.coveragerc
.gitignore
.gitmodules
.pre-commit-config.yaml
LICENSE
README.md
__init__.py
flagopen.png
pyproject.toml
run.py
setup.py
version.py
./__init__.py
./run.py
./setup.py
./version.py
.github/CODEOWNERS
.github/workflows/all-tests-metax.yml
.github/workflows/all-tests-nvidia.yml
.github/workflows/format.yml
.github/workflows/functional-tests-metax.yml
.github/workflows/functional-tests-nvidia.yml
.github/workflows/unit-tests-nvidia.yml
build_backend/__init__.py
build_backend/flagscale_build_backend.py
docker/Dockerfile.cuda
docker/Dockerfile.metax
docker/Dockerfile.ssh
docker/README.md
examples/aquila/__init__.py
examples/aquila/conf/inference.yaml
examples/aquila/conf/train.yaml
examples/aquila/conf/train_auto_tuner.yaml
examples/aquila/conf/train_cloud.yaml
examples/aquila/conf/train_hetero.yaml
examples/aquila/conf/inference/7b.yaml
examples/aquila/conf/train/34b.yaml
examples/aquila/conf/train/3b.yaml
examples/aquila/conf/train/70b.yaml
examples/aquila/conf/train/7b.yaml
examples/aquila/conf/train/hetero.yaml
examples/aquila/tokenizer/merges.txt
examples/aquila/tokenizer/special_tokens.txt
examples/aquila/tokenizer/vocab.json
examples/aquila/tokenizer_hf/special_tokens_map.json
examples/aquila/tokenizer_hf/tokenizer.json
examples/aquila/tokenizer_hf/tokenizer_config.json
examples/aquila/tokenizer_hf/vocab.json
examples/aquila/utils/__init__.py
examples/aquila/utils/convo_dataset.py
examples/aquila/utils/convo_prompt.py
examples/aquila/utils/cyg_conversation.py
examples/cloud/conf/serve.yaml
examples/cloud/conf/serve/cloud_model.yaml
examples/deepseek_r1/conf/hostfile.txt
examples/deepseek_r1/conf/serve.yaml
examples/deepseek_r1/conf/serve/671b.yaml
examples/deepseek_r1_distill_qwen/conf/serve.yaml
examples/deepseek_r1_distill_qwen/conf/serve/32b.yaml
examples/deepseek_v3/conf/serve.yaml
examples/deepseek_v3/conf/serve_auto_tuner.yaml
examples/deepseek_v3/conf/train.yaml
examples/deepseek_v3/conf/train_11b_a2b.yaml
examples/deepseek_v3/conf/train_auto_tuner.yaml
examples/deepseek_v3/conf/train_hetero.yaml
examples/deepseek_v3/conf/serve/671b.yaml
examples/deepseek_v3/conf/train/11b_a2b.yaml
examples/deepseek_v3/conf/train/16b_a3b.yaml
examples/deepseek_v3/conf/train/hetero.yaml
examples/emu3/conf/README.md
examples/emu3/conf/chat.yaml
examples/emu3/conf/compress.yaml
examples/emu3/conf/gen.yaml
examples/emu3/conf/compress/fp8_dynamic.yaml
examples/emu3/conf/compress/model.yaml
examples/emu3/conf/compress/w4a16.yaml
examples/emu3/conf/inference/i2t.yaml
examples/emu3/conf/inference/t2i.yaml
examples/ernie45/conf/hostfile.txt
examples/ernie45/conf/serve.yaml
examples/ernie45/conf/serve/300b.yaml
examples/grok2/conf/serve.yaml
examples/grok2/conf/serve/270b.yaml
examples/kimi_k2/conf/hostfile.txt
examples/kimi_k2/conf/serve.yaml
examples/kimi_k2/conf/serve/1t.yaml
examples/llama2/conf/train.yaml
examples/llama2/conf/train/7b.yaml
examples/llama3/conf/train.yaml
examples/llama3/conf/train_hetero.yaml
examples/llama3/conf/train/70b.yaml
examples/llama3/conf/train/70b_finetune.yaml
examples/llama3/conf/train/8b.yaml
examples/llama3/conf/train/8b_hetero.yaml
examples/llava1_5/conf/train.yaml
examples/llava1_5/conf/train/7b.yaml
examples/llava_onevision/conf/compress.yaml
examples/llava_onevision/conf/hostfile.txt
examples/llava_onevision/conf/serve.yaml
examples/llava_onevision/conf/train.yaml
examples/llava_onevision/conf/compress/fp8_dynamic.yaml
examples/llava_onevision/conf/compress/model.yaml
examples/llava_onevision/conf/compress/w4a16.yaml
examples/llava_onevision/conf/serve/7b.yaml
examples/llava_onevision/conf/train/1_5b.yaml
examples/llava_onevision/conf/train/7b.yaml
examples/minicpm_o_2.6/conf/serve.yaml
examples/minicpm_o_2.6/conf/serve/7b.yaml
examples/minicpm_v_4/conf/serve.yaml
examples/minicpm_v_4/conf/serve/4b.yaml
examples/mixtral/conf/train.yaml
examples/mixtral/conf/train/8x7b.yaml
examples/openjourney/conf/inference.yaml
examples/openjourney/conf/inference/t2i.yaml
