├── CODEOWNERS ├── CONTRIBUTING.md ├── Dockerfile.ci ├── LICENSE ├── MANIFEST.in ├── README.md ├── dataset-preprocessing ├── the-pile │ ├── download.py │ ├── download.sh │ └── process.sh └── wikitext-2 │ ├── download.py │ └── preprocess.sh ├── docs ├── llama2.md └── source │ ├── api-guide │ ├── dist_checkpointing.rst │ ├── dist_checkpointing.strategies.rst │ ├── distributed.rst │ ├── fusions.rst │ ├── index.rst │ ├── models.gpt.rst │ ├── models.rst │ ├── pipeline_parallel.rst │ ├── tensor_parallel.rst │ └── transformer.rst │ ├── distrib_optimizer.md │ ├── images │ └── distrib_optimizer │ │ ├── data_flow.png │ │ └── sharding_scheme.png │ ├── index.rst │ └── user-guide │ └── index.rst ├── examples ├── bert │ ├── README.md │ └── train_bert_340m_distributed.sh ├── detxoify_lm │ ├── README.md │ ├── annotations │ │ ├── filter-selfgeneration.py │ │ ├── perspective_api_annotate.py │ │ └── preprocess.sh │ ├── finetune_gpt.py │ ├── finetune_gpt_distributed-1.3b.sh │ ├── generate-1.3b.sh │ ├── generate_samples_gpt.py │ ├── perspective_api.py │ └── self_generation │ │ └── selfgenerate-1.3b-unconditional.sh ├── evaluate_retriever_nq.sh ├── evaluate_zeroshot_gpt.sh ├── finetune_mnli_distributed.sh ├── finetune_race_distributed.sh ├── finetune_retriever_distributed.sh ├── gpt3 │ ├── README.md │ └── train_gpt3_175b_distributed.sh ├── merge_mp_bert.sh ├── msdp │ ├── README.md │ ├── data_processing.sh │ ├── eval_knwl_generation.sh │ ├── eval_resp_generation.sh │ ├── prep_resp_gen.sh │ ├── prompt_knwl_gen.sh │ └── prompt_resp_gen.sh ├── pretrain_bert.sh ├── pretrain_bert_distributed.sh ├── pretrain_bert_distributed_with_mp.sh ├── pretrain_gpt.sh ├── pretrain_gpt3_175B.sh ├── pretrain_gpt_distributed.sh ├── pretrain_gpt_distributed_with_mp.sh ├── pretrain_ict.sh ├── pretrain_t5.sh ├── pretrain_t5_distributed.sh ├── pretrain_t5_distributed_with_mp.sh ├── pretrain_vision_classify.sh ├── pretrain_vision_dino.sh ├── pretrain_vision_inpaint.sh ├── run_text_generation_server_345M.sh ├── run_text_generation_server_345M_8_tensor_parallel.sh ├── sc21 │ ├── CONFIG.sh │ ├── README.md │ ├── SBATCH.sh │ ├── SRUN.sh │ ├── run_figure_11.sh │ ├── run_figure_12.sh │ ├── run_figure_13.sh │ ├── run_figure_14.sh │ ├── run_figure_15.sh │ ├── run_figure_16.sh │ ├── run_figure_17.sh │ ├── run_figure_18.sh │ └── run_table_1.sh └── t5 │ ├── README.md │ ├── t5_mcore_train_curve.png │ └── train_t5_220m_distributed.sh ├── images ├── Achieved_petaFLOPs.png └── cases_april2021.png ├── jet-tests.yml ├── licenses ├── LICENSE-deepspeed.txt ├── LICENSE-fast-hadamard-transform.txt └── LICENSE-megatron-lm.txt ├── megatron ├── __init__.py ├── arguments.py ├── checkpointing.py ├── core │ ├── README.md │ ├── __init__.py │ ├── datasets │ │ ├── Makefile │ │ ├── __init__.py │ │ ├── blended_dataset.py │ │ ├── blended_megatron_dataset_builder.py │ │ ├── blended_megatron_dataset_config.py │ │ ├── gpt_dataset.py │ │ ├── helpers.cpp │ │ ├── indexed_dataset.py │ │ ├── megatron_dataset.py │ │ ├── megatron_tokenizer.py │ │ ├── readme.md │ │ └── utils.py │ ├── dist_checkpointing │ │ ├── __init__.py │ │ ├── core.py │ │ ├── dict_utils.py │ │ ├── mapping.py │ │ ├── optimizer.py │ │ ├── serialization.py │ │ ├── strategies │ │ │ ├── __init__.py │ │ │ ├── base.py │ │ │ ├── tensorstore.py │ │ │ ├── two_stage.py │ │ │ └── zarr.py │ │ └── utils.py │ ├── distributed │ │ ├── __init__.py │ │ ├── distributed_data_parallel.py │ │ ├── finalize_model_grads.py │ │ └── grad_buffer.py │ ├── enums.py │ ├── fusions │ │ ├── __init__.py │ │ ├── fused_bias_dropout.py │ │ ├── fused_bias_gelu.py │ │ ├── fused_bias_swiglu.py │ │ ├── fused_layer_norm.py │ │ └── fused_softmax.py │ ├── inference_params.py │ ├── jit.py │ ├── model_parallel_config.py │ ├── models │ │ ├── T5 │ │ │ ├── __init__.py │ │ │ ├── t5_model.py │ │ │ └── t5_spec.py │ │ ├── __init__.py │ │ ├── bert │ │ │ ├── __init__.py │ │ │ ├── bert_layer_specs.py │ │ │ ├── bert_lm_head.py │ │ │ ├── bert_model.py │ │ │ └── pooler.py │ │ ├── common │ │ │ ├── __init__.py │ │ │ ├── embeddings │ │ │ │ ├── __init__.py │ │ │ │ ├── language_model_embedding.py │ │ │ │ └── rotary_pos_embedding.py │ │ │ └── language_module │ │ │ │ ├── __init__.py │ │ │ │ └── language_module.py │ │ ├── gpt │ │ │ ├── __init__.py │ │ │ ├── gpt_layer_specs.py │ │ │ └── gpt_model.py │ │ └── retro │ │ │ ├── __init__.py │ │ │ ├── base_attention.py │ │ │ ├── config.py │ │ │ ├── decoder_attention.py │ │ │ ├── decoder_spec.py │ │ │ ├── encoder_attention.py │ │ │ ├── encoder_spec.py │ │ │ └── model.py │ ├── package_info.py │ ├── packed_seq_params.py │ ├── parallel_state.py │ ├── pipeline_parallel │ │ ├── __init__.py │ │ ├── p2p_communication.py │ │ └── schedules.py │ ├── requirements.txt │ ├── tensor_parallel │ │ ├── __init__.py │ │ ├── cross_entropy.py │ │ ├── data.py │ │ ├── layers.py │ │ ├── mappings.py │ │ ├── random.py │ │ └── utils.py │ ├── transformer │ │ ├── __init__.py │ │ ├── attention.py │ │ ├── custom_layers │ │ │ ├── __init__.py │ │ │ └── transformer_engine.py │ │ ├── dot_product_attention.py │ │ ├── enums.py │ │ ├── identity_op.py │ │ ├── mlp.py │ │ ├── module.py │ │ ├── moe │ │ │ ├── __init__.py │ │ │ ├── base_moe_layer.py │ │ │ ├── grouped_gemm_util.py │ │ │ ├── grouped_mlp.py │ │ │ └── switch_mlp.py │ │ ├── spec_utils.py │ │ ├── transformer_block.py │ │ ├── transformer_config.py │ │ ├── transformer_layer.py │ │ └── utils.py │ └── utils.py ├── data │ ├── __init__.py │ ├── autoaugment.py │ ├── bert_dataset.py │ ├── biencoder_dataset_utils.py │ ├── data_samplers.py │ ├── dataset_utils.py │ ├── ict_dataset.py │ ├── image_folder.py │ ├── multimodal_dataset.py │ ├── orqa_wiki_dataset.py │ ├── realm_dataset_utils.py │ ├── realm_index.py │ ├── t5_dataset.py │ └── vit_dataset.py ├── dist_signal_handler.py ├── fp16_deprecated │ └── loss_scaler.py ├── fused_kernels │ ├── __init__.py │ ├── compat.h │ ├── tests │ │ ├── __init__.py │ │ └── test_fused_kernels.py │ └── type_shim.h ├── global_vars.py ├── indexer.py ├── initialize.py ├── log_handler.py ├── memory.py ├── microbatches.py ├── model │ ├── __init__.py │ ├── bert_model.py │ ├── biencoder_model.py │ ├── classification.py │ ├── enums.py │ ├── fused_bias_gelu.py │ ├── fused_layer_norm.py │ ├── fused_softmax.py │ ├── gpt_model.py │ ├── language_model.py │ ├── module.py │ ├── multiple_choice.py │ ├── realm_model.py │ ├── rms_norm.py │ ├── t5_model.py │ ├── transformer.py │ ├── utils.py │ └── vision │ │ ├── classification.py │ │ ├── dino.py │ │ ├── esvit_swin_backbone.py │ │ ├── inpainting.py │ │ ├── knn_monitor.py │ │ ├── mit_backbone.py │ │ ├── swin_backbone.py │ │ ├── utils.py │ │ └── vit_backbone.py ├── mpu │ └── tests │ │ ├── __init__.py │ │ ├── commons.py │ │ ├── test_cross_entropy.py │ │ ├── test_data.py │ │ ├── test_initialize.py │ │ ├── test_layers.py │ │ └── test_random.py ├── optimizer │ ├── __init__.py │ ├── clip_grads.py │ ├── distrib_optimizer.py │ ├── grad_scaler.py │ ├── optimizer.py │ └── utils.py ├── optimizer_param_scheduler.py ├── quantization_cuda_builder.py ├── quantization_helper.py ├── static │ └── index.html ├── text_generation │ ├── __init__.py │ ├── api.py │ ├── beam_utils.py │ ├── communication.py │ ├── forward_step.py │ ├── generation.py │ ├── sampling.py │ └── tokenization.py ├── text_generation_server.py ├── theoretical_memory_usage.py ├── timers.py ├── tokenizer │ ├── __init__.py │ ├── bert_tokenization.py │ ├── gpt2_tokenization.py │ └── tokenizer.py ├── training.py └── utils.py ├── pretrain_bert.py ├── pretrain_gpt.py ├── pretrain_ict.py ├── pretrain_retro.py ├── pretrain_t5.py ├── pretrain_vision_classify.py ├── pretrain_vision_dino.py ├── pretrain_vision_inpaint.py ├── pyproject.toml ├── report_theoretical_memory.py ├── sample_scripts ├── accuracy │ ├── 125M │ │ ├── baseline │ │ │ └── train.sh │ │ ├── quantGrad │ │ │ └── train.sh │ │ ├── quantGradwithoutHT │ │ │ └── train.sh │ │ ├── quantWeightDiff │ │ │ └── train.sh │ │ ├── quantWeightDiff_Grad │ │ │ └── train.sh │ │ └── quantWeightDiff_GradwithoutHT │ │ │ └── train.sh │ ├── 1_3B │ │ ├── grad_quant │ │ │ └── train.sh │ │ └── grad_quant_without_hadamard │ │ │ └── train.sh │ └── 6_7B │ │ ├── baseline │ │ └── train.sh │ │ ├── quantWeightDiff │ │ └── train.sh │ │ └── quantWeightDiff_Grad │ │ └── train.sh └── speed │ ├── Exp1 │ ├── 13B_Baseline.sh │ ├── 13B_QWG.sh │ ├── 18B_Baseline.sh │ ├── 18B_QWG.sh │ ├── 1_3B_Baseline.sh │ ├── 1_3B_QWG.sh │ ├── 2_7B_Baseline.sh │ ├── 2_7B_QWG.sh │ ├── 6_7B_Baseline.sh │ ├── 6_7B_QWG.sh │ └── ReadME.md │ ├── Exp2 │ ├── 2_7B_Baseline.sh │ ├── 2_7B_QGrad.sh │ ├── 2_7B_QWG.sh │ ├── 2_7B_QWeightdiff.sh │ └── ReadME.md │ ├── Exp3 │ └── ReadME.md │ ├── ReadME.md │ ├── starter_exp1.sh │ ├── starter_exp2.sh │ └── starter_exp3.sh ├── setup.py ├── tasks ├── data_utils.py ├── ensemble_classifier.py ├── eval_utils.py ├── finetune_utils.py ├── glue │ ├── data.py │ ├── finetune.py │ ├── mnli.py │ └── qqp.py ├── main.py ├── msdp │ ├── README.md │ ├── evaluate.py │ ├── main.py │ ├── metrics.py │ ├── preprocessing.py │ └── prompt.py ├── orqa │ ├── README.md │ ├── evaluate_orqa.py │ ├── evaluate_utils.py │ ├── supervised │ │ ├── data.py │ │ ├── eval_utils.py │ │ └── finetune.py │ └── unsupervised │ │ ├── nq.py │ │ ├── qa_utils.py │ │ └── tokenizers.py ├── race │ ├── data.py │ └── finetune.py ├── vision │ ├── classification │ │ ├── classification.py │ │ └── eval_utils.py │ ├── finetune_utils.py │ ├── main.py │ └── segmentation │ │ ├── cityscapes.py │ │ ├── data.py │ │ ├── finetune_segformer.py │ │ ├── finetune_setr.py │ │ ├── metrics.py │ │ ├── seg_heads.py │ │ ├── seg_models.py │ │ ├── transforms.py │ │ └── utils.py └── zeroshot_gpt │ ├── datasets.py │ ├── detokenizer.py │ └── evaluate.py ├── tests ├── __init__.py ├── functional_tests │ ├── __init__.py │ ├── python_test_utils │ │ ├── __init__.py │ │ ├── check_slurm_job_completion.py │ │ ├── common.py │ │ ├── get_test_results_from_tensorboard_logs.py │ │ ├── jet_test_pipeline.py │ │ ├── multitest_ci_pipeline.py │ │ ├── test_ci_pipeline.py │ │ └── test_resume_checkpoint_pipeline.py │ ├── shell_test_utils │ │ ├── jobwait.sh │ │ ├── run_selene_test_launcher_script.sh │ │ └── run_selene_test_resume_checkpoint_launcher_script.sh │ ├── test_results │ │ ├── bert │ │ │ ├── bert_tp1_pp2_1nodes_50steps.json │ │ │ ├── bert_tp1_pp2_1nodes_50steps_core_enabled.json │ │ │ ├── bert_tp1_pp2_1nodes_50steps_core_enabled_rope_embeddings.json │ │ │ ├── bert_tp1_pp2_1nodes_50steps_core_enabled_sequence_parallel.json │ │ │ ├── bert_tp1_pp4_1nodes_50steps.json │ │ │ ├── bert_tp1_pp4_interleaved_1nodes_50steps.json │ │ │ ├── bert_tp1_pp4_interleaved_1nodes_50steps_core_enabled.json │ │ │ ├── bert_tp2_pp2_1nodes_50steps.json │ │ │ ├── bert_tp2_pp2_1nodes_50steps_core_enabled.json │ │ │ ├── bert_tp4_pp1_1nodes_50steps.json │ │ │ └── bert_tp4_pp1_1nodes_50steps_core_enabled.json │ │ ├── gpt3 │ │ │ ├── gpt3_tp1_pp1_1nodes_50steps_dist_optimizer.json │ │ │ ├── gpt3_tp1_pp1_1nodes_50steps_dist_optimizer_overlap_grad_reduce.json │ │ │ ├── gpt3_tp1_pp1_1nodes_50steps_dist_optimizer_overlap_grad_reduce_param_gather.json │ │ │ ├── gpt3_tp1_pp1_1nodes_50steps_overlap_grad_reduce.json │ │ │ ├── gpt3_tp1_pp2_1nodes_50steps.json │ │ │ ├── gpt3_tp1_pp2_1nodes_50steps_core_enabled.json │ │ │ ├── gpt3_tp1_pp2_1nodes_50steps_core_enabled_rope_embeddings.json │ │ │ ├── gpt3_tp1_pp4_1nodes_50steps.json │ │ │ ├── gpt3_tp1_pp4_1nodes_50steps_core_enabled.json │ │ │ ├── gpt3_tp1_pp4_1nodes_50steps_core_enabled_disable_bias_linear.json │ │ │ ├── gpt3_tp1_pp4_1nodes_50steps_core_enabled_sequence_parallel.json │ │ │ ├── gpt3_tp1_pp4_1nodes_50steps_core_enabled_swiglu.json │ │ │ ├── gpt3_tp1_pp4_1nodes_50steps_core_enabled_untie_embeddings_and_outputs.json │ │ │ ├── gpt3_tp1_pp4_1nodes_50steps_overlap_grad_reduce.json │ │ │ ├── gpt3_tp1_pp4_interleaved_1nodes_50steps.json │ │ │ ├── gpt3_tp1_pp4_interleaved_1nodes_50steps_core_enabled.json │ │ │ ├── gpt3_tp1_pp4_interleaved_1nodes_50steps_dist_optimizer_overlap_grad_reduce.json │ │ │ ├── gpt3_tp1_pp4_interleaved_1nodes_50steps_dist_optimizer_overlap_grad_reduce_param_gather.json │ │ │ ├── gpt3_tp1_pp4_interleaved_1nodes_50steps_overlap_grad_reduce.json │ │ │ ├── gpt3_tp2_pp1_1nodes_50steps_core_enabled_context_parallelism_cp2.json │ │ │ ├── gpt3_tp2_pp1_1nodes_50steps_core_enabled_te_8experts2parallel.json │ │ │ ├── gpt3_tp2_pp1_1nodes_50steps_core_enabled_te_8experts2parallel_groupedGEMM.json │ │ │ ├── gpt3_tp2_pp2_1nodes_50steps.json │ │ │ ├── gpt3_tp2_pp2_1nodes_50steps_4experts.json │ │ │ ├── gpt3_tp2_pp2_1nodes_50steps_core_enabled.json │ │ │ ├── gpt3_tp2_pp2_1nodes_50steps_core_enabled_context_parallelism_cp2.json │ │ │ ├── gpt3_tp2_pp2_1nodes_50steps_core_enabled_te_2experts.json │ │ │ ├── gpt3_tp2_pp2_1nodes_50steps_core_enabled_te_4experts2parallel.json │ │ │ ├── gpt3_tp2_pp2_1nodes_50steps_overlap_grad_reduce.json │ │ │ ├── gpt3_tp2_pp2_1nodes_50steps_te_enabled.json │ │ │ ├── gpt3_tp4_pp1_1nodes_50steps.json │ │ │ ├── gpt3_tp4_pp1_1nodes_50steps_core_enabled.json │ │ │ ├── gpt3_tp4_pp1_1nodes_50steps_dist_optimizer_overlap_grad_reduce.json │ │ │ ├── gpt3_tp4_pp1_1nodes_50steps_dist_optimizer_overlap_grad_reduce_param_gather.json │ │ │ └── gpt3_tp4_pp1_1nodes_50steps_overlap_grad_reduce.json │ │ ├── jet │ │ │ ├── gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-1_args--recompute-granularity-full-recompute-method-uniform-recompute-num-layers-1-.json │ │ │ ├── gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-2.json │ │ │ ├── gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-2_args--position-embedding-type-rope-.json │ │ │ ├── gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4.json │ │ │ ├── gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4_args--disable-bias-linear.json │ │ │ ├── gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4_args--sequence-parallel.json │ │ │ ├── gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4_args--swiglu.json │ │ │ ├── gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4_args--untie-embeddings-and-output-weights.json │ │ │ ├── gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-2_pp-2.json │ │ │ └── gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-4_pp-1.json │ │ ├── retro │ │ │ └── retro_tp1_pp1_1nodes_50steps_core_enabled.json │ │ └── t5 │ │ │ └── t5_tp1_pp1_interleaved_1nodes_100steps_te_enabled_core_enabled.json │ └── test_scripts │ │ ├── bert │ │ ├── pretrain_bert_distributed_resume_checkpoint_test.sh │ │ ├── pretrain_bert_distributed_test.sh │ │ ├── sbatch_bert_distributed_resume_checkpoint_test.sh │ │ └── sbatch_bert_distributed_test.sh │ │ ├── gpt3 │ │ ├── pretrain_gpt3_distributed_resume_checkpoint_test.sh │ │ ├── pretrain_gpt3_distributed_test.sh │ │ ├── sbatch_gpt3_distributed_resume_checkpoint_test.sh │ │ └── sbatch_gpt3_distributed_test.sh │ │ ├── retro │ │ ├── pretrain_retro_distributed_resume_checkpoint_test.sh │ │ ├── pretrain_retro_distributed_test.sh │ │ ├── sbatch_retro_distributed_resume_checkpoint_test.sh │ │ └── sbatch_retro_distributed_test.sh │ │ └── t5 │ │ ├── pretrain_t5_distributed_resume_checkpoint_test.sh │ │ ├── pretrain_t5_distributed_test.sh │ │ ├── sbatch_t5_distributed_resume_checkpoint_test.sh │ │ └── sbatch_t5_distributed_test.sh └── unit_tests │ ├── __init__.py │ ├── data │ ├── test_builder.py │ ├── test_builder_mock_gpt_dataset.py │ ├── test_preprocess_data.py │ └── test_preprocess_mmdata.py │ ├── dist_checkpointing │ ├── __init__.py │ ├── conftest.py │ ├── models │ │ ├── test_gpt_model.py │ │ └── test_mlp_glu.py │ ├── test_mapping.py │ ├── test_optimizer.py │ └── test_serialization.py │ ├── models │ ├── __init__.py │ ├── test_base_embedding.py │ ├── test_bert_model.py │ ├── test_gpt_model.py │ └── test_t5_model.py │ ├── pipeline_parallel │ ├── __init__.py │ └── test_schedules.py │ ├── tensor_parallel │ ├── test_cross_entropy.py │ ├── test_data.py │ ├── test_mappings.py │ ├── test_random.py │ └── test_tensor_parallel_utils.py │ ├── test_basic.py │ ├── test_imports.py │ ├── test_parallel_state.py │ ├── test_utilities.py │ ├── test_utils.py │ └── transformer │ ├── __init__.py │ ├── moe │ ├── test_grouped_mlp.py │ └── test_switch_mlp.py │ ├── test_attention.py │ ├── test_attention_packed_seq.py │ ├── test_core_attention.py │ ├── test_mlp.py │ ├── test_module.py │ ├── test_retro_attention.py │ ├── test_spec_customization.py │ ├── test_transformer_block.py │ └── test_transformer_layer.py └── tools ├── autoformat.sh ├── bert_embedding ├── __init__.py ├── dataset.py ├── embed.py ├── external_libs.py ├── huggingface.py └── utils.py ├── checkpoint ├── loader_llama2_hf.py ├── loader_megatron.py ├── saver_megatron.py └── util.py ├── jet_quant_cuda ├── hadamard │ ├── fast_hadamard_transform_cuda.cu │ └── hadamard_binding.cpp ├── includes │ ├── conversion_utils.h │ ├── dequantization_utils.h │ ├── ds_kernel_utils.h │ ├── fast_hadamard_transform.h │ ├── fast_hadamard_transform_common.h │ ├── fast_hadamard_transform_special.h │ ├── hadamard_binding.h │ ├── memory_access_utils.h │ ├── quantization.h │ ├── quantization_utils.h │ ├── reduction_utils.h │ └── static_switch.h ├── quantization │ ├── dequant_reduce_quant.cu │ ├── dequantize.cu │ ├── dequantize_add.cu │ ├── pt_binding.cpp │ ├── quant_reduce.cu │ ├── quant_reduce_ht.cu │ ├── stochastic_quantize.cu │ ├── stochastic_quantize_ht.cu │ ├── sub_quantize.cu │ ├── swizzled_quantize.cu │ └── swizzled_quantize_ht.cu ├── setup.py └── unit_test │ ├── 2hop_quant_test.py │ ├── dequant_function.py │ ├── fused_dequantize_add.ipynb │ ├── fused_sto_test.ipynb │ ├── sto_quant_test.py │ └── tool_function.py ├── linter.py ├── merge_datasets.py ├── openwebtext ├── README.md ├── add_id.py ├── blacklist_urls.py ├── cleanup_dataset.py ├── cleanup_fix_dataset.py ├── filter_ngrams.py ├── find_duplicates.py ├── group_duplicate_url.py ├── merge_jsons.py └── remove_group_duplicates.py ├── preprocess_data.py ├── preprocess_data_nmt.py ├── preprocess_mmdata.py ├── retro ├── README.md ├── build_db.md ├── cli │ ├── __init__.py │ ├── __main__.py │ └── cli.py ├── db │ ├── __init__.py │ ├── build.py │ ├── dataset.py │ └── utils.py ├── examples │ ├── Dockerfile │ ├── preprocess_data.sh │ └── pretrain_model.sh ├── external_libs.py ├── index │ ├── __init__.py │ ├── build.py │ ├── factory.py │ ├── index.py │ ├── indexes │ │ ├── __init__.py │ │ ├── faiss_base.py │ │ └── faiss_par_add.py │ └── utils.py ├── main.py ├── query │ ├── __init__.py │ ├── chunk_dataset.py │ ├── multi_split_gpt_dataset.py │ ├── query.py │ ├── retro_dataset.py │ └── utils.py ├── sft │ ├── README.md │ ├── dataset_conv.py │ ├── open_inst.sh │ ├── sft_retro.py │ └── sft_retro_lm.sh ├── text_generation │ ├── evaluate.py │ ├── metrics.py │ ├── retro_api.py │ ├── retro_generate.sh │ ├── retro_generation.py │ └── retro_text_generation.py └── utils.py ├── run_text_generation_server.py └── text_generation_cli.py /CODEOWNERS: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/CODEOWNERS -------------------------------------------------------------------------------- /CONTRIBUTING.