megatron-core 0.16.0rc0.dev130734__tar.gz → 0.16.0rc0.dev131530__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Potentially problematic release.

This version of megatron-core might be problematic. Click here for more details.

Files changed (358) hide show
  1. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/PKG-INFO +1 -1
  2. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/package_info.py +1 -1
  3. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/cuda_graphs.py +17 -6
  4. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron_core.egg-info/PKG-INFO +1 -1
  5. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/MANIFEST.in +0 -0
  6. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/README.md +0 -0
  7. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/README.md +0 -0
  8. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/__init__.py +0 -0
  9. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/activations.py +0 -0
  10. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/config.py +0 -0
  11. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/config_logger.py +0 -0
  12. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/__init__.py +0 -0
  13. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/bert_dataset.py +0 -0
  14. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/blended_dataset.py +0 -0
  15. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/blended_megatron_dataset_builder.py +0 -0
  16. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/blended_megatron_dataset_config.py +0 -0
  17. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/gpt_dataset.py +0 -0
  18. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/helpers.cpp +0 -0
  19. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/helpers.py +0 -0
  20. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/indexed_dataset.py +0 -0
  21. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/masked_dataset.py +0 -0
  22. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/megatron_dataset.py +0 -0
  23. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/megatron_tokenizer.py +0 -0
  24. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/multimodal_dataset.py +0 -0
  25. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/object_storage_utils.py +0 -0
  26. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/__init__.py +0 -0
  27. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/config/__init__.py +0 -0
  28. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/config/bert_embedders.py +0 -0
  29. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/config/config.py +0 -0
  30. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/config/gpt_chunk_datasets.py +0 -0
  31. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/config/tokenizers.py +0 -0
  32. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/db/__init__.py +0 -0
  33. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/db/build.py +0 -0
  34. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/db/dataset.py +0 -0
  35. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/db/utils.py +0 -0
  36. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/external_libs.py +0 -0
  37. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/index/__init__.py +0 -0
  38. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/index/build.py +0 -0
  39. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/index/factory.py +0 -0
  40. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/index/index.py +0 -0
  41. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/index/indexes/__init__.py +0 -0
  42. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/index/indexes/faiss_base.py +0 -0
  43. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/index/indexes/faiss_par_add.py +0 -0
  44. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/index/utils.py +0 -0
  45. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/index/validate.py +0 -0
  46. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/query/__init__.py +0 -0
  47. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/query/gpt_chunk_dataset.py +0 -0
  48. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/query/multi_split_gpt_dataset.py +0 -0
  49. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/query/query.py +0 -0
  50. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/query/retro_dataset.py +0 -0
  51. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/query/utils.py +0 -0
  52. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/retro/utils.py +0 -0
  53. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/t5_dataset.py +0 -0
  54. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/utils.py +0 -0
  55. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/datasets/utils_s3.py +0 -0
  56. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/__init__.py +0 -0
  57. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/core.py +0 -0
  58. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/dict_utils.py +0 -0
  59. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/exchange_utils.py +0 -0
  60. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/mapping.py +0 -0
  61. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/optimizer.py +0 -0
  62. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/serialization.py +0 -0
  63. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/state_dict_utils.py +0 -0
  64. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/__init__.py +0 -0
  65. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/async_utils.py +0 -0
  66. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/base.py +0 -0
  67. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/cached_metadata_filesystem_reader.py +0 -0
  68. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/checkpointable.py +0 -0
  69. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/common.py +0 -0
  70. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/filesystem_async.py +0 -0
  71. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/fully_parallel.py +0 -0
  72. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/resharding.py +0 -0
  73. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/state_dict_saver.py +0 -0
  74. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/tensorstore.py +0 -0
  75. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/torch.py +0 -0
  76. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/two_stage.py +0 -0
  77. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/strategies/zarr.py +0 -0
  78. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/tensor_aware_state_dict.py +0 -0
