XPK Start: Sat Apr 18 22:18:35 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-18 22:18:59.436383: 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) I0418 22:18:59.613757 140313178949440 max_utils.py:273] Attempting to initialize the jax distributed system... I0418 22:19:08.653997 140313178949440 distributed.py:149] Starting JAX distributed service on [::]:8482 I0418 22:19:08.656313 140313178949440 distributed.py:172] Connecting to JAX distributed service on mt-02-synthetic-jfun5-slice-job-0-0.mt-02-synthetic-jfun5:8482 I0418 22:19:10.713223 140313178949440 max_utils.py:284] Jax distributed system initialized! I0418 22:19:16.733688 140313178949440 max_utils.py:800] System Information: Jax Version: 0.9.2 I0418 22:19:16.733791 140313178949440 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0418 22:19:16.733830 140313178949440 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 I0418 22:19:16.733864 140313178949440 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0418 22:19:17.789580 140313178949440 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0418 22:19:17.903450 140313178949440 checkpointing.py:688] Setting up checkpoint logger... I0418 22:19:17.903562 140313178949440 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0418 22:19:17.903607 140313178949440 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0418 22:19:17.903796 140313178949440 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 0x7f9c94314410>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0418 22:19:21.409503 140313178949440 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0418 22:19:21.409690 140313178949440 checkpoint_manager.py:708] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f9b73561bb0>}, handler_registry=None I0418 22:19:21.409922 140313178949440 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f9b73561bb0>` 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`. I0418 22:19:21.409968 140313178949440 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f871c74a5a0>` 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`. I0418 22:19:21.410003 140313178949440 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 0x7f9b73561bb0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f9b73561bb0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f871c74a5a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f871c74a5a0>}). I0418 22:19:21.410358 140313178949440 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0418 22:19:21.410429 140313178949440 async_checkpointer.py:192] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f88302545e0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0418 22:19:23.575292 140313178949440 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints I0418 22:19:23.590440 140313178949440 checkpoint_manager.py:929] [process=3][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f8748559cd0> I0418 22:19:23.590565 140313178949440 checkpointing.py:302] Checkpoint manager created! I0418 22:19:24.569970 140313178949440 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0418 22:19:24.570087 140313178949440 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0418 22:19:25.946499 140313178949440 maxtext_utils.py:1836] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0418 22:19:25.946613 140313178949440 maxtext_utils.py:1836] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0418 22:19:25.946664 140313178949440 maxtext_utils.py:1836] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0418 22:19:25.946707 140313178949440 maxtext_utils.py:1836] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0418 22:19:25.946744 140313178949440 maxtext_utils.py:1836] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0418 22:19:25.946778 140313178949440 maxtext_utils.py:1836] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0418 22:19:25.946813 140313178949440 maxtext_utils.py:1836] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0418 22:19:25.946849 140313178949440 maxtext_utils.py:1836] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0418 22:19:25.946882 140313178949440 maxtext_utils.py:1836] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0418 22:19:25.946913 140313178949440 maxtext_utils.py:1836] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0418 22:19:25.946944 140313178949440 maxtext_utils.py:1836] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0418 22:19:25.946976 140313178949440 maxtext_utils.py:1836] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0418 22:19:26.346724 140313178949440 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0418 22:19:26.346813 140313178949440 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 22:19:26.353159 140313178949440 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0418 22:19:26.353216 140313178949440 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 22:19:26.369581 140313178949440 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 22:19:26.369640 140313178949440 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 22:19:26.385184 140313178949440 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 22:19:26.385281 140313178949440 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 22:19:26.409037 140313178949440 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 22:19:26.409106 140313178949440 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 22:19:26.424795 140313178949440 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 22:19:26.424860 140313178949440 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 22:19:26.440434 140313178949440 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 22:19:26.440493 140313178949440 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 22:19:26.470386 140313178949440 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0418 22:19:26.470457 140313178949440 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 22:19:26.490168 140313178949440 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0418 22:19:26.490228 140313178949440 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0418 22:19:37.736916 140313178949440 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. I0418 22:19:37.739326 140313178949440 metric_logger.py:301] number parameters: 1.104 billion I0418 22:19:49.023422 140313178949440 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0418 22:19:49.174360 140313178949440 checkpointing.py:798] Waited 0.1509265899658203 seconds for step 0 to finish before starting checkpointing. I0418 22:19:49.176782 140313178949440 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 22:19:49.178460 140313178949440 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 0 I0418 22:19:49.179800 140313178949440 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0. I0418 22:19:50.332937 140313178949440 signaling_client.py:364] Using JaxDistributedSignalingClient I0418 22:19:50.333847 140313178949440 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0418 22:19:50.333904 140313178949440 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 I0418 22:19:50.620087 140313178949440 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.287908s I0418 22:19:50.620283 140313178949440 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 5.197 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2968413829803467 s) (per-host) I0418 22:19:50.620342 140313178949440 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.296912s (batch_requests_ready=0.002909s, total_serialization_initiated=0.293924s, others=0.000079s) I0418 22:19:50.620446 140313178949440 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.301313s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.301294) I0418 22:19:50.621222 140313178949440 event_tracking.py:125] [process=3] [async] Finished blocking save in 1.44 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0. I0418 22:19:50.621536 140185182402304 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 22:39:50.621499 I0418 22:19:50.623557 140313178949440 