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/CONTRIBUTING.md -------------------------------------------------------------------------------- /Dockerfile.ci: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/Dockerfile.ci -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/LICENSE -------------------------------------------------------------------------------- /MANIFEST.in: -------------------------------------------------------------------------------- 1 | include megatron/core/requirements.txt 2 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/README.md -------------------------------------------------------------------------------- /dataset-preprocessing/the-pile/download.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/dataset-preprocessing/the-pile/download.py -------------------------------------------------------------------------------- /dataset-preprocessing/the-pile/download.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/dataset-preprocessing/the-pile/download.sh -------------------------------------------------------------------------------- /dataset-preprocessing/the-pile/process.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/dataset-preprocessing/the-pile/process.sh -------------------------------------------------------------------------------- /dataset-preprocessing/wikitext-2/download.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/dataset-preprocessing/wikitext-2/download.py -------------------------------------------------------------------------------- /dataset-preprocessing/wikitext-2/preprocess.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/dataset-preprocessing/wikitext-2/preprocess.sh -------------------------------------------------------------------------------- /docs/llama2.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/llama2.md -------------------------------------------------------------------------------- /docs/source/api-guide/dist_checkpointing.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/api-guide/dist_checkpointing.rst -------------------------------------------------------------------------------- /docs/source/api-guide/dist_checkpointing.strategies.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/api-guide/dist_checkpointing.strategies.rst -------------------------------------------------------------------------------- /docs/source/api-guide/distributed.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/api-guide/distributed.rst -------------------------------------------------------------------------------- /docs/source/api-guide/fusions.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/api-guide/fusions.rst -------------------------------------------------------------------------------- /docs/source/api-guide/index.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/api-guide/index.rst -------------------------------------------------------------------------------- /docs/source/api-guide/models.gpt.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/api-guide/models.gpt.rst -------------------------------------------------------------------------------- /docs/source/api-guide/models.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/api-guide/models.rst -------------------------------------------------------------------------------- /docs/source/api-guide/pipeline_parallel.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/api-guide/pipeline_parallel.rst -------------------------------------------------------------------------------- /docs/source/api-guide/tensor_parallel.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/api-guide/tensor_parallel.rst -------------------------------------------------------------------------------- /docs/source/api-guide/transformer.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/api-guide/transformer.rst -------------------------------------------------------------------------------- /docs/source/distrib_optimizer.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/distrib_optimizer.md -------------------------------------------------------------------------------- /docs/source/images/distrib_optimizer/data_flow.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/images/distrib_optimizer/data_flow.png -------------------------------------------------------------------------------- /docs/source/images/distrib_optimizer/sharding_scheme.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/images/distrib_optimizer/sharding_scheme.png -------------------------------------------------------------------------------- /docs/source/index.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/index.rst -------------------------------------------------------------------------------- /docs/source/user-guide/index.rst: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/docs/source/user-guide/index.rst -------------------------------------------------------------------------------- /examples/bert/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/bert/README.md -------------------------------------------------------------------------------- /examples/bert/train_bert_340m_distributed.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/bert/train_bert_340m_distributed.sh -------------------------------------------------------------------------------- /examples/detxoify_lm/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/detxoify_lm/README.md -------------------------------------------------------------------------------- /examples/detxoify_lm/annotations/filter-selfgeneration.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/detxoify_lm/annotations/filter-selfgeneration.py -------------------------------------------------------------------------------- /examples/detxoify_lm/annotations/perspective_api_annotate.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/detxoify_lm/annotations/perspective_api_annotate.py -------------------------------------------------------------------------------- /examples/detxoify_lm/annotations/preprocess.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/detxoify_lm/annotations/preprocess.sh -------------------------------------------------------------------------------- /examples/detxoify_lm/finetune_gpt.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/detxoify_lm/finetune_gpt.py -------------------------------------------------------------------------------- /examples/detxoify_lm/finetune_gpt_distributed-1.3b.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/detxoify_lm/finetune_gpt_distributed-1.3b.sh -------------------------------------------------------------------------------- /examples/detxoify_lm/generate-1.3b.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/detxoify_lm/generate-1.3b.sh -------------------------------------------------------------------------------- /examples/detxoify_lm/generate_samples_gpt.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/detxoify_lm/generate_samples_gpt.py -------------------------------------------------------------------------------- /examples/detxoify_lm/perspective_api.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/detxoify_lm/perspective_api.py -------------------------------------------------------------------------------- /examples/detxoify_lm/self_generation/selfgenerate-1.3b-unconditional.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/detxoify_lm/self_generation/selfgenerate-1.3b-unconditional.sh -------------------------------------------------------------------------------- /examples/evaluate_retriever_nq.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/evaluate_retriever_nq.sh -------------------------------------------------------------------------------- /examples/evaluate_zeroshot_gpt.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/evaluate_zeroshot_gpt.sh -------------------------------------------------------------------------------- /examples/finetune_mnli_distributed.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/finetune_mnli_distributed.sh -------------------------------------------------------------------------------- /examples/finetune_race_distributed.