examples/qwen2/conf/rl.yaml
examples/qwen2/conf/rl/7b.yaml
examples/qwen2_5/__init__.py
examples/qwen2_5/models.py
examples/qwen2_5/conf/hostfile.txt
examples/qwen2_5/conf/hostfile_nodes_envs.txt
examples/qwen2_5/conf/rl.yaml
examples/qwen2_5/conf/serve.yaml
examples/qwen2_5/conf/serve_auto_tuner.yaml
examples/qwen2_5/conf/serve_disagg_xpyd.yaml
examples/qwen2_5/conf/serve_multiple_instance.yaml
examples/qwen2_5/conf/serve_multiple_models.yaml
examples/qwen2_5/conf/serve_nodes_envs.yaml
examples/qwen2_5/conf/serve_ssh.yaml
examples/qwen2_5/conf/train.yaml
examples/qwen2_5/conf/rl/0_5b.yaml
examples/qwen2_5/conf/serve/72b.yaml
examples/qwen2_5/conf/serve/72b_nodes_envs.yaml
examples/qwen2_5/conf/serve/7b.yaml
examples/qwen2_5/conf/serve/7b_multiple_instance.yaml
examples/qwen2_5/conf/serve/multiple_models.yaml
examples/qwen2_5/conf/train/1_5b.yaml
examples/qwen2_5/utils/__init__.py
examples/qwen2_5/utils/convo_dataset.py
examples/qwen2_5/utils/convo_prompt.py
examples/qwen2_5/utils/cyg_conversation.py
examples/qwen2_5_vl/conf/serve.yaml
examples/qwen2_5_vl/conf/train.yaml
examples/qwen2_5_vl/conf/serve/32b_instruct.yaml
examples/qwen2_5_vl/conf/train/32b.yaml
examples/qwen2_5_vl/conf/train/7b.yaml
examples/qwen3/conf/serve.yaml
examples/qwen3/conf/serve_atmb.yaml
examples/qwen3/conf/train.yaml
examples/qwen3/conf/train_auto_tuner.yaml
examples/qwen3/conf/serve/0_6b.yaml
examples/qwen3/conf/serve/8b.yaml
examples/qwen3/conf/train/0_6b.yaml
examples/qwen3/conf/train/14b.yaml
examples/qwen3/conf/train/235b_a22b.yaml
examples/qwen3/conf/train/30b_a3b.yaml
examples/qwen3/conf/train/32b.yaml
examples/qwq/conf/serve.yaml
examples/qwq/conf/serve/32b.yaml
examples/robobrain/compress.yaml
examples/robobrain/conf/model.yaml
examples/robobrain/conf/serve.yaml
examples/robobrain/conf/w8a16.yaml
examples/robobrain/conf/serve/7b.yaml
examples/robobrain2/conf/serve.yaml
examples/robobrain2/conf/serve/32b.yaml
examples/robobrain2/conf/serve/7b.yaml
flag_scale.egg-info/PKG-INFO
flag_scale.egg-info/SOURCES.txt
flag_scale.egg-info/dependency_links.txt
flag_scale.egg-info/entry_points.txt
flag_scale.egg-info/requires.txt
flag_scale.egg-info/top_level.txt
flagscale/__init__.py
flagscale/cli.py
flagscale/logger.py
flagscale/patches_utils.py
flagscale/utils.py
flagscale/agent/__init__.py
flagscale/agent/collaboration/__init__.py
flagscale/agent/collaboration/collaborator.py
flagscale/backends/Megatron-LM/examples/multimodal/mlp_converter.py.patch
flagscale/backends/Megatron-LM/megatron/__init__.py.patch
flagscale/backends/Megatron-LM/megatron/core/model_parallel_config.py.patch
flagscale/backends/Megatron-LM/megatron/core/optimizer_param_scheduler.py.patch
flagscale/backends/Megatron-LM/megatron/core/parallel_state.py.patch
flagscale/backends/Megatron-LM/megatron/core/timers.py.patch
flagscale/backends/Megatron-LM/megatron/core/datasets/blended_dataset.py.patch
flagscale/backends/Megatron-LM/megatron/core/datasets/blended_megatron_dataset_builder.py.patch
flagscale/backends/Megatron-LM/megatron/core/datasets/gpt_dataset.py.patch
flagscale/backends/Megatron-LM/megatron/core/datasets/utils.py.patch
flagscale/backends/Megatron-LM/megatron/core/dist_checkpointing/exchange_utils.py.patch
flagscale/backends/Megatron-LM/megatron/core/dist_checkpointing/mapping.py.patch
flagscale/backends/Megatron-LM/megatron/core/dist_checkpointing/serialization.py.patch
flagscale/backends/Megatron-LM/megatron/core/dist_checkpointing/validation.py.patch
flagscale/backends/Megatron-LM/megatron/core/dist_checkpointing/strategies/filesystem_async.py.patch
flagscale/backends/Megatron-LM/megatron/core/dist_checkpointing/strategies/torch.py.patch
flagscale/backends/Megatron-LM/megatron/core/distributed/finalize_model_grads.py.patch
flagscale/backends/Megatron-LM/megatron/core/extensions/transformer_engine.py.patch
flagscale/backends/Megatron-LM/megatron/core/models/common/embeddings/rotary_pos_embedding.py.patch
