MaxView

XPK Start: Wed Apr 22 14:26:47 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
`rope_parameters`'s factor field must be a float >= 1, got 40
`rope_parameters`'s beta_fast field must be a float, got 32
`rope_parameters`'s beta_slow field must be a float, got 1
DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config. 
2026-04-22 14:27:12.288100: 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 14:27:12.501736 135203305785152 max_utils.py:273] Attempting to initialize the jax distributed system...
I0422 14:27:21.542769 135203305785152 distributed.py:149] Starting JAX distributed service on [::]:8482
I0422 14:27:21.545088 135203305785152 distributed.py:172] Connecting to JAX distributed service on mt-06-grad-accum-ea4yx-slice-job-0-0.mt-06-grad-accum-ea4yx:8482
I0422 14:27:22.975564 135203305785152 max_utils.py:284] Jax distributed system initialized!
I0422 14:27:29.198854 135203305785152 max_utils.py:800] System Information: Jax Version: 0.9.2
I0422 14:27:29.198961 135203305785152 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0422 14:27:29.199001 135203305785152 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 14:27:29.199037 135203305785152 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0422 14:27:29.893210 135203305785152 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0422 14:27:30.007065 135203305785152 checkpointing.py:688] Setting up checkpoint logger...
I0422 14:27:30.007179 135203305785152 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0422 14:27:30.007224 135203305785152 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0422 14:27:30.007425 135203305785152 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 0x7af6c154cbc0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0422 14:27:32.845560 135203305785152 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0422 14:27:32.845807 135203305785152 checkpoint_manager.py:708] [process=5][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ae298528440>}, handler_registry=None
I0422 14:27:32.846045 135203305785152 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ae298528440>` 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 14:27:32.846094 135203305785152 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ae1b055f830>` 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 14:27:32.846130 135203305785152 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 0x7ae298528440>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ae298528440>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ae1b055f830>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ae1b055f830>}).
I0422 14:27:32.846453 135203305785152 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0422 14:27:32.846525 135203305785152 async_checkpointer.py:192] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ae1b0653600> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0422 14:27:34.313142 135203305785152 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_06_grad_accum/checkpoints
I0422 14:27:34.316476 135203305785152 checkpoint_manager.py:929] [process=5][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_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_06_grad_accum/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ae1b06bd040>
I0422 14:27:34.316590 135203305785152 checkpointing.py:302] Checkpoint manager created!
I0422 14:27:36.059727 135203305785152 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0422 14:27:36.059846 135203305785152 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0422 14:27:37.465697 135203305785152 maxtext_utils.py:1889]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0422 14:27:37.465812 135203305785152 maxtext_utils.py:1889]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0422 14:27:37.465863 135203305785152 maxtext_utils.py:1889]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0422 14:27:37.465905 135203305785152 maxtext_utils.py:1889]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0422 14:27:37.465941 135203305785152 maxtext_utils.py:1889]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0422 14:27:37.465975 135203305785152 maxtext_utils.py:1889]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0422 14:27:37.466011 135203305785152 maxtext_utils.py:1889]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 14:27:37.466046 135203305785152 maxtext_utils.py:1889]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0422 14:27:37.466080 135203305785152 maxtext_utils.py:1889]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 14:27:37.466113 135203305785152 maxtext_utils.py:1889]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 14:27:37.466145 135203305785152 maxtext_utils.py:1889]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0422 14:27:37.466176 135203305785152 maxtext_utils.py:1889]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0422 14:27:37.852871 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048]............................................... Unknown.
I0422 14:27:37.852961 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048]............................................... (None,).
I0422 14:27:37.868153 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,7168]....................................... Unknown.
I0422 14:27:37.868226 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,7168]....................................... ('fsdp', None, None).
I0422 14:27:37.898461 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[7168,16,2048]....................................... Unknown.
I0422 14:27:37.898527 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[7168,16,2048]....................................... (None, None, 'fsdp').
I0422 14:27:37.913659 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16]............................................ Unknown.
I0422 14:27:37.913713 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16]............................................ (None, None).
I0422 14:27:37.943708 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,16,128]..................................... Unknown.
I0422 14:27:37.943775 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,16,128]..................................... ('fsdp', None, None, None).
I0422 14:27:37.958873 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[16,16,128,2048]..................................... Unknown.
I0422 14:27:37.958935 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[16,16,128,2048]..................................... (None, None, None, 'fsdp').
I0422 14:27:38.004286 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,32000]......................................... Unknown.
I0422 14:27:38.004355 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,32000]......................................... ('fsdp', None).
I0422 14:27:38.019478 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[32000,2048]......................................... Unknown.
I0422 14:27:38.019538 135203305785152 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[32000,2048]......................................... (None, 'fsdp').