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/finetune_race_distributed.sh -------------------------------------------------------------------------------- /examples/finetune_retriever_distributed.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/finetune_retriever_distributed.sh -------------------------------------------------------------------------------- /examples/gpt3/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/gpt3/README.md -------------------------------------------------------------------------------- /examples/gpt3/train_gpt3_175b_distributed.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/gpt3/train_gpt3_175b_distributed.sh -------------------------------------------------------------------------------- /examples/merge_mp_bert.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/merge_mp_bert.sh -------------------------------------------------------------------------------- /examples/msdp/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/msdp/README.md -------------------------------------------------------------------------------- /examples/msdp/data_processing.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/msdp/data_processing.sh -------------------------------------------------------------------------------- /examples/msdp/eval_knwl_generation.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/msdp/eval_knwl_generation.sh -------------------------------------------------------------------------------- /examples/msdp/eval_resp_generation.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/msdp/eval_resp_generation.sh -------------------------------------------------------------------------------- /examples/msdp/prep_resp_gen.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/msdp/prep_resp_gen.sh -------------------------------------------------------------------------------- /examples/msdp/prompt_knwl_gen.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/msdp/prompt_knwl_gen.sh -------------------------------------------------------------------------------- /examples/msdp/prompt_resp_gen.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/msdp/prompt_resp_gen.sh -------------------------------------------------------------------------------- /examples/pretrain_bert.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_bert.sh -------------------------------------------------------------------------------- /examples/pretrain_bert_distributed.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_bert_distributed.sh -------------------------------------------------------------------------------- /examples/pretrain_bert_distributed_with_mp.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_bert_distributed_with_mp.sh -------------------------------------------------------------------------------- /examples/pretrain_gpt.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_gpt.sh -------------------------------------------------------------------------------- /examples/pretrain_gpt3_175B.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_gpt3_175B.sh -------------------------------------------------------------------------------- /examples/pretrain_gpt_distributed.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_gpt_distributed.sh -------------------------------------------------------------------------------- /examples/pretrain_gpt_distributed_with_mp.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_gpt_distributed_with_mp.sh -------------------------------------------------------------------------------- /examples/pretrain_ict.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_ict.sh -------------------------------------------------------------------------------- /examples/pretrain_t5.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_t5.sh -------------------------------------------------------------------------------- /examples/pretrain_t5_distributed.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_t5_distributed.sh -------------------------------------------------------------------------------- /examples/pretrain_t5_distributed_with_mp.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_t5_distributed_with_mp.sh -------------------------------------------------------------------------------- /examples/pretrain_vision_classify.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_vision_classify.sh -------------------------------------------------------------------------------- /examples/pretrain_vision_dino.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_vision_dino.sh -------------------------------------------------------------------------------- /examples/pretrain_vision_inpaint.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/pretrain_vision_inpaint.sh -------------------------------------------------------------------------------- /examples/run_text_generation_server_345M.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/run_text_generation_server_345M.sh -------------------------------------------------------------------------------- /examples/run_text_generation_server_345M_8_tensor_parallel.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/run_text_generation_server_345M_8_tensor_parallel.sh -------------------------------------------------------------------------------- /examples/sc21/CONFIG.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/CONFIG.sh -------------------------------------------------------------------------------- /examples/sc21/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/README.md -------------------------------------------------------------------------------- /examples/sc21/SBATCH.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/SBATCH.sh -------------------------------------------------------------------------------- /examples/sc21/SRUN.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/SRUN.sh -------------------------------------------------------------------------------- /examples/sc21/run_figure_11.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/run_figure_11.sh -------------------------------------------------------------------------------- /examples/sc21/run_figure_12.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/run_figure_12.sh -------------------------------------------------------------------------------- /examples/sc21/run_figure_13.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/run_figure_13.sh -------------------------------------------------------------------------------- /examples/sc21/run_figure_14.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/run_figure_14.sh -------------------------------------------------------------------------------- /examples/sc21/run_figure_15.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/run_figure_15.sh -------------------------------------------------------------------------------- /examples/sc21/run_figure_16.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/run_figure_16.sh -------------------------------------------------------------------------------- /examples/sc21/run_figure_17.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/run_figure_17.sh -------------------------------------------------------------------------------- /examples/sc21/run_figure_18.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/run_figure_18.sh -------------------------------------------------------------------------------- /examples/sc21/run_table_1.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/sc21/run_table_1.sh -------------------------------------------------------------------------------- /examples/t5/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/t5/README.md -------------------------------------------------------------------------------- /examples/t5/t5_mcore_train_curve.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/t5/t5_mcore_train_curve.png -------------------------------------------------------------------------------- /examples/t5/train_t5_220m_distributed.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/examples/t5/train_t5_220m_distributed.sh -------------------------------------------------------------------------------- /images/Achieved_petaFLOPs.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/images/Achieved_petaFLOPs.png -------------------------------------------------------------------------------- /images/cases_april2021.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/images/cases_april2021.png -------------------------------------------------------------------------------- /jet-tests.yml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/jet-tests.yml -------------------------------------------------------------------------------- /licenses/LICENSE-deepspeed.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/licenses/LICENSE-deepspeed.txt -------------------------------------------------------------------------------- /licenses/LICENSE-fast-hadamard-transform.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/licenses/LICENSE-fast-hadamard-transform.txt -------------------------------------------------------------------------------- /licenses/LICENSE-megatron-lm.