flagscale/backends/Megatron-LM/megatron/core/models/common/language_module/language_module.py.patch
flagscale/backends/Megatron-LM/megatron/core/models/gpt/gpt_layer_specs.py.patch
flagscale/backends/Megatron-LM/megatron/core/models/multimodal/llava_model.py.patch
flagscale/backends/Megatron-LM/megatron/core/optimizer/__init__.py.patch
flagscale/backends/Megatron-LM/megatron/core/optimizer/clip_grads.py.patch
flagscale/backends/Megatron-LM/megatron/core/optimizer/optimizer.py.patch
flagscale/backends/Megatron-LM/megatron/core/optimizer/optimizer_config.py.patch
flagscale/backends/Megatron-LM/megatron/core/pipeline_parallel/p2p_communication.py.patch
flagscale/backends/Megatron-LM/megatron/core/pipeline_parallel/schedules.py.patch
flagscale/backends/Megatron-LM/megatron/core/transformer/attention.py.patch
flagscale/backends/Megatron-LM/megatron/core/transformer/module.py.patch
flagscale/backends/Megatron-LM/megatron/core/transformer/transformer_block.py.patch
flagscale/backends/Megatron-LM/megatron/core/transformer/transformer_config.py.patch
flagscale/backends/Megatron-LM/megatron/core/transformer/transformer_layer.py.patch
flagscale/backends/Megatron-LM/megatron/core/transformer/moe/experts.py.patch
flagscale/backends/Megatron-LM/megatron/core/transformer/moe/moe_utils.py.patch
flagscale/backends/Megatron-LM/megatron/inference/text_generation/sampling.py.patch
flagscale/backends/Megatron-LM/megatron/inference/text_generation/tokenization.py.patch
flagscale/backends/Megatron-LM/megatron/training/arguments.py.patch
flagscale/backends/Megatron-LM/megatron/training/checkpointing.py.patch
flagscale/backends/Megatron-LM/megatron/training/global_vars.py.patch
flagscale/backends/Megatron-LM/megatron/training/initialize.py.patch
flagscale/backends/Megatron-LM/megatron/training/utils.py.patch
flagscale/backends/Megatron-LM/megatron/training/tokenizer/gpt2_tokenization.py.patch
flagscale/backends/Megatron-LM/megatron/training/tokenizer/tokenization_utils.py.patch
flagscale/backends/Megatron-LM/megatron/training/tokenizer/tokenizer.py.patch
flagscale/backends/Megatron-LM/tasks/main.py.patch
flagscale/backends/Megatron-LM/tasks/aquila/datasets.py.patch
flagscale/backends/Megatron-LM/tasks/aquila/evaluate.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/test_parallel_state.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/test_utils.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/data/__init__.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/data/test_builder.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/dist_checkpointing/__init__.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/dist_checkpointing/test_flattened_resharding.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/dist_checkpointing/test_optimizer.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/dist_checkpointing/test_serialization.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/dist_checkpointing/utils.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/dist_checkpointing/models/test_bert_model.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/dist_checkpointing/models/test_mlp_glu.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/dist_checkpointing/models/test_moe_experts.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/distributed/test_grad_sync_with_expert_parallel.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/export/trtllm/test_distributed_fp8.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/export/trtllm/test_single_device_fp8.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/models/test_bert_model.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/models/test_llava_model.py.patch
flagscale/backends/Megatron-LM/tests/unit_tests/transformer/test_transformer_block.py.patch
flagscale/backends/Megatron-LM/tools/preprocess_data.py.patch
flagscale/backends/Megatron-LM/tools/checkpoint/saver_legacy.py.patch
flagscale/backends/llama.cpp/README.md
