XPK Start: Sun Apr 19 02:04:39 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-19 02:05:03.618421: 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) I0419 02:05:03.797850 137447666476864 max_utils.py:273] Attempting to initialize the jax distributed system... I0419 02:05:12.838316 137447666476864 distributed.py:149] Starting JAX distributed service on [::]:8482 I0419 02:05:12.840645 137447666476864 distributed.py:172] Connecting to JAX distributed service on mt-13-scan-layers-true-2qx19-slice-job-0-0.mt-13-scan-layers-true-2qx19:8482 I0419 02:05:14.187839 137447666476864 max_utils.py:284] Jax distributed system initialized! I0419 02:05:20.393068 137447666476864 max_utils.py:800] System Information: Jax Version: 0.9.2 I0419 02:05:20.393171 137447666476864 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0419 02:05:20.393214 137447666476864 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 I0419 02:05:20.393250 137447666476864 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 02:05:21.080511 137447666476864 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 02:05:21.193822 137447666476864 checkpointing.py:688] Setting up checkpoint logger... I0419 02:05:21.193936 137447666476864 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 02:05:21.193981 137447666476864 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0419 02:05:21.194188 137447666476864 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 0x7d014f847950>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 02:05:24.071505 137447666476864 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 02:05:24.071697 137447666476864 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cec10736e10>}, handler_registry=None I0419 02:05:24.071945 137447666476864 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cec10736e10>` 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`. I0419 02:05:24.071995 137447666476864 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cec1074c350>` 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`. I0419 02:05:24.072031 137447666476864 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 0x7cec10736e10>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cec10736e10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cec1074c350>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cec1074c350>}). I0419 02:05:24.072342 137447666476864 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0419 02:05:24.072443 137447666476864 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7cec103ccfe0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0419 02:05:26.567997 137447666476864 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints I0419 02:05:26.983871 137447666476864 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_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7cec10782de0> I0419 02:05:26.984030 137447666476864 checkpointing.py:302] Checkpoint manager created! I0419 02:05:27.935730 137447666476864 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 02:05:27.935896 137447666476864 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0419 02:05:29.349372 137447666476864 maxtext_utils.py:1836] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0419 02:05:29.349486 137447666476864 maxtext_utils.py:1836] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 02:05:29.349538 137447666476864 maxtext_utils.py:1836] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 02:05:29.349581 137447666476864 maxtext_utils.py:1836] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0419 02:05:29.349618 137447666476864 maxtext_utils.py:1836] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 02:05:29.349653 137447666476864 maxtext_utils.py:1836] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 02:05:29.349689 137447666476864 maxtext_utils.py:1836] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 02:05:29.349725 137447666476864 maxtext_utils.py:1836] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0419 02:05:29.349758 137447666476864 maxtext_utils.py:1836] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 02:05:29.349792 137447666476864 maxtext_utils.py:1836] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 02:05:29.349837 137447666476864 maxtext_utils.py:1836] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0419 02:05:29.349872 137447666476864 maxtext_utils.py:1836] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0419 02:05:29.752187 137447666476864 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 02:05:29.752274 137447666476864 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 02:05:29.757647 137447666476864 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 02:05:29.757702 137447666476864 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 02:05:29.774078 137447666476864 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 02:05:29.774141 137447666476864 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 02:05:29.789736 137447666476864 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 02:05:29.789803 137447666476864 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 02:05:29.814778 137447666476864 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 02:05:29.814857 137447666476864 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 02:05:29.830565 137447666476864 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 02:05:29.830626 137447666476864 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 02:05:29.846446 137447666476864 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 02:05:29.846507 137447666476864 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 02:05:29.877498 137447666476864 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 02:05:29.877570 137447666476864 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 02:05:29.897549 137447666476864 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 02:05:29.897613 137447666476864 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 02:05:41.237375 137447666476864 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. I0419 02:05:41.246753 137447666476864 metric_logger.py:301] number parameters: 1.104 billion I0419 02:05:52.630900 137447666476864 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0419 02:05:52.782994 137447666476864 