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/licenses/LICENSE-megatron-lm.txt -------------------------------------------------------------------------------- /megatron/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/__init__.py -------------------------------------------------------------------------------- /megatron/arguments.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/arguments.py -------------------------------------------------------------------------------- /megatron/checkpointing.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/checkpointing.py -------------------------------------------------------------------------------- /megatron/core/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/README.md -------------------------------------------------------------------------------- /megatron/core/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/__init__.py -------------------------------------------------------------------------------- /megatron/core/datasets/Makefile: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/datasets/Makefile -------------------------------------------------------------------------------- /megatron/core/datasets/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /megatron/core/datasets/blended_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/datasets/blended_dataset.py -------------------------------------------------------------------------------- /megatron/core/datasets/blended_megatron_dataset_builder.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/datasets/blended_megatron_dataset_builder.py -------------------------------------------------------------------------------- /megatron/core/datasets/blended_megatron_dataset_config.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/datasets/blended_megatron_dataset_config.py -------------------------------------------------------------------------------- /megatron/core/datasets/gpt_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/datasets/gpt_dataset.py -------------------------------------------------------------------------------- /megatron/core/datasets/helpers.cpp: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/datasets/helpers.cpp -------------------------------------------------------------------------------- /megatron/core/datasets/indexed_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/datasets/indexed_dataset.py -------------------------------------------------------------------------------- /megatron/core/datasets/megatron_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/datasets/megatron_dataset.py -------------------------------------------------------------------------------- /megatron/core/datasets/megatron_tokenizer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/datasets/megatron_tokenizer.py -------------------------------------------------------------------------------- /megatron/core/datasets/readme.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/datasets/readme.md -------------------------------------------------------------------------------- /megatron/core/datasets/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/datasets/utils.py -------------------------------------------------------------------------------- /megatron/core/dist_checkpointing/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/dist_checkpointing/__init__.py -------------------------------------------------------------------------------- /megatron/core/dist_checkpointing/core.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/dist_checkpointing/core.py -------------------------------------------------------------------------------- /megatron/core/dist_checkpointing/dict_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/dist_checkpointing/dict_utils.py -------------------------------------------------------------------------------- /megatron/core/dist_checkpointing/mapping.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/dist_checkpointing/mapping.py -------------------------------------------------------------------------------- /megatron/core/dist_checkpointing/optimizer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/dist_checkpointing/optimizer.py -------------------------------------------------------------------------------- /megatron/core/dist_checkpointing/serialization.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/dist_checkpointing/serialization.py -------------------------------------------------------------------------------- /megatron/core/dist_checkpointing/strategies/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/dist_checkpointing/strategies/__init__.py -------------------------------------------------------------------------------- /megatron/core/dist_checkpointing/strategies/base.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/dist_checkpointing/strategies/base.py -------------------------------------------------------------------------------- /megatron/core/dist_checkpointing/strategies/tensorstore.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/dist_checkpointing/strategies/tensorstore.py -------------------------------------------------------------------------------- /megatron/core/dist_checkpointing/strategies/two_stage.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/dist_checkpointing/strategies/two_stage.py -------------------------------------------------------------------------------- /megatron/core/dist_checkpointing/strategies/zarr.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/dist_checkpointing/strategies/zarr.py -------------------------------------------------------------------------------- /megatron/core/dist_checkpointing/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/dist_checkpointing/utils.py -------------------------------------------------------------------------------- /megatron/core/distributed/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/distributed/__init__.py -------------------------------------------------------------------------------- /megatron/core/distributed/distributed_data_parallel.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/distributed/distributed_data_parallel.py -------------------------------------------------------------------------------- /megatron/core/distributed/finalize_model_grads.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/distributed/finalize_model_grads.py -------------------------------------------------------------------------------- /megatron/core/distributed/grad_buffer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/distributed/grad_buffer.py -------------------------------------------------------------------------------- /megatron/core/enums.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/enums.py -------------------------------------------------------------------------------- /megatron/core/fusions/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /megatron/core/fusions/fused_bias_dropout.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/fusions/fused_bias_dropout.py -------------------------------------------------------------------------------- /megatron/core/fusions/fused_bias_gelu.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/fusions/fused_bias_gelu.py -------------------------------------------------------------------------------- /megatron/core/fusions/fused_bias_swiglu.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/fusions/fused_bias_swiglu.py -------------------------------------------------------------------------------- /megatron/core/fusions/fused_layer_norm.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/fusions/fused_layer_norm.py -------------------------------------------------------------------------------- /megatron/core/fusions/fused_softmax.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/fusions/fused_softmax.py -------------------------------------------------------------------------------- /megatron/core/inference_params.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/inference_params.py -------------------------------------------------------------------------------- /megatron/core/jit.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/jit.py -------------------------------------------------------------------------------- /megatron/core/model_parallel_config.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/model_parallel_config.py -------------------------------------------------------------------------------- /megatron/core/models/T5/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/T5/__init__.py -------------------------------------------------------------------------------- /megatron/core/models/T5/t5_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/T5/t5_model.py -------------------------------------------------------------------------------- /megatron/core/models/T5/t5_spec.