XPK Start: Mon Apr 20 12:18:51 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 12:19:25.265301: 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 12:19:25.443884 135911003440960 max_utils.py:273] Attempting to initialize the jax distributed system... I0420 12:19:34.485382 135911003440960 distributed.py:149] Starting JAX distributed service on [::]:8482 I0420 12:19:34.487765 135911003440960 distributed.py:172] Connecting to JAX distributed service on mt-02-synthetic-r2tvn-slice-job-0-0.mt-02-synthetic-r2tvn:8482 I0420 12:19:43.599377 135911003440960 max_utils.py:284] Jax distributed system initialized! I0420 12:19:48.727515 135911003440960 max_utils.py:800] System Information: Jax Version: 0.9.2 I0420 12:19:48.727619 135911003440960 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0420 12:19:48.727659 135911003440960 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 12:19:48.727693 135911003440960 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 12:19:49.416909 135911003440960 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 12:19:49.417508 135911003440960 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 12:19:49.417691 135911003440960 checkpointing.py:688] Setting up checkpoint logger... I0420 12:19:49.417743 135911003440960 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 12:19:49.417785 135911003440960 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0420 12:19:49.418119 135911003440960 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 0x7b9b8764c170>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 12:19:52.262522 135911003440960 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 12:19:52.262755 135911003440960 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 0x7b87ec2e60c0>}, handler_registry=None I0420 12:19:52.262999 135911003440960 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b87ec2e60c0>` 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 12:19:52.263048 135911003440960 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b87ec2e9be0>` 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 12:19:52.263098 135911003440960 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 0x7b87ec2e60c0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b87ec2e60c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b87ec2e9be0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b87ec2e9be0>}). I0420 12:19:52.263403 135911003440960 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0420 12:19:52.263472 135911003440960 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7b868868b880> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0420 12:19:53.522068 135911003440960 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints I0420 12:19:54.349396 135911003440960 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/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b87ec2e7c80> I0420 12:19:54.349565 135911003440960 checkpointing.py:302] Checkpoint manager created! I0420 12:19:55.259322 135911003440960 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 12:19:55.259421 135911003440960 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 12:19:55.641539 135911003440960 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 12:19:55.641627 135911003440960 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 12:19:55.658105 135911003440960 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 12:19:55.658161 135911003440960 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 12:19:55.681964 135911003440960 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 12:19:55.682034 135911003440960 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 12:19:55.698518 135911003440960 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 12:19:55.698583 135911003440960 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 12:19:55.715021 135911003440960 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 12:19:55.715090 135911003440960 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 12:19:55.739768 135911003440960 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 12:19:55.739833 135911003440960 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 12:19:55.760708 135911003440960 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 12:19:55.760770 135911003440960 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 12:19:55.978722 135911003440960 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 12:19:55.978824 135911003440960 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 12:19:57.401411 135911003440960 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0420 12:19:57.401536 135911003440960 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0420 12:19:57.401591 135911003440960 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0420 12:19:57.401648 135911003440960 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0420 12:19:57.401701 135911003440960 maxtext_utils.py:1827] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0420 12:19:57.401750 135911003440960 maxtext_utils.py:1827] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0420 12:19:57.401802 135911003440960 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0420 12:19:57.401861 135911003440960 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0420 12:19:57.401903 135911003440960 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0420 12:19:57.401940 135911003440960 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0420 12:19:57.401987 135911003440960 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0420 12:19:57.402033 135911003440960 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0420 12:19:57.905212 135911003440960 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 12:19:57.905303 135911003440960 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0420 12:19:57.920624 135911003440960 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 12:19:57.920685 135911003440960 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0420 12:20:08.471971 135911003440960 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 12:20:08.480625 135911003440960 metric_logger.py:301] number parameters: 1.104 billion I0420 12:20:19.932099 135911003440960 