flagscale/backends/sglang/python/sglang/__init__.py.patch
flagscale/backends/sglang/python/sglang/srt/model_executor/model_runner.py.patch
flagscale/backends/vllm/vllm/sampling_params.py.patch
flagscale/backends/vllm/vllm/sequence.py.patch
flagscale/backends/vllm/vllm/core/block_manager.py.patch
flagscale/backends/vllm/vllm/core/scheduler.py.patch
flagscale/backends/vllm/vllm/distributed/kv_transfer/kv_connector/factory.py.patch
flagscale/backends/vllm/vllm/distributed/kv_transfer/kv_connector/p2p_connector.py.patch
flagscale/backends/vllm/vllm/distributed/kv_transfer/kv_connector/v1/p2p/flagcx_p2p_nccl_engine.py.patch
flagscale/backends/vllm/vllm/distributed/kv_transfer/kv_connector/v1/p2p/p2p_nccl_connector.py.patch
flagscale/backends/vllm/vllm/distributed/kv_transfer/kv_pipe/flagcx_p2p_nccl_pipe.py.patch
flagscale/backends/vllm/vllm/distributed/kv_transfer/kv_pipe/p2p_nccl_pipe.py.patch
flagscale/backends/vllm/vllm/engine/llm_engine.py.patch
flagscale/backends/vllm/vllm/engine/output_processor/single_step.py.patch
flagscale/backends/vllm/vllm/inputs/data.py.patch
flagscale/backends/vllm/vllm/inputs/preprocess.py.patch
flagscale/backends/vllm/vllm/model_executor/sampling_metadata.py.patch
flagscale/backends/vllm/vllm/model_executor/layers/logits_processor.py.patch
flagscale/backends/vllm/vllm/model_executor/models/llava_onevision.py.patch
flagscale/backends/vllm/vllm/model_executor/models/minicpmv.py.patch
flagscale/backends/vllm/vllm/model_executor/models/registry.py.patch
flagscale/backends/vllm/vllm/model_executor/models/siglip.py.patch
flagscale/backends/vllm/vllm/v1/worker/gpu_model_runner.py.patch
flagscale/backends/vllm/vllm/worker/model_runner.py.patch
flagscale/compress/__init__.py
flagscale/compress/adapter.py
flagscale/compress/combined_algo.py
flagscale/compress/compressor.py
flagscale/compress/compressor_emu3.py
flagscale/compress/compressor_llava_ov.py
flagscale/compress/algo/__init__.py
flagscale/compress/algo/algo_base.py
flagscale/inference/__init__.py
flagscale/inference/arguments.py
flagscale/inference/diffusion_entrypoint.py
flagscale/inference/inference_aquila.py
flagscale/inference/inference_emu3.py
flagscale/inference/inference_engine.py
flagscale/inference/processing_emu3.py
flagscale/inference/runtime_context.py
flagscale/runner/__init__.py
flagscale/runner/runner_base.py
flagscale/runner/runner_compress.py
flagscale/runner/runner_inference.py
flagscale/runner/runner_rl.py
flagscale/runner/runner_serve.py
flagscale/runner/runner_train.py
flagscale/runner/utils.py
flagscale/runner/auto_tuner/__init__.py
flagscale/runner/auto_tuner/generate.py
flagscale/runner/auto_tuner/memory_model.py
flagscale/runner/auto_tuner/platform.py
flagscale/runner/auto_tuner/tuner.py
flagscale/runner/auto_tuner/utils.py
flagscale/runner/auto_tuner/prune/history.py
flagscale/runner/auto_tuner/prune/memory.py
flagscale/runner/auto_tuner/prune/pruner.py
flagscale/runner/auto_tuner/record/recorder.py
flagscale/runner/auto_tuner/search/algorithm.py
flagscale/runner/auto_tuner/search/searcher.py
flagscale/runner/estimator/__init__.py
flagscale/runner/estimator/meta_attention.py
flagscale/runner/estimator/meta_base.py
flagscale/runner/estimator/meta_functional.py
flagscale/runner/estimator/meta_gpt.py
flagscale/runner/estimator/meta_mlp.py
flagscale/runner/estimator/meta_modules.py
flagscale/runner/estimator/meta_registry.py
flagscale/runner/estimator/meta_tensor.py
flagscale/runner/estimator/meta_transformer_layer.py
flagscale/runner/estimator/utils.py
flagscale/serve/README.md
flagscale/serve/__init__.py
flagscale/serve/arguments.py
flagscale/serve/core.py
flagscale/serve/dag_utils.py
flagscale/serve/engine.py
flagscale/serve/run_disagg_xpyd_router.py
flagscale/serve/run_fs_serve_vllm.py
flagscale/serve/run_inference_engine.py
flagscale/serve/run_serve.py
flagscale/serve/utils.py
flagscale/serve/args_mapping/README.md
flagscale/serve/args_mapping/common_args.yaml
flagscale/serve/args_mapping/gen_template_conf.py
