XPK Start: Mon Apr 20 18:01:41 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-20 18:02:05.835321: 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 18:02:06.014656 138084015765312 max_utils.py:273] Attempting to initialize the jax distributed system... I0420 18:02:15.056138 138084015765312 distributed.py:149] Starting JAX distributed service on [::]:8482 I0420 18:02:15.058497 138084015765312 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-true-rq2e7-slice-job-0-0.mt-10-shardy-true-rq2e7:8482 I0420 18:02:16.645196 138084015765312 max_utils.py:284] Jax distributed system initialized! I0420 18:02:22.799309 138084015765312 max_utils.py:800] System Information: Jax Version: 0.9.2 I0420 18:02:22.799412 138084015765312 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0420 18:02:22.799453 138084015765312 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 I0420 18:02:22.799487 138084015765312 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 18:02:23.505934 138084015765312 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 18:02:23.620711 138084015765312 checkpointing.py:688] Setting up checkpoint logger... I0420 18:02:23.620824 138084015765312 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 18:02:23.620870 138084015765312 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0420 18:02:23.621125 138084015765312 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 0x7d957933c1d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 18:02:27.280554 138084015765312 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 18:02:27.280794 138084015765312 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 0x7d7ffc752cf0>}, handler_registry=None I0420 18:02:27.281029 138084015765312 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d7ffc752cf0>` 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 18:02:27.281089 138084015765312 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d807c51bc50>` 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 18:02:27.281130 138084015765312 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 0x7d7ffc752cf0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d7ffc752cf0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d807c51bc50>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d807c51bc50>}). I0420 18:02:27.281445 138084015765312 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0420 18:02:27.281517 138084015765312 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d80fc2dcfe0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0420 18:02:27.990316 138084015765312 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints I0420 18:02:28.434784 138084015765312 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d80fc4694f0> I0420 18:02:28.434956 138084015765312 checkpointing.py:302] Checkpoint manager created! I0420 18:02:29.392338 138084015765312 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 18:02:29.392454 138084015765312 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 18:02:30.747213 138084015765312 maxtext_utils.py:1836] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0420 18:02:30.747325 138084015765312 maxtext_utils.py:1836] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 18:02:30.747377 138084015765312 maxtext_utils.py:1836] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 18:02:30.747419 138084015765312 maxtext_utils.py:1836] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0420 18:02:30.747456 138084015765312 maxtext_utils.py:1836] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 18:02:30.747491 138084015765312 maxtext_utils.py:1836] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 18:02:30.747528 138084015765312 maxtext_utils.py:1836] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 18:02:30.747564 138084015765312 maxtext_utils.py:1836] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0420 18:02:30.747598 138084015765312 maxtext_utils.py:1836] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 18:02:30.747634 138084015765312 maxtext_utils.py:1836] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 18:02:30.747667 138084015765312 maxtext_utils.py:1836] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0420 18:02:30.747698 138084015765312 maxtext_utils.py:1836] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0420 18:02:31.150599 138084015765312 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 18:02:31.150700 138084015765312 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 18:02:31.156054 138084015765312 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 18:02:31.156135 138084015765312 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 18:02:31.172466 138084015765312 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 18:02:31.172535 138084015765312 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 18:02:31.188154 138084015765312 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 18:02:31.188215 138084015765312 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 18:02:31.213417 138084015765312 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 18:02:31.213487 138084015765312 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 18:02:31.229234 138084015765312 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 18:02:31.229294 138084015765312 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 18:02:31.245169 138084015765312 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 18:02:31.245229 138084015765312 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 18:02:31.275197 138084015765312 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 18:02:31.275265 138084015765312 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 18:02:31.294955 138084015765312 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 18:02:31.295019 138084015765312 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 18:02:42.301692 138084015765312 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 18:02:42.310913 138084015765312 metric_logger.py:301] number parameters: 1.104 billion I0420 18:02:53.733029 138084015765312 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 18:02:53.883958 138084015765312 checkpointing.py:798] Waited 