I0422 14:27:38.243932 135203305785152 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 14:27:38.244076 135203305785152 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 14:27:38.250823 135203305785152 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 14:27:38.250921 135203305785152 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 14:27:38.267886 135203305785152 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 14:27:38.268018 135203305785152 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 14:27:38.284307 135203305785152 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 14:27:38.284435 135203305785152 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 14:27:38.315305 135203305785152 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 14:27:38.315447 135203305785152 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 14:27:38.331775 135203305785152 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 14:27:38.331898 135203305785152 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 14:27:38.348119 135203305785152 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 14:27:38.348243 135203305785152 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 14:27:38.376093 135203305785152 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0422 14:27:38.376248 135203305785152 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 14:27:38.398344 135203305785152 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0422 14:27:38.398453 135203305785152 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0422 14:27:49.554684 135203305785152 max_utils.py:791] Total memory size: 1.7 GB, Output size: 0.4 GB, Temp size: 1.3 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0422 14:27:49.557080 135203305785152 metric_logger.py:301] number parameters: 1.104 billion
I0422 14:27:51.232679 135203305785152 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0422 14:28:03.124161 135203305785152 checkpointing.py:798] Waited 11.891479253768921 seconds for step 0 to finish before starting checkpointing.
I0422 14:28:03.126497 135203305785152 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 14:28:03.128142 135203305785152 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 0
I0422 14:28:03.129568 135203305785152 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_06_grad_accum/checkpoints/0.
I0422 14:28:03.690061 135203305785152 signaling_client.py:364] Using JaxDistributedSignalingClient
I0422 14:28:03.691085 135203305785152 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0422 14:28:03.691141 135203305785152 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 14:28:03.965694 135203305785152 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.276406s
I0422 14:28:03.965859 135203305785152 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 5.401 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2855956554412842 s) (per-host)
I0422 14:28:03.965910 135203305785152 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.285657s (batch_requests_ready=0.003158s, total_serialization_initiated=0.282430s, others=0.000069s)
I0422 14:28:03.966003 135203305785152 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.289754s (all_items=0.000018s, per_item={'items': '0.00001812'}, temp_paths=0.289736)
I0422 14:28:03.966813 135203305785152 event_tracking.py:125] [process=5] [async] Finished blocking save in 0.84 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_06_grad_accum/checkpoints/0.
I0422 14:28:03.967146 135076067120896 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 14:48:03.967108
I0422 14:28:03.978495 135203305785152 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 14:28:03.978825 135074464311040 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save.
I0422 14:28:03.978982 135203305785152 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776868083.1264799, 'wait_for_prev_duration_secs': 6.031990051269531e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776868083.1281807, 'checkpointer_blocking_duration_secs': 0.839123010635376, 'get_old_steps_start_time': 1776868083.9673293, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1776868083.1245446, 'checkpoint_manager_blocking_duration_secs': 0.8543980121612549}
I0422 14:28:03.979085 135203305785152 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0422 14:28:03.979150 135203305785152 max_utils.py:750] 
Memstats: After params initialized:
I0422 14:28:03.979202 135203305785152 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_18(process=5,(2,4,0,0))
I0422 14:28:03.979234 135203305785152 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_19(process=5,(3,4,0,0))
I0422 14:28:03.979260 135203305785152 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_22(process=5,(2,5,0,0))
I0422 14:28:03.979289 135203305785152 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_23(process=5,(3,5,0,0))
I0422 14:28:04.290932 135203305785152 metric_logger.py:196] completed step: 0, seconds: 1.675, TFLOP/s/device: 32.455, Tokens/s/device: 4892.005, total_weights: 262144, loss: 10.876, lm_loss: 10.876, perplexity: 52883.598
I0422 14:28:04.909713 135203305785152 metric_logger.py:196] completed step: 1, seconds: 13.056, TFLOP/s/device: 4.163, Tokens/s/device: 627.432, total_weights: 262144, loss: 10.876, lm_loss: 10.876, perplexity: 52883.598
I0422 14:28:05.488130 135203305785152 metric_logger.py:196] completed step: 2, seconds: 0.042, TFLOP/s/device: 1301.667, Tokens/s/device: 196201.471, total_weights: 262144, loss: 10.561, lm_loss: 10.561, perplexity: 38601.980
I0422 14:28:06.066463 135203305785152 metric_logger.py:196] completed step: 3, seconds: 0.585, TFLOP/s/device: 92.966, Tokens/s/device: 14012.809, total_weights: 262144, loss: 10.270, lm_loss: 10.270, perplexity: 28865.557
I0422 14:28:07.183753    2588 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0422 14:28:07.223763 135203305785152 metric_logger.py:196] completed step: 4, seconds: 0.579, TFLOP/s/device: 93.897, Tokens/s/device: 14153.250, total_weights: 262144, loss: 10.021, lm_loss: 10.021, perplexity: 22497.775
I0422 14:28:07.231209 135203305785152 metric_logger.py:196] completed step: 5, seconds: 0.578, TFLOP/s/device: 94.026, Tokens/s/device: 14172.692, total_weights: 262144, loss: 9.819, lm_loss: 9.819, perplexity: 18383.324
I0422 14:28:09.696951 135075542849280 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_06_grad_accum/checkpoints/0/items/array_metadatas/process_5
I0422 14:28:41.680629 135076067120896 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 41.568 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.000327825546265 s) (per-host)
I0422 14:28:41.680764 135076067120896 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.713504s.