flagscale/serve/args_mapping/gen_template_funcs.py
flagscale/serve/args_mapping/mapping.py
flagscale/serve/args_mapping/mapping.yaml
flagscale/serve/args_mapping/mapping_funcs/llama_cpp.py
flagscale/serve/metric/__init__.py
flagscale/serve/metric/serve_metric.py
flagscale/train/__init__.py
flagscale/train/arguments.py
flagscale/train/extra_valid.py
flagscale/train/global_vars.py
flagscale/train/spiky_loss.py
flagscale/train/stablelm2_scheduler.py
flagscale/train/theoretical_memory_usage.py
flagscale/train/train.py
flagscale/train/train_aquila_sft.py
flagscale/train/train_gpt.py
flagscale/train/train_llava.py
flagscale/train/train_llava_onevision.py
flagscale/train/train_qwen2_5_vl.py
flagscale/train/datasets/concated_indexed_dataset.py
flagscale/train/datasets/sft_dataset.py
flagscale/train/dualpipev/dualpipev_schedules.py
flagscale/train/dualpipev/fb_overlap/gpt_model.py
flagscale/train/dualpipev/fb_overlap/transformer_block.py
flagscale/train/dualpipev/fb_overlap/transformer_layer.py
flagscale/train/dualpipev/fb_overlap/modules/attention.py
flagscale/train/dualpipev/fb_overlap/modules/token_dispatcher.py
flagscale/train/dualpipev/fb_overlap/modules/utils.py
flagscale/train/dualpipev/fb_overlap/overlap_funcs/bwd.py
flagscale/train/dualpipev/fb_overlap/overlap_funcs/fwd.py
flagscale/train/dualpipev/fb_overlap/overlap_funcs/fwdbwd.py
flagscale/train/hetero/__init__.py
flagscale/train/hetero/p2p_communication.py
flagscale/train/hetero/parallel_context.py
flagscale/train/models/__init__.py
flagscale/train/models/llava_onevision/__init__.py
flagscale/train/models/llava_onevision/clip_vit_model.py
flagscale/train/models/llava_onevision/config.py
flagscale/train/models/llava_onevision/dataloader_provider.py
flagscale/train/models/llava_onevision/dataset_helpers.py
flagscale/train/models/llava_onevision/layer_specs.py
flagscale/train/models/llava_onevision/llava_onevision_model.py
flagscale/train/models/qwen2_5_vl/QuickStart.md
flagscale/train/models/qwen2_5_vl/__init__.py
flagscale/train/models/qwen2_5_vl/language_module.py
flagscale/train/models/qwen2_5_vl/layer_specs.py
flagscale/train/models/qwen2_5_vl/qwen2_5_vl_model.py
flagscale/train/models/qwen2_5_vl/tensor_parallel.py
flagscale/train/models/qwen2_5_vl/transformer_config.py
flagscale/train/models/qwen2_5_vl/vision_attention.py
flagscale/train/models/qwen2_5_vl/vision_transformer_block.py
flagscale/train/models/qwen2_5_vl/vit_model.py
flagscale/transforms/__init__.py
flagscale/transforms/hook.py
flagscale/transforms/log_io_transformation.py
flagscale/transforms/state_scope_transformation.py
flagscale/transforms/state_store.py
flagscale/transforms/transformation.py
hardware/LICENSE
hardware/LICENSE_CN
hardware/README.md
hardware/README_CN.md
hardware/patch_history.yaml
hardware/BI_V150/FlagScale/diff.patch.encrypted
hardware/BI_V150/FlagScale/diff.yaml
hardware/BI_V150/Megatron-LM/diff.patch.encrypted
hardware/BI_V150/Megatron-LM/diff.yaml
hardware/BI_V150/vllm/diff.patch.encrypted
hardware/BI_V150/vllm/diff.yaml
hardware/Cambricon_MLU/FlagScale/diff.patch
hardware/Cambricon_MLU/FlagScale/diff.yaml
hardware/Cambricon_MLU/vllm/diff.patch
hardware/Cambricon_MLU/vllm/diff.yaml
hardware/Huawei_Atlas800TA3/FlagScale/diff.patch.encrypted
hardware/Huawei_Atlas800TA3/FlagScale/diff.yaml
hardware/Huawei_Atlas800TA3/Megatron-LM/diff.patch.encrypted
hardware/Huawei_Atlas800TA3/Megatron-LM/diff.yaml
hardware/Huawei_Atlas800TA3/vllm/diff.patch.encrypted
hardware/Huawei_Atlas800TA3/vllm/diff.yaml
hardware/Hygon_BW1000/FlagScale/diff.patch.encrypted
hardware/Hygon_BW1000/FlagScale/diff.yaml
hardware/Hygon_BW1000/Megatron-LM/diff.patch
hardware/Hygon_BW1000/Megatron-LM/diff.yaml
hardware/Hygon_BW1000/vllm/diff.patch.encrypted
hardware/Hygon_BW1000/vllm/diff.yaml
hardware/Kunlunxin_R310p/FlagScale/diff.patch.encrypted
hardware/Kunlunxin_R310p/FlagScale/diff.yaml
hardware/Kunlunxin_R310p/Megatron-LM/diff.patch.encrypted