checkpointing.py:798] Waited 0.15207982063293457 seconds for step 0 to finish before starting checkpointing. I0419 02:05:52.785243 137447666476864 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 02:05:52.787007 137447666476864 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0419 02:05:52.788497 137447666476864 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0. I0419 02:05:53.540009 137447666476864 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 02:05:53.560259 137447666476864 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0419 02:05:53.560355 137447666476864 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 I0419 02:05:53.831892 137318059116288 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0 I0419 02:05:53.833516 137447666476864 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.274650s I0419 02:05:53.834328 137447666476864 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.399 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2857046127319336 s) (per-host) I0419 02:05:53.834416 137447666476864 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.285813s (batch_requests_ready=0.003140s, total_serialization_initiated=0.281923s, others=0.000750s) I0419 02:05:53.834505 137447666476864 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.290494s (all_items=0.000024s, per_item={'items': '0.00002384'}, temp_paths=0.290470) I0419 02:05:53.835323 137447666476864 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.05 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0. I0419 02:05:53.835609 137318067508992 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:25:53.835576 I0419 02:05:53.845521 137447666476864 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0419 02:05:53.845791 137315366328064 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 02:05:53.845941 137447666476864 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776564352.7852242, 'wait_for_prev_duration_secs': 6.389617919921875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776564352.787047, 'checkpointer_blocking_duration_secs': 1.0487029552459717, 'get_old_steps_start_time': 1776564353.8357768, 'get_old_steps_duration_secs': 2.6702880859375e-05, 'checkpoint_manager_blocking_start_time': 1776564352.7833853, 'checkpoint_manager_blocking_duration_secs': 1.0625174045562744} I0419 02:05:53.846047 137447666476864 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0419 02:05:53.846118 137447666476864 max_utils.py:750] Memstats: After params initialized: I0419 02:05:53.846181 137447666476864 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0419 02:05:53.846219 137447666476864 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0419 02:05:53.846244 137447666476864 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0419 02:05:53.846270 137447666476864 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0419 02:05:54.195332 137447666476864 metric_logger.py:196] completed step: 0, seconds: 11.064, TFLOP/s/device: 1.228, Tokens/s/device: 185.107, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0419 02:05:54.198129 137447666476864 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/tensorboard/' I0419 02:05:54.589661 137447666476864 metric_logger.py:196] completed step: 1, seconds: 1.561, TFLOP/s/device: 8.706, Tokens/s/device: 1312.196, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0419 02:05:54.731487 137447666476864 metric_logger.py:196] completed step: 2, seconds: 0.395, TFLOP/s/device: 34.380, Tokens/s/device: 5182.081, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133 I0419 02:05:54.878853 137447666476864 metric_logger.py:196] completed step: 3, seconds: 0.012, TFLOP/s/device: 1108.249, Tokens/s/device: 167047.308, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16855.789 I0419 02:05:56.605629 137313229203200 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776564356184906056, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776564356184906056, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0419 02:05:56.870469 137318042330880 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0/items I0419 02:05:58.002179 2833 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 02:06:00.576980 137316461459200 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_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0/items/array_metadatas/process_0 I0419 02:06:19.609117 137447666476864 metric_logger.py:196] completed step: 4, seconds: 0.143, TFLOP/s/device: 94.852, Tokens/s/device: 14297.083, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10661.594 I0419 02:06:19.621074 137447666476864 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.553, Tokens/s/device: 13950.669, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7245.978 I0419 02:06:19.764369 137447666476864 metric_logger.py:196] completed step: 6, seconds: 24.731, TFLOP/s/device: 0.549, Tokens/s/device: 82.813, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.156 I0419 02:06:19.912206 137447666476864 metric_logger.py:196] completed step: 7, seconds: 0.008, TFLOP/s/device: 1646.326, Tokens/s/device: 248152.187, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.465 I0419 02:06:20.059547 137447666476864 metric_logger.py:196] completed step: 8, seconds: 0.145, TFLOP/s/device: 93.853, Tokens/s/device: 14146.480, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.256 I0419 02:06:20.207862 137447666476864 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0419 02:06:20.208647 137447666476864 checkpointing.py:798] Waited 0.0007944107055664062 seconds for step 9 to finish before starting checkpointing. I0419 02:06:20.210629 137447666476864 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 02:06:35.302336 137315406051072 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.307365s (commit=36.124912s, array_metadata_write=1.182453s) I0419 02:06:35.303447 137318067508992 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.830 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.754820585250854 s) (per-host) I0419 02:06:35.303516 137318067508992 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.467800s. I0419 02:06:36.042496 137318067508992 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776564356184906056, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0419 