I0422 14:28:50.814447 135076067120896 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 46.847171s.
I0422 14:28:50.814725 135074464311040 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 14:28:50.814848 135074464311040 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save.
I0422 14:28:50.814899 135074464311040 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0422 14:28:51.899577 135074464311040 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0422 14:28:52.801752 135203305785152 metric_logger.py:196] completed step: 6, seconds: 1.158, TFLOP/s/device: 46.944, Tokens/s/device: 7075.910, total_weights: 262144, loss: 9.666, lm_loss: 9.666, perplexity: 15775.117
I0422 14:28:53.380095 135203305785152 metric_logger.py:196] completed step: 7, seconds: 44.993, TFLOP/s/device: 1.208, Tokens/s/device: 182.072, total_weights: 262144, loss: 9.560, lm_loss: 9.560, perplexity: 14192.143
I0422 14:28:53.958381 135203305785152 metric_logger.py:196] completed step: 8, seconds: 0.585, TFLOP/s/device: 92.948, Tokens/s/device: 14010.101, total_weights: 262144, loss: 9.495, lm_loss: 9.495, perplexity: 13292.714
I0422 14:28:54.537483 135203305785152 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0422 14:28:54.538175 135203305785152 checkpointing.py:798] Waited 0.0007028579711914062 seconds for step 9 to finish before starting checkpointing.
I0422 14:28:54.540324 135203305785152 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 14:28:54.541919 135203305785152 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 9
I0422 14:28:54.543406 135203305785152 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_06_grad_accum/checkpoints/9.
I0422 14:28:54.838001 135203305785152 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0422 14:28:54.838100 135203305785152 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 14:28:54.871636 135203305785152 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.035206s
I0422 14:28:54.871815 135203305785152 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 36.578 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04217243194580078 s) (per-host)
I0422 14:28:54.871867 135203305785152 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.042233s (batch_requests_ready=0.002595s, total_serialization_initiated=0.039569s, others=0.000069s)
I0422 14:28:54.871962 135203305785152 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.046258s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.046242)
I0422 14:28:54.872698 135203305785152 event_tracking.py:125] [process=5] [async] Finished blocking save in 0.33 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_06_grad_accum/checkpoints/9.
I0422 14:28:54.873041 135074464311040 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 14:48:54.873000
I0422 14:28:54.875020 135203305785152 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 14:28:54.875276 135075553470208 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save.
I0422 14:28:54.875407 135203305785152 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776868134.5402935, 'wait_for_prev_duration_secs': 7.200241088867188e-05, 'time_between_consecutive_saves_sec': 2.640605926513672, 'checkpointer_blocking_start_time': 1776868134.5419571, 'checkpointer_blocking_duration_secs': 0.3312356472015381, 'get_old_steps_start_time': 1776868134.873217, 'get_old_steps_duration_secs': 2.6702880859375e-05, 'checkpoint_manager_blocking_start_time': 1776868134.538397, 'checkpoint_manager_blocking_duration_secs': 0.3369779586791992}
I0422 14:28:54.875576 135203305785152 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0422 14:28:54.875623 135203305785152 checkpoint_manager.py:2020] [process=5][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 14:29:00.396253 135069121369856 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_06_grad_accum/checkpoints/9/items/array_metadatas/process_5
I0422 14:29:37.389490 135074464311040 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 37.115 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.559807538986206 s) (per-host)
I0422 14:29:37.389619 135074464311040 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.516463s.
I0422 14:29:47.106419 135074464311040 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 52.233247s.
I0422 14:29:47.106734 135075553470208 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 14:29:47.106858 135075553470208 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save.
I0422 14:29:47.106904 135075553470208 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0422 14:29:47.108441 135075553470208 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0422 14:29:47.108613 135203305785152 checkpoint_manager.py:2032] [process=5][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0422 14:29:47.108794 135203305785152 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 14:29:47.109774 135203305785152 metric_logger.py:196] completed step: 9, seconds: 0.578, TFLOP/s/device: 94.032, Tokens/s/device: 14173.525, total_weights: 262144, loss: 9.457, lm_loss: 9.457, perplexity: 12793.173
Per train step:
 Total TFLOPs: 54.35 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Wed Apr 22 14:29:57 UTC 2026
EXIT_CODE=0