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 12:20:20.119260 135911003440960 checkpointing.py:798] Waited 0.18715739250183105 seconds for step 0 to finish before starting checkpointing. I0420 12:20:20.121650 135911003440960 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 12:20:20.123233 135911003440960 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0420 12:20:20.124624 135911003440960 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0. I0420 12:20:21.268931 135911003440960 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 12:20:21.283465 135911003440960 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 12:20:21.283530 135911003440960 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 12:20:21.553886 135783159682816 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0 I0420 12:20:21.557859 135911003440960 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.275134s I0420 12:20:21.558690 135911003440960 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.479 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28153181076049805 s) (per-host) I0420 12:20:21.558755 135911003440960 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281615s (batch_requests_ready=0.002124s, total_serialization_initiated=0.278754s, others=0.000737s) I0420 12:20:21.558832 135911003440960 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.285686s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.285670) I0420 12:20:21.559629 135911003440960 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.44 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0. I0420 12:20:21.559944 135783168075520 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 12:40:21.559898 I0420 12:20:21.570412 135911003440960 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 12:20:21.570716 135781557462784 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 12:20:21.570875 135911003440960 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776687620.1216319, 'wait_for_prev_duration_secs': 5.984306335449219e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776687620.1232715, 'checkpointer_blocking_duration_secs': 1.436833381652832, 'get_old_steps_start_time': 1776687621.560128, 'get_old_steps_duration_secs': 2.765655517578125e-05, 'checkpoint_manager_blocking_start_time': 1776687620.1198006, 'checkpoint_manager_blocking_duration_secs': 1.451038122177124} I0420 12:20:21.570993 135911003440960 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 12:20:21.571042 135911003440960 max_utils.py:750] Memstats: After params initialized: I0420 12:20:21.571101 135911003440960 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0420 12:20:21.571136 135911003440960 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0420 12:20:21.571162 135911003440960 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0420 12:20:21.571187 135911003440960 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0420 12:20:21.908416 135911003440960 metric_logger.py:196] completed step: 0, seconds: 11.148, TFLOP/s/device: 1.219, Tokens/s/device: 183.712, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0420 12:20:21.911641 135911003440960 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/tensorboard/' I0420 12:20:22.340110 135911003440960 metric_logger.py:196] completed step: 1, seconds: 1.975, TFLOP/s/device: 6.881, Tokens/s/device: 1037.214, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0420 12:20:22.486120 135911003440960 metric_logger.py:196] completed step: 2, seconds: 0.432, TFLOP/s/device: 31.424, Tokens/s/device: 4736.563, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129 I0420 12:20:22.633580 135911003440960 metric_logger.py:196] completed step: 3, seconds: 0.010, TFLOP/s/device: 1353.030, Tokens/s/device: 203943.438, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.203 I0420 12:20:23.935317 135779925878528 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776687623499116253, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776687623499116253, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/_CHECKPOINT_METADATA I0420 12:20:24.185227 135783142897408 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/items I0420 12:20:25.329889 2856 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 12:20:27.136790 135782584862464 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/items/array_metadatas/process_0 I0420 12:20:45.293413 135911003440960 metric_logger.py:196] completed step: 4, seconds: 0.146, TFLOP/s/device: 93.055, Tokens/s/device: 14026.244, total_weights: 65536, loss: 9.277, lm_loss: 9.277, perplexity: 10694.614 I0420 12:20:45.302260 135911003440960 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.302, Tokens/s/device: 13912.760, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.717 I0420 12:20:45.448446 135911003440960 metric_logger.py:196] completed step: 6, seconds: 22.659, TFLOP/s/device: 0.600, Tokens/s/device: 90.384, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.759 I0420 12:20:45.595826 135911003440960 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 2076.273, Tokens/s/device: 312958.435, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4376.905 I0420 12:20:45.743132 135911003440960 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.152, Tokens/s/device: 13890.113, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3846.406 I0420 12:20:45.889718 135911003440960 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 12:20:45.890432 135911003440960 checkpointing.py:798] Waited 0.0007376670837402344 seconds for step 9 to finish before starting checkpointing. I0420 12:20:45.892481 135911003440960 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 12:21:03.090209 135781565855488 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.759164s (commit=36.112665s, array_metadata_write=1.646499s) I0420 12:21:03.091376 135783168075520 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.777 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.81418752670288 s) (per-host) I0420 12:21:03.091516 135783168075520 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.531458s. I0420 12:21:03.816428 135783168075520 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776687623499116253, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/_CHECKPOINT_METADATA