XPK Start: Wed Apr 22 13:57:25 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-22 13:57:49.401761: 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) I0422 13:57:49.582879 134671428339520 max_utils.py:273] Attempting to initialize the jax distributed system... I0422 13:57:58.625603 134671428339520 distributed.py:149] Starting JAX distributed service on [::]:8482 I0422 13:57:58.630996 134671428339520 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-true-sv-belzk-slice-job-0-0.mt-16-zarr3-true-sv-belzk:8482 I0422 13:58:00.265037 134671428339520 max_utils.py:284] Jax distributed system initialized! I0422 13:58:06.356760 134671428339520 max_utils.py:800] System Information: Jax Version: 0.9.2 I0422 13:58:06.356863 134671428339520 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0422 13:58:06.356902 134671428339520 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 I0422 13:58:06.356933 134671428339520 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 13:58:07.057922 134671428339520 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 13:58:07.058529 134671428339520 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 13:58:07.058711 134671428339520 checkpointing.py:688] Setting up checkpoint logger... I0422 13:58:07.058762 134671428339520 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 13:58:07.058804 134671428339520 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0422 13:58:07.059143 134671428339520 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 0x7a7b0bd14920>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 13:58:10.115877 134671428339520 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 13:58:10.116122 134671428339520 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a67742b8e60>}, handler_registry=None I0422 13:58:10.116362 134671428339520 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a67742b8e60>` 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`. I0422 13:58:10.116412 134671428339520 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a677430dbb0>` 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`. I0422 13:58:10.116447 134671428339520 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 0x7a67742b8e60>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a67742b8e60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a677430dbb0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a677430dbb0>}). I0422 13:58:10.116760 134671428339520 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0422 13:58:10.116831 134671428339520 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a66c47c4680> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0422 13:58:11.192721 134671428339520 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints I0422 13:58:11.195872 134671428339520 checkpoint_manager.py:929] [process=6][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=5), 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_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a677430f620> I0422 13:58:11.195988 134671428339520 checkpointing.py:302] Checkpoint manager created! I0422 13:58:12.139895 134671428339520 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 13:58:12.140000 134671428339520 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 13:58:12.521825 134671428339520 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 13:58:12.521918 134671428339520 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 13:58:12.538378 134671428339520 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 13:58:12.538435 134671428339520 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 13:58:12.561985 134671428339520 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 13:58:12.562050 134671428339520 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 13:58:12.578470 134671428339520 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 13:58:12.578536 134671428339520 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 13:58:12.594849 134671428339520 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 13:58:12.594910 134671428339520 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 13:58:12.619540 134671428339520 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 13:58:12.619609 134671428339520 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 13:58:12.641251 134671428339520 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 13:58:12.641316 134671428339520 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 13:58:12.853197 134671428339520 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 13:58:12.853305 134671428339520 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0422 13:58:14.273456 134671428339520 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0422 13:58:14.273600 134671428339520 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) I0422 13:58:14.273673 134671428339520 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) I0422 13:58:14.273734 134671428339520 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') I0422 13:58:14.273786 134671428339520 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) I0422 13:58:14.273827 134671428339520 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) I0422 13:58:14.273881 134671428339520 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) I0422 13:58:14.273935 134671428339520 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') I0422 13:58:14.273974 134671428339520 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) I0422 13:58:14.274008 134671428339520 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) I0422 13:58:14.274055 134671428339520 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0422 13:58:14.274115 134671428339520 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0422 13:58:14.765964 134671428339520 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0422 13:58:14.766059 134671428339520 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0422 13:58:14.781554 134671428339520 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0422 13:58:14.781618 134671428339520 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0422 13:58:25.763895 134671428339520 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. I0422 13:58:25.764755 134671428339520 metric_logger.py:301] number parameters: 1.104 billion I0422 13:58:37.102130 134671428339520 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0422 13:58:37.255749 134671428339520 checkpointing.py:798] Waited 0.15361642837524414 seconds for step 0 to finish before