hardware/Kunlunxin_R310p/Megatron-LM/diff.yaml
hardware/Kunlunxin_R310p/vllm/diff.patch.encrypted
hardware/Kunlunxin_R310p/vllm/diff.yaml
hardware/MUSA_S5000/FlagScale/diff.patch
hardware/MUSA_S5000/FlagScale/diff.yaml
hardware/MUSA_S5000/Megatron-LM/diff.patch
hardware/MUSA_S5000/Megatron-LM/diff.yaml
hardware/MUSA_S5000/vllm/diff.patch
hardware/MUSA_S5000/vllm/diff.yaml
hardware/Metax_C550/FlagScale/diff.patch
hardware/Metax_C550/FlagScale/diff.yaml
hardware/Metax_C550/Megatron-LM/diff.patch
hardware/Metax_C550/Megatron-LM/diff.yaml
hardware/Metax_C550/vllm/diff.patch
hardware/Metax_C550/vllm/diff.yaml
hardware/Tsing_micro/FlagScale/diff.patch.encrypted
hardware/Tsing_micro/FlagScale/diff.yaml
hardware/Tsing_micro/Megatron-LM/diff.patch.encrypted
hardware/Tsing_micro/Megatron-LM/diff.yaml
hardware/Tsing_micro/vllm/diff.patch.encrypted
hardware/Tsing_micro/vllm/diff.yaml
install/README.md
install/install-requirements-metax.sh
install/install-requirements.sh
requirements/requirements-base.txt
requirements/requirements-common.txt
requirements/compress/requirements-dev.txt
requirements/inference/requirements.txt
requirements/inference/vllm/README.md
requirements/serving/requirements.txt
requirements/train/megatron/requirements-cuda.txt
tests/README.md
tests/functional_tests/test_cases/hetero_train/aquila/conf/dp2dp4_shared_embedding.yaml
tests/functional_tests/test_cases/hetero_train/aquila/conf/tp2dp1pp1_tp2dp2pp1_tp1dp2pp1.yaml
tests/functional_tests/test_cases/hetero_train/aquila/conf/tp2pp1_tp4pp1_tp2pp1.yaml
tests/functional_tests/test_cases/hetero_train/aquila/conf/train/dp2dp4_shared_embedding.yaml
tests/functional_tests/test_cases/hetero_train/aquila/conf/train/tp2dp1pp1_tp2dp2pp1_tp1dp2pp1.yaml
tests/functional_tests/test_cases/hetero_train/aquila/conf/train/tp2pp1_tp4pp1_tp2pp1.yaml
tests/functional_tests/test_cases/hetero_train/aquila/results_gold/dp2dp4_shared_embedding.json
tests/functional_tests/test_cases/hetero_train/aquila/results_gold/tp2dp1pp1_tp2dp2pp1_tp1dp2pp1.json
tests/functional_tests/test_cases/hetero_train/aquila/results_gold/tp2pp1_tp4pp1_tp2pp1.json
tests/functional_tests/test_cases/inference/deepseek_r1_distill_qwen-flaggems-metax/conf/7b-tp2.yaml
tests/functional_tests/test_cases/inference/deepseek_r1_distill_qwen-flaggems-metax/conf/inference/7b-tp2.yaml
tests/functional_tests/test_cases/inference/deepseek_r1_distill_qwen-flaggems-metax/results_gold/7b-tp2
tests/functional_tests/test_cases/inference/deepseek_r1_distill_qwen-flaggems/conf/7b-tp2.yaml
tests/functional_tests/test_cases/inference/deepseek_r1_distill_qwen-flaggems/conf/inference/7b-tp2.yaml
tests/functional_tests/test_cases/inference/deepseek_r1_distill_qwen-flaggems/results_gold/7b-tp2
tests/functional_tests/test_cases/inference/deepseek_r1_distill_qwen-metax/conf/7b-tp2.yaml
tests/functional_tests/test_cases/inference/deepseek_r1_distill_qwen-metax/conf/inference/7b-tp2.yaml
tests/functional_tests/test_cases/inference/deepseek_r1_distill_qwen-metax/results_gold/7b-tp2
tests/functional_tests/test_cases/inference/deepseek_r1_distill_qwen/conf/7b-tp2.yaml
tests/functional_tests/test_cases/inference/deepseek_r1_distill_qwen/conf/inference/7b-tp2.yaml
tests/functional_tests/test_cases/inference/deepseek_r1_distill_qwen/results_gold/7b-tp2
tests/functional_tests/test_cases/inference/opi_llama3_1_instruct-flaggems-metax/conf/8b-tp2.yaml
tests/functional_tests/test_cases/inference/opi_llama3_1_instruct-flaggems-metax/conf/inference/8b-tp2.yaml
tests/functional_tests/test_cases/inference/opi_llama3_1_instruct-flaggems-metax/results_gold/8b-tp2
tests/functional_tests/test_cases/inference/opi_llama3_1_instruct-metax/conf/8b-tp2.yaml
tests/functional_tests/test_cases/inference/opi_llama3_1_instruct-metax/conf/inference/8b-tp2.yaml
tests/functional_tests/test_cases/inference/opi_llama3_1_instruct-metax/results_gold/8b-tp2
tests/functional_tests/test_cases/inference/qwen3-flaggems-metax/conf/4b-tp2.yaml