  79. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/utils.py +0 -0
  80. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/dist_checkpointing/validation.py +0 -0
  81. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/__init__.py +0 -0
  82. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/data_parallel_base.py +0 -0
  83. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/distributed_data_parallel.py +0 -0
  84. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/distributed_data_parallel_config.py +0 -0
  85. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/finalize_model_grads.py +0 -0
  86. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/fsdp/__init__.py +0 -0
  87. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/fsdp/mcore_fsdp_adapter.py +0 -0
  88. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/fsdp/src/__init__.py +0 -0
  89. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/fsdp/src/megatron_fsdp/__init__.py +0 -0
  90. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/fsdp/src/megatron_fsdp/distributed_data_parallel_config.py +0 -0
  91. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/fsdp/src/megatron_fsdp/fully_shard.py +0 -0
  92. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/fsdp/src/megatron_fsdp/megatron_fsdp.py +0 -0
  93. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/fsdp/src/megatron_fsdp/package_info.py +0 -0
  94. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/fsdp/src/megatron_fsdp/param_and_grad_buffer.py +0 -0
  95. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/fsdp/src/megatron_fsdp/uneven_dtensor.py +0 -0
  96. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/fsdp/src/megatron_fsdp/utils.py +0 -0
  97. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/param_and_grad_buffer.py +0 -0
  98. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/reduce_scatter_with_fp32_accumulation.py +0 -0
  99. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/torch_fully_sharded_data_parallel.py +0 -0
  100. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/distributed/torch_fully_sharded_data_parallel_config.py +0 -0
  101. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/energy_monitor.py +0 -0
  102. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/enums.py +0 -0
  103. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/__init__.py +0 -0
  104. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/data_type.py +0 -0
  105. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/export_config.py +0 -0
  106. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/model_type.py +0 -0
  107. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/__init__.py +0 -0
  108. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/engine_builder/__init__.py +0 -0
  109. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/engine_builder/trtllm_engine_builder.py +0 -0
  110. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/model_to_trllm_mapping/__init__.py +0 -0
  111. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/model_to_trllm_mapping/default_conversion_dict.py +0 -0
  112. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/trt_model_config.py +0 -0
  113. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/trt_model_type.py +0 -0
  114. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/trtllm_helper.py +0 -0
  115. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/trtllm_layers.py +0 -0
  116. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/trtllm_weights_converter/__init__.py +0 -0
  117. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/trtllm_weights_converter/distributed_trtllm_model_weights_converter.py +0 -0
  118. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/trtllm_weights_converter/single_device_trtllm_model_weights_converter.py +0 -0
  119. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/export/trtllm/trtllm_weights_converter/utils.py +0 -0
  120. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/extensions/__init__.py +0 -0
  121. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/extensions/kitchen.py +0 -0
  122. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/extensions/transformer_engine.py +0 -0
  123. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/extensions/transformer_engine_spec_provider.py +0 -0
  124. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fp4_utils.py +0 -0
  125. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fp8_utils.py +0 -0
  126. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/full_cuda_graph.py +0 -0
  127. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fusions/__init__.py +0 -0
  128. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fusions/fused_bias_dropout.py +0 -0
  129. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fusions/fused_bias_geglu.py +0 -0
  130. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fusions/fused_bias_gelu.py +0 -0
  131. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fusions/fused_bias_swiglu.py +0 -0
  132. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fusions/fused_cross_entropy.py +0 -0
  133. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fusions/fused_indices_converter.py +0 -0
  134. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fusions/fused_layer_norm.py +0 -0
  135. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fusions/fused_mla_yarn_rope_apply.py +0 -0
  136. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fusions/fused_pad_routing_map.py +0 -0
  137. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fusions/fused_softmax.py +0 -0
  138. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/fusions/fused_weighted_squared_relu.py +0 -0
  139. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/hyper_comm_grid.py +0 -0
  140. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/__init__.py +0 -0
  141. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/async_stream.py +0 -0
  142. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/common_inference_params.py +0 -0
  143. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/communication_utils.py +0 -0
  144. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/contexts/__init__.py +0 -0
  145. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/contexts/base_context.py +0 -0
  146. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/contexts/dynamic_block_allocator.py +0 -0
  147. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/contexts/dynamic_context.py +0 -0
  148. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/contexts/fused_kv_append_kernel.py +0 -0