I0420 12:21:04.998168 135779925878528 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776687623499116253, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/_CHECKPOINT_METADATA I0420 12:21:05.192070 135783168075520 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 12:21:05.343137 135783168075520 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000256s. I0420 12:21:06.150575 135783168075520 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.210262s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/items I0420 12:21:06.152175 135783168075520 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0/items I0420 12:21:06.693262 135783168075520 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0 I0420 12:21:08.097063 135783168075520 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0`. I0420 12:21:08.097945 135783168075520 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 47.97 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/0 I0420 12:21:08.099475 135783168075520 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 46.539415s. I0420 12:21:08.099686 135781557462784 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 12:21:08.099803 135781557462784 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 12:21:08.099867 135781557462784 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 12:21:08.101771 135781557462784 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 12:21:08.101959 135911003440960 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 12:21:08.102106 135911003440960 checkpoint_manager.py:1452] Waiting for previous save to complete took 22.209608 seconds. If this number is high, consider checkpointing less frequently. I0420 12:21:08.103802 135911003440960 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0420 12:21:08.105854 135911003440960 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9. I0420 12:21:08.814032 135911003440960 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 12:21:08.814139 135911003440960 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 12:21:08.840421 135911003440960 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027117s I0420 12:21:08.841315 135911003440960 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 48.541 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.031778812408447266 s) (per-host) I0420 12:21:08.841385 135911003440960 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031865s (batch_requests_ready=0.001900s, total_serialization_initiated=0.029157s, others=0.000808s) I0420 12:21:08.841467 135911003440960 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035964s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.035948) I0420 12:21:08.842247 135911003440960 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9. I0420 12:21:08.842509 135781557462784 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 12:41:08.842488 I0420 12:21:08.857722 135911003440960 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 12:21:08.857974 135782052329216 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 12:21:08.858116 135911003440960 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776687645.8924494, 'wait_for_prev_duration_secs': 22.20960807800293, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776687668.1038423, 'checkpointer_blocking_duration_secs': 0.7387616634368896, 'get_old_steps_start_time': 1776687668.8426268, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1776687645.8906558, 'checkpoint_manager_blocking_duration_secs': 22.967417001724243} I0420 12:21:08.858232 135911003440960 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 12:21:08.858276 135911003440960 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 12:21:09.060873 135783168075520 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9 I0420 12:21:11.261028 135783142897408 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9/items I0420 12:21:14.218188 135782069114624 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9/items/array_metadatas/process_0 I0420 12:21:55.252045 135782060721920 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.845471s (commit=41.613860s, array_metadata_write=1.231610s) I0420 12:21:55.253047 135781557462784 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.011 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.44350051879883 s) (per-host) I0420 12:21:55.253129 135781557462784 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.410546s. I0420 12:21:56.757499 135781557462784 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000244s. I0420 12:21:56.980625 135781557462784 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 12:21:58.053137 135781557462784 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.948584s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9/items I0420 12:21:58.055001 135781557462784 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9/items I0420 12:21:58.582047 135781557462784 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9 I0420 12:22:00.520127 135781557462784 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9`. I0420 12:22:00.520936 135781557462784 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.42 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_02_synthetic/checkpoints/9 I0420 12:22:00.522520 135781557462784 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.679934s. I0420 12:22:00.522708 135782052329216 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 12:22:00.522826 135782052329216 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 12:22:00.522892 135782052329216 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 12:22:00.525391 135782052329216 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 12:22:00.525534 135911003440960 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 12:22:00.525616 135911003440960 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 12:22:00.526609 135911003440960 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.260, Tokens/s/device: 13906.430, total_weights: 65536, loss: 8.179, lm_loss: 8.179, perplexity: 3564.635 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Mon Apr 20 12:22:11 UTC 2026 EXIT_CODE=0