tests/functional_tests/test_cases/inference/qwen3-flaggems-metax/conf/inference/4b-tp2.yaml
tests/functional_tests/test_cases/inference/qwen3-flaggems-metax/results_gold/4b-tp2
tests/functional_tests/test_cases/inference/qwen3-flaggems/conf/4b-tp2.yaml
tests/functional_tests/test_cases/inference/qwen3-flaggems/conf/inference/4b-tp2.yaml
tests/functional_tests/test_cases/inference/qwen3-flaggems/results_gold/4b-tp2
tests/functional_tests/test_cases/inference/qwen3-metax/conf/4b-tp2.yaml
tests/functional_tests/test_cases/inference/qwen3-metax/conf/inference/4b-tp2.yaml
tests/functional_tests/test_cases/inference/qwen3-metax/results_gold/4b-tp2
tests/functional_tests/test_cases/inference/qwen3/conf/4b-tp2.yaml
tests/functional_tests/test_cases/inference/qwen3/conf/inference/4b-tp2.yaml
tests/functional_tests/test_cases/inference/qwen3/results_gold/4b-tp2
tests/functional_tests/test_cases/rl/qwen2_5/conf/0_5b.yaml
tests/functional_tests/test_cases/rl/qwen2_5/conf/rl/0_5b.yaml
tests/functional_tests/test_cases/rl/qwen2_5/results_gold/0_5b.json
tests/functional_tests/test_cases/serve/base/conf/multiple_model.yaml
tests/functional_tests/test_cases/serve/base/conf/serve/multiple_model.yaml
tests/functional_tests/test_cases/serve/qwen2_5/conf/0.5b.yaml
tests/functional_tests/test_cases/serve/qwen2_5/conf/0.5b_multiple_instance.yaml
tests/functional_tests/test_cases/serve/qwen2_5/conf/serve/0.5b.yaml
tests/functional_tests/test_cases/serve/qwen2_5/conf/serve/0.5b_multiple_instance.yaml
tests/functional_tests/test_cases/serve/utils/models/base_model.py
tests/functional_tests/test_cases/serve/utils/models/util_models/util_model.py
tests/functional_tests/test_cases/train/aquila/conf/tp2_pp2.yaml
tests/functional_tests/test_cases/train/aquila/conf/tp4_pp2.yaml
tests/functional_tests/test_cases/train/aquila/conf/train/tp2_pp2.yaml
tests/functional_tests/test_cases/train/aquila/conf/train/tp4_pp2.yaml
tests/functional_tests/test_cases/train/aquila/results_gold/tp2_pp2.json
tests/functional_tests/test_cases/train/aquila/results_gold/tp4_pp2.json
tests/functional_tests/test_cases/train/deepseek/conf/tp2_pp2_ep2.yaml
tests/functional_tests/test_cases/train/deepseek/conf/train/tp2_pp2_ep2.yaml
tests/functional_tests/test_cases/train/deepseek/results_gold/tp2_pp2_ep2.json
tests/functional_tests/test_cases/train/llava_onevision/pretrain_dataset.yaml
tests/functional_tests/test_cases/train/llava_onevision/conf/tp2.yaml
tests/functional_tests/test_cases/train/llava_onevision/conf/tp4.yaml
tests/functional_tests/test_cases/train/llava_onevision/conf/train/tp2.yaml
tests/functional_tests/test_cases/train/llava_onevision/conf/train/tp4.yaml
tests/functional_tests/test_cases/train/llava_onevision/results_gold/tp2.json
tests/functional_tests/test_cases/train/llava_onevision/results_gold/tp4.json
tests/functional_tests/test_cases/train/mixtral/conf/tp2_pp1_ep2.yaml
tests/functional_tests/test_cases/train/mixtral/conf/tp4_pp1_ep2.yaml
tests/functional_tests/test_cases/train/mixtral/conf/train/tp2_pp1_ep2.yaml
tests/functional_tests/test_cases/train/mixtral/conf/train/tp4_pp1_ep2.yaml
tests/functional_tests/test_cases/train/mixtral/results_gold/tp2_pp1_ep2.json
tests/functional_tests/test_cases/train/mixtral/results_gold/tp4_pp1_ep2.json
tests/functional_tests/test_utils/conftest.py
tests/functional_tests/test_utils/test_call.py
tests/functional_tests/test_utils/test_result.py
tests/scripts/_gpu_check.sh
tests/scripts/_tests_log.sh
tests/scripts/_tests_stop.sh
tests/scripts/utils.sh
tests/scripts/functional_tests/config.yml
tests/scripts/functional_tests/parse_config.py
tests/scripts/functional_tests/test_all.sh
tests/scripts/functional_tests/test_task.sh
tests/scripts/unit_tests/config.yml
tests/scripts/unit_tests/parse_config.py
tests/scripts/unit_tests/test_all.sh
tests/scripts/unit_tests/test_coverage.sh
tests/scripts/unit_tests/test_subset.sh
tests/unit_tests/test_basic.py
tests/unit_tests/test_parallel_context.py