  149. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/contexts/static_context.py +0 -0
  150. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/data_parallel_inference_coordinator.py +0 -0
  151. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/engines/__init__.py +0 -0
  152. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/engines/abstract_engine.py +0 -0
  153. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/engines/dynamic_engine.py +0 -0
  154. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/engines/mcore_engine.py +0 -0
  155. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/engines/static_engine.py +0 -0
  156. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/headers.py +0 -0
  157. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/inference_client.py +0 -0
  158. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/inference_request.py +0 -0
  159. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/model_inference_wrappers/__init__.py +0 -0
  160. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/model_inference_wrappers/abstract_model_inference_wrapper.py +0 -0
  161. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/model_inference_wrappers/gpt/__init__.py +0 -0
  162. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/model_inference_wrappers/gpt/gpt_inference_wrapper.py +0 -0
  163. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/model_inference_wrappers/inference_wrapper_config.py +0 -0
  164. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/model_inference_wrappers/multimodal/vlm_inference_wrapper.py +0 -0
  165. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/model_inference_wrappers/t5/__init__.py +0 -0
  166. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/model_inference_wrappers/t5/t5_inference_wrapper.py +0 -0
  167. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/sampling_params.py +0 -0
  168. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/scheduler.py +0 -0
  169. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/text_generation_controllers/__init__.py +0 -0
  170. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/text_generation_controllers/encoder_decoder_text_generation_controller.py +0 -0
  171. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/text_generation_controllers/simple_text_generation_controller.py +0 -0
  172. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/text_generation_controllers/text_generation_controller.py +0 -0
  173. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/text_generation_controllers/vlm_text_generation_controller.py +0 -0
  174. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/text_generation_server/__init__.py +0 -0
  175. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/text_generation_server/endpoints/common.py +0 -0
  176. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/text_generation_server/endpoints/completions.py +0 -0
  177. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/text_generation_server/run_mcore_engine.py +0 -0
  178. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/text_generation_server/text_generation_server.py +0 -0
  179. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/text_generation_server/tokenization.py +0 -0
  180. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/unified_memory.py +0 -0
  181. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference/utils.py +0 -0
  182. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/inference_params.py +0 -0
  183. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/jit.py +0 -0
  184. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/model_parallel_config.py +0 -0
  185. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/T5/__init__.py +0 -0
  186. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/T5/t5_model.py +0 -0
  187. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/T5/t5_spec.py +0 -0
  188. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/__init__.py +0 -0
  189. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/backends.py +0 -0
  190. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/bert/__init__.py +0 -0
  191. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/bert/bert_layer_specs.py +0 -0
  192. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/bert/bert_lm_head.py +0 -0
  193. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/bert/bert_model.py +0 -0
  194. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/bert/pooler.py +0 -0
  195. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/common/__init__.py +0 -0
  196. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/common/embeddings/__init__.py +0 -0
  197. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/common/embeddings/language_model_embedding.py +0 -0
  198. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/common/embeddings/relative_pos_embedding.py +0 -0
  199. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/common/embeddings/rope_utils.py +0 -0
  200. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/common/embeddings/rotary_pos_embedding.py +0 -0
  201. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/common/embeddings/yarn_rotary_pos_embedding.py +0 -0
  202. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/common/language_module/__init__.py +0 -0
  203. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/common/language_module/language_module.py +0 -0
  204. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/common/model_chunk_schedule_plan.py +0 -0
  205. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/common/vision_module/__init__.py +0 -0
  206. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/common/vision_module/vision_module.py +0 -0
  207. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/gpt/__init__.py +0 -0
  208. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/gpt/fine_grained_callables.py +0 -0
  209. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/gpt/gpt_layer_specs.py +0 -0
  210. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/gpt/gpt_model.py +0 -0
  211. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/gpt/heterogeneous/heterogeneous_layer_specs.py +0 -0
  212. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/gpt/moe_module_specs.py +0 -0
  213. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/huggingface/__init__.py +0 -0
  214. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/huggingface/clip_model.py +0 -0
  215. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/huggingface/module.py +0 -0
  216. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/huggingface/qwen_model.py +0 -0
  217. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/mamba/__init__.py +0 -0