0.15091681480407715 seconds for step 0 to finish before starting checkpointing. I0420 18:02:53.886353 138084015765312 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 18:02:53.888297 138084015765312 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0420 18:02:53.890015 138084015765312 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0. I0420 18:02:54.187579 138084015765312 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 18:02:54.207200 138084015765312 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0420 18:02:54.207272 138084015765312 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 18:02:54.479765 138084015765312 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.274056s I0420 18:02:54.480581 138084015765312 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.416 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28482913970947266 s) (per-host) I0420 18:02:54.480651 138084015765312 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.284911s (batch_requests_ready=0.003007s, total_serialization_initiated=0.281182s, others=0.000722s) I0420 18:02:54.480730 138084015765312 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.289441s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.289425) I0420 18:02:54.481565 138084015765312 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.59 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0. I0420 18:02:54.481836 137954529482496 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 18:22:54.481814 I0420 18:02:54.492486 138084015765312 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 18:02:54.492746 137950204733184 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 18:02:54.492866 138084015765312 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776708173.8863347, 'wait_for_prev_duration_secs': 6.198883056640625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776708173.888336, 'checkpointer_blocking_duration_secs': 0.5936002731323242, 'get_old_steps_start_time': 1776708174.4819632, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776708173.8843553, 'checkpoint_manager_blocking_duration_secs': 0.6084716320037842} I0420 18:02:54.492969 138084015765312 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 18:02:54.493039 138084015765312 max_utils.py:750] Memstats: After params initialized: I0420 18:02:54.493103 138084015765312 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0420 18:02:54.493139 138084015765312 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0420 18:02:54.493168 138084015765312 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0420 18:02:54.493194 138084015765312 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0420 18:02:54.502014 137954521089792 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0 I0420 18:02:54.850136 138084015765312 metric_logger.py:196] completed step: 0, seconds: 11.136, TFLOP/s/device: 1.220, Tokens/s/device: 183.905, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0420 18:02:54.853618 138084015765312 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/tensorboard/' I0420 18:02:55.337072 138084015765312 metric_logger.py:196] completed step: 1, seconds: 1.113, TFLOP/s/device: 12.203, Tokens/s/device: 1839.399, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0420 18:02:55.478854 138084015765312 metric_logger.py:196] completed step: 2, seconds: 0.488, TFLOP/s/device: 27.847, Tokens/s/device: 4197.358, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133 I0420 18:02:55.626315 138084015765312 metric_logger.py:196] completed step: 3, seconds: 0.013, TFLOP/s/device: 1068.003, Tokens/s/device: 160980.978, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16855.789 I0420 18:02:56.690826 137954504304384 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0/items I0420 18:02:57.235049 137948615927552 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776708176395693109, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776708176395693109, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0/_CHECKPOINT_METADATA I0420 18:02:57.850157 2843 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 18:03:00.510771 137950255904512 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0/items/array_metadatas/process_0 I0420 18:03:18.304910 138084015765312 metric_logger.py:196] completed step: 4, seconds: 0.143, TFLOP/s/device: 94.936, Tokens/s/device: 14309.770, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10661.594 I0420 18:03:18.316845 138084015765312 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.497, Tokens/s/device: 13942.121, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7245.978 I0420 18:03:18.460112 138084015765312 metric_logger.py:196] completed step: 6, seconds: 22.678, TFLOP/s/device: 0.599, Tokens/s/device: 90.306, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.156 I0420 18:03:18.607890 138084015765312 metric_logger.py:196] completed step: 7, seconds: 0.008, TFLOP/s/device: 1607.754, Tokens/s/device: 242338.185, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.465 I0420 18:03:18.755167 138084015765312 metric_logger.py:196] completed step: 8, seconds: 0.144, TFLOP/s/device: 94.066, Tokens/s/device: 14178.701, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.256 I0420 18:03:18.903277 138084015765312 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 18:03:18.903933 138084015765312 checkpointing.py:798] Waited 0.0006678104400634766 seconds for step 9 to finish before starting checkpointing. I0420 18:03:18.905977 138084015765312 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 18:03:35.045043 137950213125888 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.192267s (commit=35.979213s, array_metadata_write=1.213055s) I0420 18:03:35.046231 137954529482496 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 38.668 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.85044836997986 s) (per-host) I0420 18:03:35.046357 137954529482496 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.564447s. I0420 18:03:35.797486 137954529482496 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776708176395693109, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0/_CHECKPOINT_METADATA