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/T5/t5_spec.py -------------------------------------------------------------------------------- /megatron/core/models/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /megatron/core/models/bert/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /megatron/core/models/bert/bert_layer_specs.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/bert/bert_layer_specs.py -------------------------------------------------------------------------------- /megatron/core/models/bert/bert_lm_head.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/bert/bert_lm_head.py -------------------------------------------------------------------------------- /megatron/core/models/bert/bert_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/bert/bert_model.py -------------------------------------------------------------------------------- /megatron/core/models/bert/pooler.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/bert/pooler.py -------------------------------------------------------------------------------- /megatron/core/models/common/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /megatron/core/models/common/embeddings/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /megatron/core/models/common/embeddings/language_model_embedding.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/common/embeddings/language_model_embedding.py -------------------------------------------------------------------------------- /megatron/core/models/common/embeddings/rotary_pos_embedding.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/common/embeddings/rotary_pos_embedding.py -------------------------------------------------------------------------------- /megatron/core/models/common/language_module/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /megatron/core/models/common/language_module/language_module.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/common/language_module/language_module.py -------------------------------------------------------------------------------- /megatron/core/models/gpt/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/gpt/__init__.py -------------------------------------------------------------------------------- /megatron/core/models/gpt/gpt_layer_specs.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/gpt/gpt_layer_specs.py -------------------------------------------------------------------------------- /megatron/core/models/gpt/gpt_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/gpt/gpt_model.py -------------------------------------------------------------------------------- /megatron/core/models/retro/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/retro/__init__.py -------------------------------------------------------------------------------- /megatron/core/models/retro/base_attention.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/retro/base_attention.py -------------------------------------------------------------------------------- /megatron/core/models/retro/config.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/retro/config.py -------------------------------------------------------------------------------- /megatron/core/models/retro/decoder_attention.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/retro/decoder_attention.py -------------------------------------------------------------------------------- /megatron/core/models/retro/decoder_spec.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/retro/decoder_spec.py -------------------------------------------------------------------------------- /megatron/core/models/retro/encoder_attention.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/retro/encoder_attention.py -------------------------------------------------------------------------------- /megatron/core/models/retro/encoder_spec.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/retro/encoder_spec.py -------------------------------------------------------------------------------- /megatron/core/models/retro/model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/models/retro/model.py -------------------------------------------------------------------------------- /megatron/core/package_info.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/package_info.py -------------------------------------------------------------------------------- /megatron/core/packed_seq_params.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/packed_seq_params.py -------------------------------------------------------------------------------- /megatron/core/parallel_state.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/parallel_state.py -------------------------------------------------------------------------------- /megatron/core/pipeline_parallel/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/pipeline_parallel/__init__.py -------------------------------------------------------------------------------- /megatron/core/pipeline_parallel/p2p_communication.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/pipeline_parallel/p2p_communication.py -------------------------------------------------------------------------------- /megatron/core/pipeline_parallel/schedules.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/pipeline_parallel/schedules.py -------------------------------------------------------------------------------- /megatron/core/requirements.txt: -------------------------------------------------------------------------------- 1 | torch -------------------------------------------------------------------------------- /megatron/core/tensor_parallel/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/tensor_parallel/__init__.py -------------------------------------------------------------------------------- /megatron/core/tensor_parallel/cross_entropy.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/tensor_parallel/cross_entropy.py -------------------------------------------------------------------------------- /megatron/core/tensor_parallel/data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/tensor_parallel/data.py -------------------------------------------------------------------------------- /megatron/core/tensor_parallel/layers.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/tensor_parallel/layers.py -------------------------------------------------------------------------------- /megatron/core/tensor_parallel/mappings.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/tensor_parallel/mappings.py -------------------------------------------------------------------------------- /megatron/core/tensor_parallel/random.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/tensor_parallel/random.py -------------------------------------------------------------------------------- /megatron/core/tensor_parallel/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/tensor_parallel/utils.py -------------------------------------------------------------------------------- /megatron/core/transformer/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/__init__.py -------------------------------------------------------------------------------- /megatron/core/transformer/attention.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/attention.py -------------------------------------------------------------------------------- /megatron/core/transformer/custom_layers/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /megatron/core/transformer/custom_layers/transformer_engine.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/custom_layers/transformer_engine.py -------------------------------------------------------------------------------- /megatron/core/transformer/dot_product_attention.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/dot_product_attention.py -------------------------------------------------------------------------------- /megatron/core/transformer/enums.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/enums.py -------------------------------------------------------------------------------- /megatron/core/transformer/identity_op.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/identity_op.py -------------------------------------------------------------------------------- /megatron/core/transformer/mlp.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/mlp.py -------------------------------------------------------------------------------- /megatron/core/transformer/module.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/module.py -------------------------------------------------------------------------------- /megatron/core/transformer/moe/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /megatron/core/transformer/moe/base_moe_layer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/moe/base_moe_layer.py -------------------------------------------------------------------------------- /megatron/core/transformer/moe/grouped_gemm_util.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/moe/grouped_gemm_util.py -------------------------------------------------------------------------------- /megatron/core/transformer/moe/grouped_mlp.