  218. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/mamba/mamba_layer_specs.py +0 -0
  219. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/mamba/mamba_model.py +0 -0
  220. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/mimo/__init__.py +0 -0
  221. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/mimo/config/__init__.py +0 -0
  222. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/mimo/config/base_configs.py +0 -0
  223. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/mimo/model/__init__.py +0 -0
  224. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/mimo/model/base.py +0 -0
  225. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/mimo/submodules/audio.py +0 -0
  226. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/mimo/submodules/base.py +0 -0
  227. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/mimo/submodules/vision.py +0 -0
  228. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/multimodal/__init__.py +0 -0
  229. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/multimodal/context_parallel.py +0 -0
  230. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/multimodal/llava_model.py +0 -0
  231. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/multimodal/llava_spec.py +0 -0
  232. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/retro/__init__.py +0 -0
  233. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/retro/base_attention.py +0 -0
  234. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/retro/config.py +0 -0
  235. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/retro/decoder_attention.py +0 -0
  236. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/retro/decoder_spec.py +0 -0
  237. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/retro/encoder_attention.py +0 -0
  238. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/retro/encoder_spec.py +0 -0
  239. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/retro/model.py +0 -0
  240. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/retro/utils.py +0 -0
  241. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/vision/__init__.py +0 -0
  242. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/vision/clip_vit_model.py +0 -0
  243. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/vision/multimodal_projector.py +0 -0
  244. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/vision/radio.py +0 -0
  245. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/models/vision/vit_layer_specs.py +0 -0
  246. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/msc_utils.py +0 -0
  247. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/nccl_allocator.py +0 -0
  248. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/num_microbatches_calculator.py +0 -0
  249. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/optimizer/__init__.py +0 -0
  250. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/optimizer/clip_grads.py +0 -0
  251. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/optimizer/cpu_offloading/__init__.py +0 -0
  252. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/optimizer/cpu_offloading/hybrid_optimizer.py +0 -0
  253. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/optimizer/distrib_optimizer.py +0 -0
  254. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/optimizer/grad_scaler.py +0 -0
  255. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/optimizer/optimizer.py +0 -0
  256. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/optimizer/optimizer_config.py +0 -0
  257. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/optimizer_param_scheduler.py +0 -0
  258. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/packed_seq_params.py +0 -0
  259. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/parallel_state.py +0 -0
  260. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/pipeline_parallel/__init__.py +0 -0
  261. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/pipeline_parallel/bridge_communicator.py +0 -0
  262. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/pipeline_parallel/combined_1f1b.py +0 -0
  263. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/pipeline_parallel/p2p_communication.py +0 -0
  264. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/pipeline_parallel/schedules.py +0 -0
  265. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/pipeline_parallel/utils.py +0 -0
  266. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/post_training/__init__.py +0 -0
  267. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/post_training/modelopt/__init__.py +0 -0
  268. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/post_training/modelopt/gpt/__init__.py +0 -0
  269. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/post_training/modelopt/gpt/model_specs.py +0 -0
  270. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/post_training/modelopt/gpt/state_dict_hooks.py +0 -0
  271. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/post_training/modelopt/layers.py +0 -0
  272. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/post_training/modelopt/mamba/__init__.py +0 -0
  273. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/post_training/modelopt/mamba/model_specs.py +0 -0
  274. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/process_groups_config.py +0 -0
  275. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/quantization/__init__.py +0 -0
  276. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/quantization/quant_config.py +0 -0
  277. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/quantization/utils.py +0 -0
  278. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/requirements.txt +0 -0
  279. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/rerun_state_machine.py +0 -0
  280. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/safe_globals.py +0 -0
  281. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/ssm/__init__.py +0 -0
  282. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/ssm/mamba_block.py +0 -0
  283. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/ssm/mamba_context_parallel.py +0 -0
  284. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/ssm/mamba_hybrid_layer_allocation.py +0 -0
  285. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/ssm/mamba_layer.py +0 -0
  286. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/ssm/mamba_mixer.py +0 -0
  287. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/ssm/mlp_layer.py +0 -0
  288. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/ssm/triton_cache_manager.py +0 -0
  289. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tensor_parallel/__init__.py +0 -0
