/s/device: 182.481, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0421 15:43:03.806057 135937944586048 metric_logger.py:196] completed step: 1, seconds: 1.504, TFLOP/s/device: 9.035, Tokens/s/device: 1361.827, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0421 15:43:04.259789 135937944586048 metric_logger.py:196] completed step: 2, seconds: 0.033, TFLOP/s/device: 408.206, Tokens/s/device: 61529.217, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0421 15:43:04.407152 135937944586048 metric_logger.py:196] completed step: 3, seconds: 0.454, TFLOP/s/device: 29.920, Tokens/s/device: 4509.901, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0421 15:43:04.411484 135937944586048 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0421 15:43:04.701432 135937944586048 checkpointing.py:776] Waited 0.28991198539733887 seconds for step 5 to finish before starting checkpointing. I0421 15:43:04.704278 135937944586048 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 15:43:06.730053 2778 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0421 15:43:09.213247 135807469876992 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_4 I0421 15:43:39.854013 135809074378496 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 42.875 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.841973066329956 s) (per-host) I0421 15:43:39.854139 135809074378496 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.554835s. I0421 15:43:50.119358 135809074378496 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 46.820038s. I0421 15:43:50.119583 135807461484288 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 15:43:50.119673 135807461484288 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0421 15:43:50.119726 135807461484288 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0421 15:43:50.122697 135807461484288 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0421 15:43:50.122858 135937944586048 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0421 15:43:50.122990 135937944586048 checkpoint_manager.py:1452] Waiting for previous save to complete took 45.418709 seconds. If this number is high, consider checkpointing less frequently. I0421 15:43:50.125018 135937944586048 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 5 I0421 15:43:50.127415 135937944586048 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints/5. I0421 15:43:50.840300 135937944586048 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 15:43:50.840394 135937944586048 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 15:43:50.876195 135937944586048 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037064s I0421 15:43:50.876365 135937944586048 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 38.050 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.040540456771850586 s) (per-host) I0421 15:43:50.876417 135937944586048 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.040602s (batch_requests_ready=0.001815s, total_serialization_initiated=0.038717s, others=0.000070s) I0421 15:43:50.876508 135937944586048 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.045750s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.045734) I0421 15:43:50.877110 135937944586048 event_tracking.py:125] [process=4] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints/5. I0421 15:43:50.877444 135807469876992 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 16:03:50.877406 I0421 15:43:50.884005 135937944586048 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0421 15:43:50.884300 135807461484288 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0421 15:43:50.884463 135937944586048 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776786184.7042503, 'wait_for_prev_duration_secs': 45.41870903968811, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776786230.125062, 'checkpointer_blocking_duration_secs': 0.7525327205657959, 'get_old_steps_start_time': 1776786230.8776155, 'get_old_steps_duration_secs': 4.57763671875e-05, 'checkpoint_manager_blocking_start_time': 1776786184.7020621, 'checkpoint_manager_blocking_duration_secs': 46.182366132736206} I0421 15:43:50.884581 135937944586048 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0421 15:43:50.885405 135937944586048 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.931, Tokens/s/device: 13404.720, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0421 15:43:50.891456 135937944586048 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.296, Tokens/s/device: 13911.815, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0421 15:43:57.263506 135787484002048 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_4 I0421 15:44:13.395041 135937944586048 metric_logger.py:196] completed step: 6, seconds: 46.479, TFLOP/s/device: 0.292, Tokens/s/device: 44.063, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0421 15:44:13.542555 135937944586048 metric_logger.py:196] completed step: 7, seconds: 22.357, TFLOP/s/device: 0.608, Tokens/s/device: 91.603, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0421 15:44:13.689894 135937944586048 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.102, Tokens/s/device: 13430.389, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0421 15:44:13.836694 135937944586048 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0421 15:44:13.837339 135937944586048 checkpointing.py:776] Waited 0.0006668567657470703 seconds for step 9 to finish before starting checkpointing. I0421 15:44:13.839748 135937944586048 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 15:44:33.865821 135807469876992 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 36.709 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.02996492385864 s) (per-host) I0421 15:44:33.865932 135807469876992 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.988376s. I0421 15:44:43.268101 135807469876992 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 52.390529s. I0421 15:44:43.268432 135807461484288 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 15:44:43.268555 135807461484288 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0421 15:44:43.268604 135807461484288 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0421 15:44:43.270420 135807461484288 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0421 15:44:43.270636 135937944586048 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0421 15:44:43.270769 135937944586048 checkpoint_manager.py:1452] Waiting for previous save to complete took 29.431031 seconds. If this number is high, consider checkpointing less frequently. I0421 15:44:43.272327 135937944586048 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 9 I0421 15:44:43.274278 135937944586048 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints/9. I0421 15:44:43.566447 135937944586048 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 15:44:43.566545 135937944586048 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 15:44:43.602168 135937944586048 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.036757s I0421 15:44:43.602336 135937944586048 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 38.351 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.040222883224487305 s) (per-host) I0421 15:44:43.602387 135937944586048 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.040285s (batch_requests_ready=0.001699s, total_serialization_initiated=0.038519s, others=0.000067s) I0421 15:44:43.602481 135937944586048 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.044560s (all_items=0.000012s, per_item={'items': '0.00001168'}, temp_paths=0.044548) I0421 15:44:43.603201 135937944586048 event_tracking.py:125] [process=4] [async] Finished blocking save in 0.33 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints/9. I0421 15:44:43.603529 135807461484288 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 16:04:43.603492 I0421 15:44:43.605390 135937944586048 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0421 15:44:43.605693 135807436306176 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0421 15:44:43.605797 135937944586048 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776786253.8397112, 'wait_for_prev_duration_secs': 29.431031227111816, 'time_between_consecutive_saves_sec': 23.716973304748535, 'checkpointer_blocking_start_time': 1776786283.2723653, 'checkpointer_blocking_duration_secs': 0.3313279151916504, 'get_old_steps_start_time': 1776786283.6037116, 'get_old_steps_duration_secs': 2.4557113647460938e-05, 'checkpoint_manager_blocking_start_time': 1776786253.837565, 'checkpoint_manager_blocking_duration_secs': 29.76819920539856} I0421 15:44:43.605949 135937944586048 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0421 15:44:43.605993 135937944586048 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 15:44:48.910558 135787484002048 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_144122/linen_xpk_main_20260421_144122_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_4 I0421 15:45:25.827128 135807461484288 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 37.374 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.264976501464844 s) (per-host) I0421 15:45:25.827253 135807461484288 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.223612s. I0421 15:45:34.459394 135807461484288 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 50.855737s. I0421 15:45:34.459686 135807436306176 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 15:45:34.459801 135807436306176 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0421 15:45:34.459843 135807436306176 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0421 15:45:34.462264 135807436306176 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0421 15:45:34.462374 135937944586048 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0421 15:45:34.462451 135937944586048 checkpoint_manager.py:2009] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 15:45:34.463340 135937944586048 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.138, Tokens/s/device: 13888.041, total_weights: 65536, loss: 7.181, lm_loss: 7.181, perplexity: 1314.495 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Tue Apr 21 15:45:45 UTC 2026 EXIT_CODE=0