Skip to content
Closed
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 3rdparty/Megatron-Bridge-workspace/Megatron-Bridge
Submodule Megatron-Bridge updated 68 files
+4 −0 docker/Dockerfile.ci
+2 −0 docker/common/import_check_skip.txt
+2 −0 docs/training/cuda-graphs.md
+67 −30 examples/model_verification_cards/deepseek-v3/card.yaml
+19 −18 examples/model_verification_cards/gemma-3-4b-it/card.yaml
+21 −22 examples/model_verification_cards/gemma-4-26b-a4b-it/card.yaml
+15 −9 examples/model_verification_cards/gpt-oss-120b/card.yaml
+14 −11 examples/model_verification_cards/gpt-oss-20b/card.yaml
+16 −16 examples/model_verification_cards/k-exaone-2/card.yaml
+27 −18 examples/model_verification_cards/minimax-m3/card.yaml
+80 −26 examples/model_verification_cards/nemotron-3-nano-30b-a3b/card.yaml
+18 −11 examples/model_verification_cards/nemotron-3-super-120b-a12b/card.yaml
+29 −25 examples/model_verification_cards/nemotron-3-ultra-550b-a55b/card.yaml
+1 −0 scripts/performance/dump_perf_configs.py
+9 −1 scripts/performance/utils/overrides.py
+15 −5 scripts/training/README.md
+14 −0 skills/nemo-mbridge-perf-activation-recompute/SKILL.md
+7 −0 skills/nemo-mbridge-perf-activation-recompute/card.yaml
+28 −2 skills/nemo-mbridge-perf-memory-tuning/SKILL.md
+3 −0 skills/nemo-mbridge-perf-memory-tuning/card.yaml
+12 −2 skills/nemo-mbridge-perf-moe-dispatcher-selection/SKILL.md
+8 −1 skills/nemo-mbridge-perf-moe-dispatcher-selection/card.yaml
+51 −10 skills/nemo-mbridge-perf-moe-optimization-workflow/SKILL.md
+4 −1 skills/nemo-mbridge-perf-moe-optimization-workflow/card.yaml
+36 −4 skills/nemo-mbridge-recipe-recommender/SKILL.md
+15 −0 skills/nemo-mbridge-recipe-recommender/evals/evals.json
+13 −9 src/megatron/bridge/data/loaders.py
+14 −3 src/megatron/bridge/data/packing/gpt_sft.py
+4 −0 src/megatron/bridge/data/packing/in_batch.py
+7 −0 src/megatron/bridge/data/sources/hf.py
+24 −0 src/megatron/bridge/data/sources/hf_adapters.py
+1 −0 src/megatron/bridge/models/conversion/auto_bridge.py
+49 −1 src/megatron/bridge/models/deepseek/deepseek_v3_bridge.py
+25 −6 src/megatron/bridge/models/hf_pretrained/state.py
+22 −14 src/megatron/bridge/models/qwen_audio/qwen2_audio_bridge.py
+1 −0 src/megatron/bridge/perf_recipes/nemotronh/__init__.py
+34 −0 src/megatron/bridge/perf_recipes/nemotronh/vr200/nemotronh.py
+33 −16 src/megatron/bridge/recipes/qwen_vl/h100/qwen35_vl.py
+37 −0 src/megatron/bridge/recipes/utils/dataset_utils.py
+45 −0 src/megatron/bridge/training/checkpointing.py
+12 −2 src/megatron/bridge/training/config.py
+3 −2 src/megatron/bridge/training/gpt_step.py
+7 −6 src/megatron/bridge/training/setup.py
+16 −0 src/megatron/bridge/training/utils/omegaconf_utils.py
+17 −0 src/megatron/bridge/training/utils/packed_seq_utils.py
+24 −3 tests/functional_tests/test_groups/data/test_loaders.py
+56 −0 tests/unit_tests/bridge/training/test_setup.py
+21 −4 tests/unit_tests/data/datasets/test_chat_template.py
+22 −1 tests/unit_tests/data/datasets/test_gpt_sft.py
+2 −0 tests/unit_tests/data/packing/test_in_batch.py
+25 −0 tests/unit_tests/data/sources/test_hf.py
+50 −0 tests/unit_tests/data/sources/test_hf_adapters.py
+49 −0 tests/unit_tests/models/deepseek/test_deepseek_bridges.py
+47 −4 tests/unit_tests/models/hf_pretrained/test_state_dict.py
+70 −0 tests/unit_tests/models/qwen_audio/test_qwen2_audio_bridge.py
+10 −2 tests/unit_tests/models/test_auto_bridge.py
+28 −0 tests/unit_tests/recipes/nemotronh/test_nemotron_3_ultra.py
+19 −2 tests/unit_tests/recipes/qwen_vl/test_qwen35_vl_recipes.py
+73 −0 tests/unit_tests/recipes/test_override_precedence.py
+1 −0 tests/unit_tests/recipes/test_perf_recipe_environment.py
+2 −0 tests/unit_tests/recipes/utils/test_dataset_utils.py
+24 −0 tests/unit_tests/recipes/utils/test_sft_dataset_utils.py
+4 −2 tests/unit_tests/scripts/training/test_run_recipe.py
+171 −5 tests/unit_tests/training/test_checkpointing.py
+33 −2 tests/unit_tests/training/test_config.py
+4 −0 tests/unit_tests/training/test_gpt_step.py
+30 −0 tests/unit_tests/training/utils/test_omegaconf_utils.py
+2 −0 tests/unit_tests/training/utils/test_packed_seq_utils.py
Loading