  290. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tensor_parallel/cross_entropy.py +0 -0
  291. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tensor_parallel/data.py +0 -0
  292. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tensor_parallel/layers.py +0 -0
  293. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tensor_parallel/mappings.py +0 -0
  294. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tensor_parallel/random.py +0 -0
  295. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tensor_parallel/utils.py +0 -0
  296. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/timers.py +0 -0
  297. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/__init__.py +0 -0
  298. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/base_tokenizer.py +0 -0
  299. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/megatron_tokenizer.py +0 -0
  300. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/__init__.py +0 -0
  301. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/libraries/__init__.py +0 -0
  302. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/libraries/abstract_tokenizer.py +0 -0
  303. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/libraries/bytelevel_tokenizer.py +0 -0
  304. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/libraries/chat_template.py +0 -0
  305. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/libraries/huggingface_tokenizer.py +0 -0
  306. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/libraries/megatron_hf_tokenizer.py +0 -0
  307. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/libraries/null_tokenizer.py +0 -0
  308. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/libraries/sentencepiece_tokenizer.py +0 -0
  309. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/libraries/tiktoken_tokenizer.py +0 -0
  310. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/models/__init__.py +0 -0
  311. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/models/bert_tokenizer.py +0 -0
  312. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/models/default_tokenizer.py +0 -0
  313. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/models/gpt_tokenizer.py +0 -0
  314. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/models/mamba_tokenizer.py +0 -0
  315. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/models/retro_tokenizer.py +0 -0
  316. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/models/t5_tokenizer.py +0 -0
  317. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/text_tokenizer.py +0 -0
  318. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/tokenizers/text/utils/build_tokenizer.py +0 -0
  319. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/__init__.py +0 -0
  320. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/attention.py +0 -0
  321. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/custom_layers/__init__.py +0 -0
  322. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/custom_layers/transformer_engine.py +0 -0
  323. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/dot_product_attention.py +0 -0
  324. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/enums.py +0 -0
  325. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/fsdp_dtensor_checkpoint.py +0 -0
  326. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/heterogeneous/heterogeneous_config.py +0 -0
  327. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/heterogeneous/linear_replacements.py +0 -0
  328. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/identity_op.py +0 -0
  329. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/mlp.py +0 -0
  330. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/module.py +0 -0
  331. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/moe/__init__.py +0 -0
  332. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/moe/experts.py +0 -0
  333. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/moe/fused_a2a.py +0 -0
  334. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/moe/grouped_gemm_util.py +0 -0
  335. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/moe/moe_layer.py +0 -0
  336. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/moe/moe_utils.py +0 -0
  337. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/moe/router.py +0 -0
  338. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/moe/shared_experts.py +0 -0
  339. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/moe/token_dispatcher.py +0 -0
  340. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/moe/upcycling_utils.py +0 -0
  341. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/multi_latent_attention.py +0 -0
  342. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/multi_token_prediction.py +0 -0
  343. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/pipeline_parallel_layer_layout.py +0 -0
  344. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/spec_utils.py +0 -0
  345. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/torch_layer_norm.py +0 -0
  346. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/torch_norm.py +0 -0
  347. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/transformer_block.py +0 -0
  348. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/transformer_config.py +0 -0
  349. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/transformer_layer.py +0 -0
  350. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/utils.py +0 -0
  351. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/utils.py +0 -0
  352. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron_core.egg-info/SOURCES.txt +0 -0
  353. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron_core.egg-info/dependency_links.txt +0 -0
  354. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron_core.egg-info/requires.txt +0 -0
  355. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron_core.egg-info/top_level.txt +0 -0
  356. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/pyproject.toml +0 -0
  357. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/setup.cfg +0 -0
  358. {megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/setup.py +0 -0
{megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: megatron-core
3
- Version: 0.16.0rc0.dev130734
3
+ Version: 0.16.0rc0.dev131530
4
4
  Summary: Megatron Core - a library for efficient and scalable training of transformer based models
5
5
  Author-email: NVIDIA <nemo-toolkit@nvidia.com>
6
6
  Maintainer-email: NVIDIA <nemo-toolkit@nvidia.com>
{megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/package_info.py RENAMED
@@ -4,7 +4,7 @@
4
4
  MAJOR = 0
5
5
  MINOR = 16
6
6
  PATCH = 0
7
- PRE_RELEASE = 'rc0.dev130734'
7
+ PRE_RELEASE = 'rc0.dev131530'
8
8
 