02:06:36.753232 137313229203200 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776564356184906056, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0419 02:06:37.177320 137318067508992 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000428s. I0419 02:06:37.446551 137318067508992 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 02:06:38.359677 137318067508992 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.191191s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0/items I0419 02:06:38.361397 137318067508992 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0/items I0419 02:06:39.331735 137318067508992 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0 I0419 02:06:41.125562 137318067508992 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0`. I0419 02:06:41.126424 137318067508992 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 48.34 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/0 I0419 02:06:41.127961 137318067508992 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 47.292241s. I0419 02:06:41.128174 137315366328064 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 02:06:41.128296 137315366328064 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 02:06:41.128365 137315366328064 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0419 02:06:41.131391 137315366328064 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0419 02:06:41.131559 137447666476864 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0419 02:06:41.131690 137447666476864 checkpoint_manager.py:1452] Waiting for previous save to complete took 20.921060 seconds. If this number is high, consider checkpointing less frequently. I0419 02:06:41.133537 137447666476864 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0419 02:06:41.135794 137447666476864 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/9. I0419 02:06:42.285175 137447666476864 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0419 02:06:42.285273 137447666476864 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 I0419 02:06:42.312910 137447666476864 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.029133s I0419 02:06:42.313714 137447666476864 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 40.725 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.037877798080444336 s) (per-host) I0419 02:06:42.313786 137447666476864 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.037970s (batch_requests_ready=0.002635s, total_serialization_initiated=0.034616s, others=0.000719s) I0419 02:06:42.313888 137447666476864 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.042055s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.042039) I0419 02:06:42.314607 137447666476864 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.18 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/9. I0419 02:06:42.314895 137315366328064 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:26:42.314863 I0419 02:06:42.329247 137447666476864 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0419 02:06:42.329539 137317008271104 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 02:06:42.329705 137447666476864 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776564380.2105982, 'wait_for_prev_duration_secs': 20.921059608459473, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776564401.1335769, 'checkpointer_blocking_duration_secs': 1.1814463138580322, 'get_old_steps_start_time': 1776564402.3150475, 'get_old_steps_duration_secs': 3.0279159545898438e-05, 'checkpoint_manager_blocking_start_time': 1776564380.2089512, 'checkpoint_manager_blocking_duration_secs': 22.120718240737915} I0419 02:06:42.329837 137447666476864 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0419 02:06:42.329883 137447666476864 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 02:06:42.527379 137318067508992 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/9 I0419 02:06:44.724581 137317553534720 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/9/items I0419 02:06:48.413714 137317503715072 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_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/9/items/array_metadatas/process_0 I0419 02:07:28.713839 137317016663808 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.461725s (commit=40.777934s, array_metadata_write=1.683791s) I0419 02:07:28.715021 137315366328064 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.014 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.43914818763733 s) (per-host) I0419 02:07:28.715179 137315366328064 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.400175s. I0419 02:07:30.258344 137315366328064 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000438s. I0419 02:07:31.028381 137315366328064 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 02:07:31.965031 137315366328064 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.336688s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/9/items I0419 02:07:31.969994 137315366328064 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/9/items I0419 02:07:32.500825 137315366328064 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/9 I0419 02:07:34.345479 137315366328064 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/9`. I0419 02:07:34.346341 137315366328064 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.21 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_13_scan_layers_true/checkpoints/9 I0419 02:07:34.347774 137315366328064 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.032770s. I0419 02:07:34.347976 137317008271104 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 02:07:34.348090 137317008271104 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 02:07:34.348162 137317008271104 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0419 02:07:34.350258 137317008271104 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0419 02:07:34.350439 137447666476864 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0419 02:07:34.350611 137447666476864 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 02:07:34.351514 137447666476864 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.129, Tokens/s/device: 13886.629, 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: Sun Apr 19 02:07:45 UTC 2026 EXIT_CODE=0