Commit e84ecca
authored
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
1 | | - | |
| 1 | + | |
Submodule Megatron-LM updated 57 files
- .github/oncall_schedule.json+1-1
- .gitlab-ci.yml+16
- .gitlab/nemo-ci-triage.yml+14
- .gitlab/stages/02.test.yml+2-1
- .gitlab/stages/04.functional-tests.yml+12-2
- .gitlab/stages/06.triage.yml+95
- docker/Dockerfile.linting+1-1
- examples/inference/advanced/gpt_dynamic_inference.py+19-11
- examples/inference/utils.py+21-7
- megatron/core/datasets/indexed_dataset.py+8-28
- megatron/core/dist_checkpointing/core.py+7-19
- megatron/core/dist_checkpointing/serialization.py+3-10
- megatron/core/dist_checkpointing/strategies/common.py+4-17
- megatron/core/dist_checkpointing/validation.py+19-54
- megatron/core/distributed/fsdp/src/megatron_fsdp/experimental/__init__.py+2
- megatron/core/distributed/fsdp/src/megatron_fsdp/experimental/dbuffer.py+31-3
- megatron/core/distributed/fsdp/src/megatron_fsdp/experimental/module.py+1-1
- megatron/core/distributed/fsdp/src/megatron_fsdp/experimental/optimizer.py+108
- megatron/core/distributed/fsdp/src/megatron_fsdp/experimental/parameter_group.py+39-13
- megatron/core/extensions/transformer_engine.py+3
- megatron/core/inference/config.py+3
- megatron/core/inference/contexts/dynamic_context.py+105-40
- megatron/core/inference/engines/dynamic_engine.py+6-6
- megatron/core/inference/text_generation_controllers/text_generation_controller.py+524-223
- megatron/core/models/audio/__init__.py+2
- megatron/core/models/audio/audio_processor.py+338
- megatron/core/msc_utils.py+54-10
- megatron/core/packed_seq_params.py+1
- megatron/training/arguments.py+4-2
- megatron/training/checkpointing.py+18-36
- megatron/training/config/container.py+6-20
- megatron/training/config/inference_config.py+3-2
- megatron/training/distillation/utils_logits.py+2-6
- megatron/training/training.py+12-2
- megatron/training/utils/common_utils.py+3-3
- tests/performance_tests/client/static_benchmark.py+39-12
- tests/performance_tests/shell_test_utils/run_perf_test.sh+5
- tests/performance_tests/test_cases/gpt/gpt_583m_perf/baseline_values.json+28-24
- tests/test_utils/python_scripts/linear_ci.py+192
- tests/test_utils/python_scripts/notify.py+77-15
- tests/test_utils/test_ci_triage.py+335-9
- tests/unit_tests/distributed/mfsdp_v2/profiler_utils.py+55
- tests/unit_tests/distributed/mfsdp_v2/test_fully_shard.py+303-70
- tests/unit_tests/distributed/mfsdp_v2/test_optimizer.py+99
- tests/unit_tests/inference/contexts/test_dynamic_context.py+177-26
- tests/unit_tests/inference/engines/test_dynamic_engine.py+43
- tests/unit_tests/inference/engines/test_dynamic_engine_async_sched.py+4
- tests/unit_tests/inference/test_async_sched_output_metrics.py+42
- tests/unit_tests/inference/test_inference_config.py+6-4
- tests/unit_tests/inference/test_mtp_cuda_graph_inference.py+2-2
- tests/unit_tests/inference/text_generation_controllers/test_text_generation_controller.py+475-115
- tests/unit_tests/models/test_nemo_audio_processor.py+455
- tests/unit_tests/test_msc_utils.py+40
- tests/unit_tests/test_static_benchmark.py+106
- tests/unit_tests/training/config/test_container_base.py+2-2
- tests/unit_tests/transformer/test_cuda_graphs.py+108
- uv.lock+857-837
0 commit comments