9
9
  # Use the following formatting: (major, minor, patch, pre-release)
10
10
  VERSION = (MAJOR, MINOR, PATCH, PRE_RELEASE)
{megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron/core/transformer/cuda_graphs.py RENAMED
@@ -26,6 +26,7 @@ from megatron.core.transformer.module import MegatronModule
26
26
  from megatron.core.transformer.transformer_config import TransformerConfig
27
27
  from megatron.core.utils import (
28
28
  get_attr_wrapped_model,
29
+ get_torch_version,
29
30
  is_te_min_version,
30
31
  log_on_each_pipeline_stage,
31
32
  log_single_rank,
@@ -58,6 +59,18 @@ _IS_GRAPH_CAPTURING = False
58
59
 
59
60
  logger = logging.getLogger(__name__)
60
61
 
62
+ # Freeze GC during capture.
63
+ # TODO (@lmcafee): remove all freeze-GC code once most users are on PyTorch 2.9+.
64
+ FREEZE_GC = os.getenv("CUDA_GRAPH_CAPTURE_FREEZE_GC") != "0"
65
+ try:
66
+ from packaging.version import Version as PkgVersion
67
+
68
+ FREEZE_GC_MAX_TORCH_VERSION = PkgVersion("2.9.0a0")
69
+ if get_torch_version() >= FREEZE_GC_MAX_TORCH_VERSION:
70
+ FREEZE_GC = False
71
+ except ImportError:
72
+ pass
73
+
61
74
 
62
75
  def is_graph_capturing():
63
76
  """Query if currently capturing."""
@@ -616,8 +629,7 @@ class _CudaGraphRunner(torch.nn.Module):
616
629
  'create_cudagraphs()'."""
617
630
 
618
631
  # Freeze GC, to speed up capture time ~15-20x.
619
- freeze_gc = os.getenv("CUDA_GRAPH_CAPTURE_FREEZE_GC") != "0"
620
- if freeze_gc:
632
+ if FREEZE_GC:
621
633
  gc.freeze()
622
634
 
623
635
  # save grads and other variables that may be affected by graph warmup
@@ -706,7 +718,7 @@ class _CudaGraphRunner(torch.nn.Module):
706
718
  restore_fp8_tensors([self.base_module], saved_fp8_tensors)
707
719
 
708
720
  # Unfreeze GC.
709
- if freeze_gc:
721
+ if FREEZE_GC:
710
722
  gc.unfreeze()
711
723
 
712
724
  # gc.collect() drops references to unreachable tensors created during capture,
@@ -721,8 +733,7 @@ class _CudaGraphRunner(torch.nn.Module):
721
733
  'create_cudagraphs()'."""
722
734
 
723
735
  # Freeze GC, to speed up capture time ~15-20x.
724
- freeze_gc = os.getenv("CUDA_GRAPH_CAPTURE_FREEZE_GC") != "0"
725
- if freeze_gc:
736
+ if FREEZE_GC:
726
737
  gc.freeze()
727
738
 
728
739
  self.bwd_graph = torch.cuda.CUDAGraph()
@@ -782,7 +793,7 @@ class _CudaGraphRunner(torch.nn.Module):
782
793
  self.static_grad_inputs = static_grad_inputs
783
794
 
784
795
  # Unfreeze GC.
785
- if freeze_gc:
796
+ if FREEZE_GC:
786
797
  gc.unfreeze()
787
798
 
788
799
  if self.is_first_layer:
{megatron_core-0.16.0rc0.dev130734 → megatron_core-0.16.0rc0.dev131530}/megatron_core.egg-info/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: megatron-core
3
- Version: 0.16.0rc0.dev130734
3
+ Version: 0.16.0rc0.dev131530
4
4
  Summary: Megatron Core - a library for efficient and scalable training of transformer based models
5
5
  Author-email: NVIDIA <nemo-toolkit@nvidia.com>
6
6
  Maintainer-email: NVIDIA <nemo-toolkit@nvidia.com>