XPK Start: Mon Apr 20 21:28:26 UTC 2026 2026-04-20 21:28:30.293928: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered WARNING: All log messages before absl::InitializeLog() is called are written to STDERR E0000 00:00:1776720510.306802 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776720510.310491 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776720510.321940 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776720510.321961 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776720510.321964 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776720510.321966 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-20 21:28:49.519903: 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 21:28:50.044126 137902528489280 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-20 21:28:59,085:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0420 21:28:59.085464 137902528489280 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-20 21:28:59,087:jax._src.distributed:157: Connecting to JAX distributed service on mt-11-optimizer-offload-sz823-slice-job-0-0.mt-11-optimizer-offload-sz823:8482 I0420 21:28:59.087860 137902528489280 distributed.py:157] Connecting to JAX distributed service on mt-11-optimizer-offload-sz823-slice-job-0-0.mt-11-optimizer-offload-sz823:8482 I0420 21:29:00.074206 137902528489280 max_utils.py:284] Jax distributed system initialized! I0420 21:29:06.079424 137902528489280 max_utils.py:800] System Information: Jax Version: 0.8.1 I0420 21:29:06.079531 137902528489280 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0420 21:29:06.079581 137902528489280 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0420 21:29:06.079617 137902528489280 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 21:29:06.685209 137902528489280 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 21:29:06.799390 137902528489280 checkpointing.py:688] Setting up checkpoint logger... I0420 21:29:06.799510 137902528489280 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 21:29:06.799557 137902528489280 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0420 21:29:06.799766 137902528489280 base_pytree_checkpoint_handler.py:415] 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 0x7d6b65d71430>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 21:29:09.710125 137902528489280 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 21:29:09.710477 137902528489280 checkpoint_manager.py:709] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d57342f8320>}, handler_registry=None I0420 21:29:09.710715 137902528489280 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d57342f8320>` 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 21:29:09.710765 137902528489280 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d57342f8a10>` 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 21:29:09.710801 137902528489280 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 0x7d57342f8320>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d57342f8320>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d57342f8a10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d57342f8a10>}). I0420 21:29:09.711150 137902528489280 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0420 21:29:09.711227 137902528489280 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d573422cf40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 21:29:10.956155 137902528489280 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints I0420 21:29:10.958354 137902528489280 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_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d5734213e00> I0420 21:29:10.958467 137902528489280 checkpointing.py:302] Checkpoint manager created! I0420 21:29:11.893799 137902528489280 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 21:29:11.893912 137902528489280 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 21:29:13.303220 137902528489280 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0420 21:29:13.303359 137902528489280 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 21:29:13.303440 137902528489280 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 21:29:13.303511 137902528489280 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0420 21:29:13.303573 137902528489280 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 21:29:13.303632 137902528489280 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 21:29:13.303692 137902528489280 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:29:13.303750 137902528489280 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0420 21:29:13.303807 137902528489280 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:29:13.303869 137902528489280 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:29:13.303925 137902528489280 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0420 21:29:13.303978 137902528489280 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0420 21:29:13.705439 137902528489280 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 21:29:13.705589 137902528489280 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:29:13.710994 137902528489280 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 21:29:13.711107 137902528489280 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:29:13.861872 137902528489280 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 21:29:13.862008 137902528489280 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:29:14.011480 137902528489280 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 21:29:14.011612 137902528489280 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:29:14.166945 137902528489280 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 21:29:14.167097 137902528489280 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:29:14.315352 137902528489280 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 21:29:14.315486 137902528489280 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:29:14.464528 137902528489280 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 21:29:14.464666 137902528489280 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:29:14.623859 137902528489280 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 21:29:14.623998 137902528489280 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:29:14.646989 137902528489280 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 21:29:14.647099 137902528489280 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 21:29:25.532033 137902528489280 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 21:29:25.539527 137902528489280 metric_logger.py:289] number parameters: 