starting checkpointing. I0422 13:58:37.258148 134671428339520 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 13:58:37.259804 134671428339520 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0422 13:58:37.261577 134671428339520 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints/0. I0422 13:58:37.607976 134671428339520 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 13:58:37.608952 134671428339520 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0422 13:58:37.609008 134671428339520 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 I0422 13:58:37.880941 134671428339520 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.273038s I0422 13:58:37.881132 134671428339520 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.539 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2784695625305176 s) (per-host) I0422 13:58:37.881183 134671428339520 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.278532s (batch_requests_ready=0.002217s, total_serialization_initiated=0.276233s, others=0.000082s) I0422 13:58:37.881279 134671428339520 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.282597s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.282579) I0422 13:58:37.882058 134671428339520 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.62 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints/0. I0422 13:58:37.882413 134546544645888 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 14:18:37.882370 I0422 13:58:38.312368 134671428339520 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 13:58:38.312763 134541154965248 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 13:58:38.312929 134671428339520 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776866317.25813, 'wait_for_prev_duration_secs': 5.91278076171875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776866317.2598424, 'checkpointer_blocking_duration_secs': 0.6227395534515381, 'get_old_steps_start_time': 1776866317.8826096, 'get_old_steps_duration_secs': 3.075599670410156e-05, 'checkpoint_manager_blocking_start_time': 1776866317.2561738, 'checkpoint_manager_blocking_duration_secs': 1.0567119121551514} I0422 13:58:38.313036 134671428339520 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0422 13:58:38.313086 134671428339520 max_utils.py:750] Memstats: After params initialized: I0422 13:58:38.313158 134671428339520 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0422 13:58:38.313191 134671428339520 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0422 13:58:38.313218 134671428339520 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0422 13:58:38.313248 134671428339520 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0422 13:58:38.633767 134671428339520 metric_logger.py:196] completed step: 0, seconds: 11.337, TFLOP/s/device: 1.198, Tokens/s/device: 180.643, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0422 13:58:38.800501 134671428339520 metric_logger.py:196] completed step: 1, seconds: 1.530, TFLOP/s/device: 8.880, Tokens/s/device: 1338.446, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0422 13:58:39.242561 134671428339520 metric_logger.py:196] completed step: 2, seconds: 0.020, TFLOP/s/device: 692.408, Tokens/s/device: 104367.324, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0422 13:58:39.389931 134671428339520 metric_logger.py:196] completed step: 3, seconds: 0.442, TFLOP/s/device: 30.714, Tokens/s/device: 4629.525, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0422 13:58:39.394886 134671428339520 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0422 13:58:39.684129 134671428339520 checkpointing.py:798] Waited 0.28919029235839844 seconds for step 5 to finish before starting checkpointing. I0422 13:58:39.688070 134671428339520 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 13:58:40.863699 2795 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 13:58:42.934666 134541163357952 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints/0/items/array_metadatas/process_6 I0422 13:59:14.996809 134546544645888 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 42.242 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.39411997795105 s) (per-host) I0422 13:59:14.996917 134546544645888 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.114391s. I0422 13:59:22.940849 134546544645888 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 45.058306s. I0422 13:59:22.941073 134541154965248 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 13:59:22.941160 134541154965248 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 13:59:22.941210 134541154965248 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 13:59:22.942969 134541154965248 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 13:59:22.943084 134671428339520 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0422 13:59:22.943171 134671428339520 checkpoint_manager.py:1452] Waiting for previous save to complete took 43.255105 seconds. If this number is high, consider checkpointing less frequently. I0422 13:59:22.945189 134671428339520 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 5 I0422 13:59:22.947633 134671428339520 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints/5. I0422 13:59:23.261826 134671428339520 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0422 13:59:23.261919 134671428339520 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 I0422 13:59:23.302666 134671428339520 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.041923s I0422 13:59:23.302816 134671428339520 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 34.084 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0452580451965332 s) (per-host) I0422 13:59:23.302865 134671428339520 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045316s (batch_requests_ready=0.001749s, total_serialization_initiated=0.043502s, others=0.000066s) I0422 13:59:23.302946 134671428339520 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.050371s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.050354) I0422 13:59:23.303683 134671428339520 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.36 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints/5. I0422 13:59:23.303979 134541154965248 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 14:19:23.303949 I0422 13:59:23.306384 134671428339520 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0422 13:59:23.306622 134541163357952 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 13:59:23.306751 134671428339520 