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0418 22:19:50.623856 140183655651072 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0418 22:19:50.624008 140313178949440 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776550789.1767635, 'wait_for_prev_duration_secs': 6.222724914550781e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776550789.1785002, 'checkpointer_blocking_duration_secs': 1.4431867599487305, 'get_old_steps_start_time': 1776550790.6217103, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1776550789.1747437, 'checkpoint_manager_blocking_duration_secs': 1.4492263793945312} I0418 22:19:50.624125 140313178949440 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0418 22:19:50.624201 140313178949440 max_utils.py:750] Memstats: After params initialized: I0418 22:19:50.624261 140313178949440 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0)) I0418 22:19:50.624297 140313178949440 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0)) I0418 22:19:50.624323 140313178949440 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0)) I0418 22:19:50.624349 140313178949440 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0)) I0418 22:19:50.941996 140313178949440 metric_logger.py:196] completed step: 0, seconds: 11.283, TFLOP/s/device: 1.204, Tokens/s/device: 181.512, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0418 22:19:51.111012 140313178949440 metric_logger.py:196] completed step: 1, seconds: 1.917, TFLOP/s/device: 7.089, Tokens/s/device: 1068.495, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0418 22:19:51.558161 140313178949440 metric_logger.py:196] completed step: 2, seconds: 0.023, TFLOP/s/device: 602.961, Tokens/s/device: 90884.885, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133 I0418 22:19:51.706166 140313178949440 metric_logger.py:196] completed step: 3, seconds: 0.437, TFLOP/s/device: 31.068, Tokens/s/device: 4682.952, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16855.789 I0418 22:19:52.002117 140313178949440 metric_logger.py:196] completed step: 4, seconds: 0.165, TFLOP/s/device: 82.481, Tokens/s/device: 12432.390, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10661.594 I0418 22:19:52.009975 140313178949440 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.225, Tokens/s/device: 13901.144, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7245.978 I0418 22:19:55.610541 2799 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0418 22:19:57.987582 140183676643072 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/0/items/array_metadatas/process_3 I0418 22:20:16.180463 140313178949440 metric_logger.py:196] completed step: 6, seconds: 0.297, TFLOP/s/device: 45.792, Tokens/s/device: 6902.339, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.156 I0418 22:20:16.328122 140313178949440 metric_logger.py:196] completed step: 7, seconds: 24.024, TFLOP/s/device: 0.566, Tokens/s/device: 85.248, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.465 I0418 22:20:16.475636 140313178949440 metric_logger.py:196] completed step: 8, seconds: 0.154, TFLOP/s/device: 88.474, Tokens/s/device: 13335.764, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.256 I0418 22:20:16.623950 140313178949440 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0418 22:20:16.624628 140313178949440 checkpointing.py:798] Waited 0.0006949901580810547 seconds for step 9 to finish before starting checkpointing. I0418 22:20:16.626722 140313178949440 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 22:20:29.669080 140185182402304 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 40.147 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 39.34559440612793 s) (per-host) I0418 22:20:29.669199 140185182402304 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 39.047553s. I0418 22:20:38.683753 140185182402304 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 48.062090s. I0418 22:20:38.684064 140183655651072 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 22:20:38.684192 140183655651072 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0418 22:20:38.684262 140183655651072 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0418 22:20:38.686605 140183655651072 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0418 22:20:38.686785 140313178949440 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0418 22:20:38.686921 140313178949440 checkpoint_manager.py:1452] Waiting for previous save to complete took 22.060200 seconds. If this number is high, consider checkpointing less frequently. I0418 22:20:38.688756 140313178949440 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 9 I0418 22:20:38.690832 140313178949440 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9. I0418 22:20:39.404900 140313178949440 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0418 22:20:39.404996 140313178949440 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 I0418 22:20:39.440219 140313178949440 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037073s I0418 22:20:39.440407 140313178949440 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 34.969 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0441129207611084 s) (per-host) I0418 22:20:39.440458 140313178949440 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.044174s (batch_requests_ready=0.002660s, total_serialization_initiated=0.041446s, others=0.000067s) I0418 22:20:39.440558 140313178949440 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.048603s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.048587) I0418 22:20:39.441324 140313178949440 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9. I0418 22:20:39.441587 140183655651072 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 22:40:39.441563 I0418 22:20:39.445650 140313178949440 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0418 22:20:39.445890 140184242878208 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0418 22:20:39.446047 140313178949440 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776550816.6266894, 'wait_for_prev_duration_secs': 22.060199975967407, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776550838.6887941, 'checkpointer_blocking_duration_secs': 0.7529022693634033, 'get_old_steps_start_time': 1776550839.4417212, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776550816.6248584, 'checkpoint_manager_blocking_duration_secs': 22.82115387916565} I0418 22:20:39.446154 140313178949440 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0418 22:20:39.446197 140313178949440 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 22:20:45.874285 140185139939072 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152/nnx_xpk_feat_nnx_linen_converter_and_sharding_tools_20260418_204152_02_synthetic/checkpoints/9/items/array_metadatas/process_3 I0418 22:21:21.378307 140183655651072 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 37.626 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.98195481300354 s) (per-host) I0418 22:21:21.378433 140183655651072 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.936761s. I0418 22:21:27.647137 140183655651072 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 48.205449s. I0418 22:21:27.647471 140184242878208 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 22:21:27.647604 140184242878208 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0418 22:21:27.647654 140184242878208 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0418 22:21:27.650269 140184242878208 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0418 22:21:27.650446 140313178949440 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0418 22:21:27.650612 140313178949440 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 22:21:27.651619 140313178949440 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.776, Tokens/s/device: 13833.445, total_weights: 65536, loss: 8.175, lm_loss: 8.175, perplexity: 3551.438 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 18 22:21:35 UTC 2026 EXIT_CODE=0