1.104 billion I0420 21:29:36.756461 137902528489280 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 21:29:36.907795 137902528489280 checkpointing.py:798] Waited 0.1513228416442871 seconds for step 0 to finish before starting checkpointing. I0420 21:29:36.910133 137902528489280 checkpoint_manager.py:2013] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 21:29:36.911881 137902528489280 checkpoint_manager.py:1518] [process=0] Saving checkpoint at step 0 I0420 21:29:36.914660 137902528489280 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0. I0420 21:29:37.207190 137902528489280 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 21:29:37.460780 137778851473152 atomicity.py:137] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0 I0420 21:29:37.510705 137902528489280 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:29:37.510798 137902528489280 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 21:29:37.754949 137902528489280 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.246320s I0420 21:29:37.755757 137902528489280 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.850 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5411877632141113 s) (per-host) I0420 21:29:37.755826 137902528489280 base_pytree_checkpoint_handler.py:737] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.541276s (batch_requests_ready=0.286674s, total_serialization_initiated=0.253889s, others=0.000712s) I0420 21:29:37.755911 137902528489280 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.545084s (all_items=0.000019s, per_item={'items': '0.00001884'}, temp_paths=0.545065) I0420 21:29:37.756881 137778265052928 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0420 21:29:37.757061 137902528489280 async_checkpointer.py:561] Finished blocking save. Time taken: 0.845112s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0. I0420 21:29:38.157225 137902528489280 checkpoint_manager.py:1566] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:29:38.157607 137777240860416 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:29:38.157773 137902528489280 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776720576.9101145, 'wait_for_prev_duration_secs': 6.389617919921875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776720576.9119217, 'checkpointer_blocking_duration_secs': 0.8452560901641846, 'get_old_steps_start_time': 1776720577.7571971, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1776720576.9082186, 'checkpoint_manager_blocking_duration_secs': 1.249516248703003} I0420 21:29:38.157884 137902528489280 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 21:29:38.157960 137902528489280 max_utils.py:750] Memstats: After params initialized: I0420 21:29:38.158016 137902528489280 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0420 21:29:38.158048 137902528489280 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0420 21:29:38.158074 137902528489280 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0420 21:29:38.158116 137902528489280 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0420 21:29:38.456877 137902528489280 metric_logger.py:185] completed step: 0, seconds: 10.943, TFLOP/s/device: 1.242, Tokens/s/device: 187.148, total_weights: 65536, loss: 10.874 I0420 21:29:38.459541 137902528489280 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/tensorboard/' I0420 21:29:38.873174 137902528489280 metric_logger.py:185] completed step: 1, seconds: 1.697, TFLOP/s/device: 8.007, Tokens/s/device: 1206.908, total_weights: 65536, loss: 10.874 I0420 21:29:39.015188 137902528489280 metric_logger.py:185] completed step: 2, seconds: 0.418, TFLOP/s/device: 32.541, Tokens/s/device: 4904.931, total_weights: 65536, loss: 10.262 I0420 21:29:39.162661 137902528489280 metric_logger.py:185] completed step: 3, seconds: 0.011, TFLOP/s/device: 1203.999, Tokens/s/device: 181479.840, total_weights: 65536, loss: 9.732 I0420 21:29:39.333018 137778834687744 atomicity.py:137] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0/items I0420 21:29:58.608655 137777773901568 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776720579070383987, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776720579070383987, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0/_CHECKPOINT_METADATA I0420 21:29:58.610136 137902528489280 metric_logger.py:185] completed step: 4, seconds: 0.144, TFLOP/s/device: 94.588, Tokens/s/device: 14257.271, total_weights: 65536, loss: 9.274 I0420 21:29:58.620126 137902528489280 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.480, Tokens/s/device: 13939.654, total_weights: 65536, loss: 8.888 I0420 21:29:58.764776 137902528489280 metric_logger.py:185] completed step: 6, seconds: 19.446, TFLOP/s/device: 0.699, Tokens/s/device: 105.316, total_weights: 65536, loss: 8.589 I0420 21:29:58.912209 137902528489280 metric_logger.py:185] completed step: 7, seconds: 0.008, TFLOP/s/device: 1720.325, Tokens/s/device: 259306.153, total_weights: 65536, loss: 8.380 I0420 21:29:59.059825 137902528489280 metric_logger.py:185] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.606, Tokens/s/device: 13958.561, total_weights: 65536, loss: 8.251 I0420 21:29:59.208113 137902528489280 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 21:29:59.208759 137902528489280 checkpointing.py:798] Waited 0.0006723403930664062 seconds for step 9 to finish before starting checkpointing. I0420 21:29:59.210719 137902528489280 checkpoint_manager.py:2024] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:29:59.620885 2595 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 21:30:01.321526 137778817902336 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_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0/items/array_metadatas/process_0 I0420 21:30:36.778256 137778809509632 base_pytree_checkpoint_handler.py:1232] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.155048s (commit=36.006314s, array_metadata_write=1.148734s) I0420 21:30:36.779429 137778265052928 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/gbytes_per_sec: 26.519 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 59.56482696533203 s) (per-host) I0420 21:30:36.779549 137778265052928 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.022539s. I0420 21:30:37.462282 137778265052928 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776720579070383987, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0/_CHECKPOINT_METADATA I0420 21:30:38.171900 137778265052928 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000437s. I0420 21:30:38.234889 137777773901568 