I0420 18:03:36.558604 137948615927552 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776708176395693109, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0/_CHECKPOINT_METADATA I0420 18:03:37.371865 137954529482496 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000394s. I0420 18:03:37.712294 137954529482496 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 18:03:38.653193 137954529482496 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.734502s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0/items I0420 18:03:38.654913 137954529482496 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0/items I0420 18:03:39.234918 137954529482496 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0 I0420 18:03:40.669881 137954529482496 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0`. I0420 18:03:40.670806 137954529482496 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 46.78 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/0 I0420 18:03:40.672481 137954529482496 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 46.190567s. I0420 18:03:40.672620 137950204733184 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 18:03:40.672711 137950204733184 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 18:03:40.672775 137950204733184 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 18:03:40.675745 137950204733184 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 18:03:40.675908 138084015765312 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. W0420 18:03:40.676041 138084015765312 checkpoint_manager.py:1452] Waiting for previous save to complete took 21.770063 seconds. If this number is high, consider checkpointing less frequently. I0420 18:03:40.677910 138084015765312 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0420 18:03:40.679882 138084015765312 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/9. I0420 18:03:41.412388 138084015765312 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0420 18:03:41.412481 138084015765312 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 18:03:41.439046 138084015765312 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028152s I0420 18:03:41.439935 138084015765312 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 41.760 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.036939382553100586 s) (per-host) I0420 18:03:41.440011 138084015765312 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.037033s (batch_requests_ready=0.002617s, total_serialization_initiated=0.033616s, others=0.000801s) I0420 18:03:41.440108 138084015765312 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.040973s (all_items=0.000022s, per_item={'items': '0.00002217'}, temp_paths=0.040951) I0420 18:03:41.440810 138084015765312 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/9. I0420 18:03:41.441159 137950204733184 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 18:23:41.441123 I0420 18:03:41.448241 138084015765312 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 18:03:41.448493 137953458099968 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 18:03:41.448628 138084015765312 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776708198.9059453, 'wait_for_prev_duration_secs': 21.770063400268555, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776708220.6779494, 'checkpointer_blocking_duration_secs': 0.7633554935455322, 'get_old_steps_start_time': 1776708221.4413295, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776708198.904167, 'checkpoint_manager_blocking_duration_secs': 22.544426679611206} I0420 18:03:41.448730 138084015765312 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 18:03:41.448775 138084015765312 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 18:03:41.664438 137954529482496 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/9 I0420 18:03:43.961330 137953997215488 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/9/items I0420 18:03:47.444886 137953474885376 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/9/items/array_metadatas/process_0 I0420 18:04:27.932985 137953466492672 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.405908s (commit=40.350785s, array_metadata_write=2.055123s) I0420 18:04:27.934041 137950204733184 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.947 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.53101825714111 s) (per-host) I0420 18:04:27.934184 137950204733184 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.492910s. I0420 18:04:30.326028 137950204733184 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000411s. I0420 18:04:30.716103 137950204733184 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 18:04:31.779512 137950204733184 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.060878s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/9/items I0420 18:04:31.781207 137950204733184 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/9/items I0420 18:04:32.333003 137950204733184 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/9 I0420 18:04:34.405139 137950204733184 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_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/9`. I0420 18:04:34.405914 137950204733184 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.73 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_true/checkpoints/9 I0420 18:04:34.407191 137950204733184 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.965915s. I0420 18:04:34.407379 137953458099968 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 18:04:34.407500 137953458099968 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 18:04:34.407562 137953458099968 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 18:04:34.409779 137953458099968 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 18:04:34.409895 138084015765312 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. I0420 18:04:34.410004 138084015765312 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 18:04:34.411022 138084015765312 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.199, Tokens/s/device: 13897.183, 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: Mon Apr 20 18:04:46 UTC 2026 EXIT_CODE=0