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/moe/grouped_mlp.py -------------------------------------------------------------------------------- /megatron/core/transformer/moe/switch_mlp.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/moe/switch_mlp.py -------------------------------------------------------------------------------- /megatron/core/transformer/spec_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/spec_utils.py -------------------------------------------------------------------------------- /megatron/core/transformer/transformer_block.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/transformer_block.py -------------------------------------------------------------------------------- /megatron/core/transformer/transformer_config.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/transformer_config.py -------------------------------------------------------------------------------- /megatron/core/transformer/transformer_layer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/transformer_layer.py -------------------------------------------------------------------------------- /megatron/core/transformer/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/transformer/utils.py -------------------------------------------------------------------------------- /megatron/core/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/core/utils.py -------------------------------------------------------------------------------- /megatron/data/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /megatron/data/autoaugment.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/autoaugment.py -------------------------------------------------------------------------------- /megatron/data/bert_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/bert_dataset.py -------------------------------------------------------------------------------- /megatron/data/biencoder_dataset_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/biencoder_dataset_utils.py -------------------------------------------------------------------------------- /megatron/data/data_samplers.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/data_samplers.py -------------------------------------------------------------------------------- /megatron/data/dataset_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/dataset_utils.py -------------------------------------------------------------------------------- /megatron/data/ict_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/ict_dataset.py -------------------------------------------------------------------------------- /megatron/data/image_folder.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/image_folder.py -------------------------------------------------------------------------------- /megatron/data/multimodal_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/multimodal_dataset.py -------------------------------------------------------------------------------- /megatron/data/orqa_wiki_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/orqa_wiki_dataset.py -------------------------------------------------------------------------------- /megatron/data/realm_dataset_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/realm_dataset_utils.py -------------------------------------------------------------------------------- /megatron/data/realm_index.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/realm_index.py -------------------------------------------------------------------------------- /megatron/data/t5_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/t5_dataset.py -------------------------------------------------------------------------------- /megatron/data/vit_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/data/vit_dataset.py -------------------------------------------------------------------------------- /megatron/dist_signal_handler.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/dist_signal_handler.py -------------------------------------------------------------------------------- /megatron/fp16_deprecated/loss_scaler.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/fp16_deprecated/loss_scaler.py -------------------------------------------------------------------------------- /megatron/fused_kernels/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/fused_kernels/__init__.py -------------------------------------------------------------------------------- /megatron/fused_kernels/compat.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/fused_kernels/compat.h -------------------------------------------------------------------------------- /megatron/fused_kernels/tests/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /megatron/fused_kernels/tests/test_fused_kernels.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/fused_kernels/tests/test_fused_kernels.py -------------------------------------------------------------------------------- /megatron/fused_kernels/type_shim.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/fused_kernels/type_shim.h -------------------------------------------------------------------------------- /megatron/global_vars.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/global_vars.py -------------------------------------------------------------------------------- /megatron/indexer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/indexer.py -------------------------------------------------------------------------------- /megatron/initialize.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/initialize.py -------------------------------------------------------------------------------- /megatron/log_handler.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/log_handler.py -------------------------------------------------------------------------------- /megatron/memory.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/memory.py -------------------------------------------------------------------------------- /megatron/microbatches.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/microbatches.py -------------------------------------------------------------------------------- /megatron/model/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/__init__.py -------------------------------------------------------------------------------- /megatron/model/bert_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/bert_model.py -------------------------------------------------------------------------------- /megatron/model/biencoder_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/biencoder_model.py -------------------------------------------------------------------------------- /megatron/model/classification.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/classification.py -------------------------------------------------------------------------------- /megatron/model/enums.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/enums.py -------------------------------------------------------------------------------- /megatron/model/fused_bias_gelu.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/fused_bias_gelu.py -------------------------------------------------------------------------------- /megatron/model/fused_layer_norm.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/fused_layer_norm.py -------------------------------------------------------------------------------- /megatron/model/fused_softmax.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/fused_softmax.py -------------------------------------------------------------------------------- /megatron/model/gpt_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/gpt_model.py -------------------------------------------------------------------------------- /megatron/model/language_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/language_model.py -------------------------------------------------------------------------------- /megatron/model/module.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/module.py -------------------------------------------------------------------------------- /megatron/model/multiple_choice.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/multiple_choice.py -------------------------------------------------------------------------------- /megatron/model/realm_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/realm_model.py -------------------------------------------------------------------------------- /megatron/model/rms_norm.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/rms_norm.py -------------------------------------------------------------------------------- /megatron/model/t5_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/t5_model.py -------------------------------------------------------------------------------- /megatron/model/transformer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/transformer.py -------------------------------------------------------------------------------- /megatron/model/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/utils.py -------------------------------------------------------------------------------- /megatron/model/vision/classification.