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776866319.6880407, 'wait_for_prev_duration_secs': 43.25510501861572, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776866362.9452262, 'checkpointer_blocking_duration_secs': 0.35890841484069824, 'get_old_steps_start_time': 1776866363.3041584, 'get_old_steps_duration_secs': 3.24249267578125e-05, 'checkpoint_manager_blocking_start_time': 1776866319.6847742, 'checkpoint_manager_blocking_duration_secs': 43.62193989753723} I0422 13:59:23.306870 134671428339520 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0422 13:59:23.307708 134671428339520 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.911, Tokens/s/device: 13401.650, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0422 13:59:23.313830 134671428339520 metric_logger.py:196] completed step: 5, seconds: 0.148, TFLOP/s/device: 91.993, Tokens/s/device: 13866.132, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0422 13:59:28.409300 134522838431488 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints/5/items/array_metadatas/process_6 I0422 13:59:45.914157 134671428339520 metric_logger.py:196] completed step: 6, seconds: 43.917, TFLOP/s/device: 0.309, Tokens/s/device: 46.633, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0422 13:59:46.061655 134671428339520 metric_logger.py:196] completed step: 7, seconds: 22.454, TFLOP/s/device: 0.605, Tokens/s/device: 91.208, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0422 13:59:46.208935 134671428339520 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.129, Tokens/s/device: 13434.530, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0422 13:59:46.355486 134671428339520 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0422 13:59:46.356179 134671428339520 checkpointing.py:798] Waited 0.0007131099700927734 seconds for step 9 to finish before starting checkpointing. I0422 13:59:46.358346 134671428339520 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 14:00:05.841948 134541154965248 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.093 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.584349393844604 s) (per-host) I0422 14:00:05.842072 134541154965248 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.537979s. I0422 14:00:13.028644 134541154965248 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.724534s. I0422 14:00:13.028909 134541163357952 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 14:00:13.029033 134541163357952 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 14:00:13.029081 134541163357952 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0422 14:00:13.030713 134541163357952 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0422 14:00:13.030944 134671428339520 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0422 14:00:13.031079 134671428339520 checkpoint_manager.py:1452] Waiting for previous save to complete took 26.672748 seconds. If this number is high, consider checkpointing less frequently. I0422 14:00:13.032639 134671428339520 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0422 14:00:13.034619 134671428339520 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints/9. I0422 14:00:13.357681 134671428339520 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0422 14:00:13.357772 134671428339520 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 I0422 14:00:13.400062 134671428339520 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043169s I0422 14:00:13.400240 134671428339520 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 33.123 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04657149314880371 s) (per-host) I0422 14:00:13.400297 134671428339520 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.046639s (batch_requests_ready=0.001652s, total_serialization_initiated=0.044914s, others=0.000073s) I0422 14:00:13.400385 134671428339520 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.050766s (all_items=0.000012s, per_item={'items': '0.00001240'}, temp_paths=0.050754) I0422 14:00:13.401136 134671428339520 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.37 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints/9. I0422 14:00:13.401405 134545451022080 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 14:20:13.401373 I0422 14:00:13.403378 134671428339520 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 14:00:13.403730 134541163357952 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 14:00:13.403888 134671428339520 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776866386.3583064, 'wait_for_prev_duration_secs': 26.67274808883667, 'time_between_consecutive_saves_sec': 23.41529893875122, 'checkpointer_blocking_start_time': 1776866413.0326793, 'checkpointer_blocking_duration_secs': 0.36882996559143066, 'get_old_steps_start_time': 1776866413.401527, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1776866386.356431, 'checkpoint_manager_blocking_duration_secs': 27.047426462173462} I0422 14:00:13.403998 134671428339520 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0422 14:00:13.404039 134671428339520 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 14:00:18.970577 134522838431488 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/linen_xpk_feat_nnx_post_train_fixes_20260422_123906_16_zarr3_true/checkpoints/9/items/array_metadatas/process_6 I0422 14:00:55.616734 134545451022080 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.375 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.26302719116211 s) (per-host) I0422 14:00:55.616851 134545451022080 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.214843s. I0422 14:01:04.645664 134545451022080 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.243640s. I0422 14:01:04.645948 134541163357952 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 14:01:04.646072 134541163357952 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 14:01:04.646144 134541163357952 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 14:01:04.647765 134541163357952 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 14:01:04.647930 134671428339520 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0422 14:01:04.648077 134671428339520 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 14:01:04.649038 134671428339520 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.236, Tokens/s/device: 13902.749, total_weights: 65536, loss: 7.181, lm_loss: 7.181, perplexity: 1314.495 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Wed Apr 22 14:01:15 UTC 2026 EXIT_CODE=0