tests/unit_tests/test_spiky_loss_detector.py
tests/unit_tests/utilities.py
tests/unit_tests/compressor/test_adapter.py
tests/unit_tests/compressor/test_compressor.py
tests/unit_tests/compressor/test_config.yaml
tests/unit_tests/inference/test_runtime_context.py
tests/unit_tests/runner/test_parse_hostfile.py
tests/unit_tests/runner/estimator/test_meta_attention.py
tests/unit_tests/runner/estimator/test_meta_base.py
tests/unit_tests/runner/estimator/test_meta_gpt.py
tests/unit_tests/runner/estimator/test_meta_mlp.py
tests/unit_tests/runner/estimator/test_meta_modules.py
tests/unit_tests/runner/estimator/test_meta_registry.py
tests/unit_tests/runner/estimator/test_meta_tensor.py
tests/unit_tests/runner/estimator/test_meta_transformer_layer.py
tests/unit_tests/transforms/__init__.py
tests/unit_tests/transforms/test_hook.py
tests/unit_tests/transforms/test_state_scope_transform.py
tests/unit_tests/transforms/test_state_store.py
tools/checkpoint/convert.py
tools/checkpoint/loader_mcore.py
tools/checkpoint/loader_transformers.py
tools/checkpoint/run.sh
tools/checkpoint/saver_mcore.py
tools/checkpoint/saver_transformers.py
tools/checkpoint/utils.py
tools/checkpoint/aquila/args.py
tools/checkpoint/aquila/ckpt.py
tools/checkpoint/aquila/model.py
tools/checkpoint/deepseek_v3/args.py
tools/checkpoint/deepseek_v3/ckpt.py
tools/checkpoint/deepseek_v3/model.py
tools/checkpoint/llama/args.py
tools/checkpoint/llama/ckpt.py
tools/checkpoint/llama/model.py
tools/checkpoint/llava_onevision/combine.py
tools/checkpoint/llava_onevision/convert_mlp.py
tools/checkpoint/llava_onevision/convert_qwen2.5_1.5b.py
tools/checkpoint/llava_onevision/convert_qwen2.5_7b.py
tools/checkpoint/llava_onevision/convert_siglip.py
tools/checkpoint/llava_onevision/convert_to_fs_qwen2.5_1.5b.py
tools/checkpoint/llava_onevision/convert_to_fs_qwen2.5_7b.py
tools/checkpoint/llava_onevision/convert_to_hf_qwen2.5_1.5b.py
tools/checkpoint/llava_onevision/convert_to_hf_qwen2.5_7b.py
tools/checkpoint/mistral/args.py
tools/checkpoint/mistral/ckpt.py
tools/checkpoint/mistral/model.py
tools/checkpoint/mixtral/args.py
tools/checkpoint/mixtral/ckpt.py
tools/checkpoint/mixtral/model.py
tools/checkpoint/moonlight_deepseek/configuration_deepseek.py
tools/checkpoint/moonlight_deepseek/modeling_deepseek.py
tools/checkpoint/qwen2_5_vl/convert.md
tools/checkpoint/qwen2_5_vl/hf2mcore_qwen2.5_vl.py
tools/checkpoint/qwen2_5_vl/hf2mcore_qwen2.5_vl_convertor.sh
tools/checkpoint/qwen2_5_vl/utils.py
tools/checkpoint/qwen3/args.py
tools/checkpoint/qwen3/ckpt.py
tools/checkpoint/qwen3/model.py
tools/checkpoint/qwen3/modeling_hf/configuration_qwen3.py
tools/checkpoint/qwen3/modeling_hf/modeling_qwen3.py
tools/checkpoint/sfpt_ckpt/README.md
tools/checkpoint/sfpt_ckpt/dcp_to_sfpt.py
tools/checkpoint/sfpt_ckpt/sfpt_to_dcp.py
tools/codestyle/check_header.py
tools/codestyle/pre-commit.sh
tools/datasets/llava_onevision/README.md
tools/datasets/llava_onevision/llava_ov_wds.py
tools/datasets/llava_onevision/make_llava_ov_wds.sh
tools/datasets/qwenvl/build_llava_frame_dataset.py
tools/datasets/qwenvl/convert_custom_dataset_to_wds_chatml_str.py
tools/datasets/qwenvl/convert_llava_pretrain_to_wds.py
tools/datasets/qwenvl/dataset_preparation.md
tools/datasets/qwenvl/replace_llava_image_key.py
tools/datasets/qwenvl/data/__init__.py
tools/datasets/qwenvl/data/dataset_helpers.py
tools/datasets/qwenvl/data/image_processing.py
tools/datasets/qwenvl/data/utils.py
tools/datasets/qwenvl/data/energon/chatml.py
tools/estimator/estimate_gpt.py
tools/patch/__init__.py
tools/patch/encryption_utils.py
tools/patch/file_utils.py
tools/patch/git_utils.py
tools/patch/logger_utils.py
tools/patch/merge.py
tools/patch/patch.py
tools/patch/patch_v2.py
tools/patch/unpatch.py
tools/patch/unpatch_v2.py
tools/patch/__pycache__/encryption_utils.cpython-312.pyc
tools/patch/__pycache__/git_utils.cpython-312.pyc
tools/patch/__pycache__/logger_utils.cpython-312.pyc
tools/patch/__pycache__/patch.cpython-312.pyc