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776720579070383987, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0/_CHECKPOINT_METADATA I0420 21:30:38.877589 137778265052928 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 21:30:39.809910 137778265052928 base_pytree_checkpoint_handler.py:1356] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.214548s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0/items I0420 21:30:39.815225 137778265052928 atomicity.py:608] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0/items I0420 21:30:40.234335 137778265052928 atomicity.py:608] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0 I0420 21:30:41.559801 137778265052928 atomicity.py:790] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0`. I0420 21:30:41.560647 137778265052928 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 64.648700s. directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/0 I0420 21:30:41.562070 137778265052928 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 63.805060s. I0420 21:30:41.562256 137777240860416 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:30:41.562349 137777240860416 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:30:41.562413 137777240860416 checkpoint_manager.py:2133] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:30:41.564391 137777240860416 checkpoint_manager.py:2142] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 21:30:41.564526 137902528489280 checkpoint_manager.py:2036] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0420 21:30:41.564621 137902528489280 checkpoint_manager.py:1458] Waiting for previous save to complete took 42.353912 seconds. If this number is high, consider checkpointing less frequently. I0420 21:30:41.566562 137902528489280 checkpoint_manager.py:1518] [process=0] Saving checkpoint at step 9 I0420 21:30:41.569974 137902528489280 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/9. I0420 21:30:42.538336 137778265052928 atomicity.py:137] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/9 I0420 21:30:42.552553 137902528489280 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:30:42.552657 137902528489280 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 21:30:42.581681 137902528489280 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.031202s I0420 21:30:42.582503 137902528489280 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.243 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29419565200805664 s) (per-host) I0420 21:30:42.582566 137902528489280 base_pytree_checkpoint_handler.py:737] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.294273s (batch_requests_ready=0.256436s, total_serialization_initiated=0.037111s, others=0.000726s) I0420 21:30:42.582644 137902528489280 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.299195s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.299178) I0420 21:30:42.583487 137777249253120 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0420 21:30:42.583599 137902528489280 async_checkpointer.py:561] Finished blocking save. Time taken: 1.016952s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/9. I0420 21:30:43.027879 137902528489280 checkpoint_manager.py:1566] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:30:43.028258 137777240860416 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:30:43.028435 137902528489280 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776720599.2106872, 'wait_for_prev_duration_secs': 42.35391187667847, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776720641.5666056, 'checkpointer_blocking_duration_secs': 1.0171077251434326, 'get_old_steps_start_time': 1776720642.5837343, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1776720599.2089896, 'checkpoint_manager_blocking_duration_secs': 43.81940937042236} I0420 21:30:43.028558 137902528489280 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 21:30:43.028608 137902528489280 checkpoint_manager.py:2024] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:30:44.316034 137777748723456 atomicity.py:137] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/9/items I0420 21:30:47.221023 137777783879424 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_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/9/items/array_metadatas/process_0 I0420 21:31:27.884146 137777740330752 base_pytree_checkpoint_handler.py:1232] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.406079s (commit=41.313027s, array_metadata_write=1.093052s) I0420 21:31:27.886121 137777249253120 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/gbytes_per_sec: 34.642 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.597790241241455 s) (per-host) I0420 21:31:27.886187 137777249253120 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.302642s. I0420 21:31:29.798246 137777249253120 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000428s. I0420 21:31:29.955630 137777249253120 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 21:31:30.909464 137777249253120 base_pytree_checkpoint_handler.py:1356] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.211557s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/9/items I0420 21:31:30.913355 137777249253120 atomicity.py:608] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/9/items I0420 21:31:31.331414 137777249253120 atomicity.py:608] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/9 I0420 21:31:33.113644 137777249253120 atomicity.py:790] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/9`. I0420 21:31:33.114426 137777249253120 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 51.547796s. directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_false/checkpoints/9 I0420 21:31:33.115777 137777249253120 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 50.532230s. I0420 21:31:33.115917 137777240860416 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:31:33.115983 137777240860416 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:31:33.116044 137777240860416 checkpoint_manager.py:2133] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:31:33.118671 137777240860416 checkpoint_manager.py:2142] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 21:31:33.118839 137902528489280 checkpoint_manager.py:2036] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0420 21:31:33.119006 137902528489280 checkpoint_manager.py:2013] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 21:31:33.119693 137902528489280 metric_logger.py:185] completed step: 9, seconds: 0.146, TFLOP/s/device: 92.752, Tokens/s/device: 13980.572, total_weights: 65536, loss: 8.175 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Mon Apr 20 21:31:41 UTC 2026 EXIT_CODE=0