Skip to content
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion .main.commit
Original file line number Diff line number Diff line change
@@ -1 +1 @@
64c3fb86ca2b1b875c0206500195b4de2977592a
471e3e38537869f332d9c97942198693009e15d8
2 changes: 1 addition & 1 deletion 3rdparty/Megatron-LM
Submodule Megatron-LM updated 53 files
+1 −1 .github/copy-pr-bot.yaml
+1 −1 docs/developer/contribute.md
+1 −1 megatron/core/distributed/fsdp/src/megatron_fsdp/package_info.py
+25 −16 megatron/core/inference/contexts/dynamic_context.py
+26 −16 megatron/core/inference/contexts/gpu_view.py
+5 −0 megatron/core/model_parallel_config.py
+3 −1 megatron/core/models/common/language_module/language_module.py
+1 −0 megatron/core/models/gpt/gpt_model.py
+3 −0 megatron/core/models/hybrid/hybrid_model.py
+14 −1 megatron/core/optimizer/__init__.py
+29 −14 megatron/core/optimizer/emerging_optimizers.py
+19 −4 megatron/core/optimizer/optimizer.py
+1 −1 megatron/core/package_info.py
+15 −0 megatron/core/pipeline_parallel/schedules.py
+23 −20 megatron/core/tensor_parallel/cross_entropy.py
+1 −0 megatron/core/tensor_parallel/layers.py
+1 −1 megatron/core/transformer/moe/README.md
+1 −0 megatron/core/transformer/moe/ops/__init__.py
+258 −0 megatron/core/transformer/moe/ops/paged_stash.py
+8 −255 megatron/core/transformer/moe/paged_stash.py
+31 −2 megatron/core/transformer/multi_token_prediction.py
+7 −0 megatron/training/arguments.py
+1 −1 tests/functional_tests/test_cases/gpt/gpt3_15b_8t_release_gb200/model_config.yaml
+1 −1 tests/functional_tests/test_cases/gpt/gpt3_15b_8t_release_sm_gb200/model_config.yaml
+1 −1 tests/functional_tests/test_cases/mixtral/deepseekv3_proxy_flex_tp1pp4emp16etp1cp1_release/model_config.yaml
+1 −1 ...s/functional_tests/test_cases/mixtral/deepseekv3_proxy_flex_tp1pp4emp16etp1cp1_release_sm/model_config.yaml
+1 −1 ...nctional_tests/test_cases/mixtral/deepseekv3_proxy_flex_tp2pp2emp16etp1cp1_gb_200_release/model_config.yaml
+1 −1 ...ional_tests/test_cases/mixtral/deepseekv3_proxy_flex_tp2pp2emp16etp1cp1_gb_200_release_sm/model_config.yaml
+17 −1 tests/performance_tests/shell_test_utils/compare_to_baseline.py
+66 −7 tests/performance_tests/shell_test_utils/run_perf_test.sh
+75 −35 tests/performance_tests/test_cases/gpt/gpt_16b_perf/baseline_values.json
+46 −44 tests/performance_tests/test_cases/gpt/gpt_583m_perf/baseline_values.json
+52 −0 tests/performance_tests/test_cases/gpt/gpt_583m_perf_gb200_4gpu/baseline_values.json
+27 −0 tests/performance_tests/test_cases/gpt/gpt_583m_perf_gb200_4gpu/model_config.yaml
+99 −47 tests/performance_tests/test_cases/hybrid/hybrid_2b_perf/baseline_values.json
+50 −48 tests/performance_tests/test_cases/hybrid/hybrid_nanov3_3b_perf/baseline_values.json
+52 −0 tests/performance_tests/test_cases/hybrid/hybrid_nanov3_3b_perf_gb200_4gpu/baseline_values.json
+28 −0 tests/performance_tests/test_cases/hybrid/hybrid_nanov3_3b_perf_gb200_4gpu/model_config.yaml
+47 −0 tests/test_utils/recipes/gb200/gpt-perf-dp4.yaml
+49 −0 tests/test_utils/recipes/gb200/gpt-perf.yaml
+47 −0 tests/test_utils/recipes/gb200/hybrid-perf-ep4.yaml
+48 −0 tests/test_utils/recipes/gb200/hybrid-perf.yaml
+28 −0 tests/test_utils/recipes/h100/unit-tests.yaml
+9 −1 tests/unit_tests/elastification/test_hybrid_flex_router.py
+110 −0 tests/unit_tests/inference/contexts/test_fused_kv_append_kernel.py
+8 −5 tests/unit_tests/inference/contexts/test_gpu_view.py
+1 −1 tests/unit_tests/models/mimo/test_mimo_1f1b_schedule.py
+76 −0 tests/unit_tests/tensor_parallel/test_cross_entropy.py
+13 −0 tests/unit_tests/tensor_parallel/test_tp_attrs_without_init.py
+17 −0 tests/unit_tests/test_emerging_optimizers.py
+17 −0 tests/unit_tests/test_model_parallel_config.py
+4 −0 tests/unit_tests/transformer/moe/test_routers.py
+107 −0 tests/unit_tests/transformer/test_multi_token_prediction.py
Loading
Loading