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/vision/classification.py -------------------------------------------------------------------------------- /megatron/model/vision/dino.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/vision/dino.py -------------------------------------------------------------------------------- /megatron/model/vision/esvit_swin_backbone.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/vision/esvit_swin_backbone.py -------------------------------------------------------------------------------- /megatron/model/vision/inpainting.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/vision/inpainting.py -------------------------------------------------------------------------------- /megatron/model/vision/knn_monitor.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/vision/knn_monitor.py -------------------------------------------------------------------------------- /megatron/model/vision/mit_backbone.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/vision/mit_backbone.py -------------------------------------------------------------------------------- /megatron/model/vision/swin_backbone.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/vision/swin_backbone.py -------------------------------------------------------------------------------- /megatron/model/vision/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/vision/utils.py -------------------------------------------------------------------------------- /megatron/model/vision/vit_backbone.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/model/vision/vit_backbone.py -------------------------------------------------------------------------------- /megatron/mpu/tests/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /megatron/mpu/tests/commons.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/mpu/tests/commons.py -------------------------------------------------------------------------------- /megatron/mpu/tests/test_cross_entropy.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/mpu/tests/test_cross_entropy.py -------------------------------------------------------------------------------- /megatron/mpu/tests/test_data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/mpu/tests/test_data.py -------------------------------------------------------------------------------- /megatron/mpu/tests/test_initialize.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/mpu/tests/test_initialize.py -------------------------------------------------------------------------------- /megatron/mpu/tests/test_layers.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/mpu/tests/test_layers.py -------------------------------------------------------------------------------- /megatron/mpu/tests/test_random.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/mpu/tests/test_random.py -------------------------------------------------------------------------------- /megatron/optimizer/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/optimizer/__init__.py -------------------------------------------------------------------------------- /megatron/optimizer/clip_grads.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/optimizer/clip_grads.py -------------------------------------------------------------------------------- /megatron/optimizer/distrib_optimizer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/optimizer/distrib_optimizer.py -------------------------------------------------------------------------------- /megatron/optimizer/grad_scaler.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/optimizer/grad_scaler.py -------------------------------------------------------------------------------- /megatron/optimizer/optimizer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/optimizer/optimizer.py -------------------------------------------------------------------------------- /megatron/optimizer/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/optimizer/utils.py -------------------------------------------------------------------------------- /megatron/optimizer_param_scheduler.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/optimizer_param_scheduler.py -------------------------------------------------------------------------------- /megatron/quantization_cuda_builder.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/quantization_cuda_builder.py -------------------------------------------------------------------------------- /megatron/quantization_helper.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/quantization_helper.py -------------------------------------------------------------------------------- /megatron/static/index.html: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/static/index.html -------------------------------------------------------------------------------- /megatron/text_generation/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/text_generation/__init__.py -------------------------------------------------------------------------------- /megatron/text_generation/api.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/text_generation/api.py -------------------------------------------------------------------------------- /megatron/text_generation/beam_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/text_generation/beam_utils.py -------------------------------------------------------------------------------- /megatron/text_generation/communication.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/text_generation/communication.py -------------------------------------------------------------------------------- /megatron/text_generation/forward_step.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/text_generation/forward_step.py -------------------------------------------------------------------------------- /megatron/text_generation/generation.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/text_generation/generation.py -------------------------------------------------------------------------------- /megatron/text_generation/sampling.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/text_generation/sampling.py -------------------------------------------------------------------------------- /megatron/text_generation/tokenization.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/text_generation/tokenization.py -------------------------------------------------------------------------------- /megatron/text_generation_server.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/text_generation_server.py -------------------------------------------------------------------------------- /megatron/theoretical_memory_usage.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/theoretical_memory_usage.py -------------------------------------------------------------------------------- /megatron/timers.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/timers.py -------------------------------------------------------------------------------- /megatron/tokenizer/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/tokenizer/__init__.py -------------------------------------------------------------------------------- /megatron/tokenizer/bert_tokenization.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/tokenizer/bert_tokenization.py -------------------------------------------------------------------------------- /megatron/tokenizer/gpt2_tokenization.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/tokenizer/gpt2_tokenization.py -------------------------------------------------------------------------------- /megatron/tokenizer/tokenizer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/tokenizer/tokenizer.py -------------------------------------------------------------------------------- /megatron/training.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/training.py -------------------------------------------------------------------------------- /megatron/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/megatron/utils.py -------------------------------------------------------------------------------- /pretrain_bert.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/pretrain_bert.py -------------------------------------------------------------------------------- /pretrain_gpt.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/pretrain_gpt.py -------------------------------------------------------------------------------- /pretrain_ict.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/pretrain_ict.py -------------------------------------------------------------------------------- /pretrain_retro.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/pretrain_retro.py -------------------------------------------------------------------------------- /pretrain_t5.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/pretrain_t5.py -------------------------------------------------------------------------------- /pretrain_vision_classify.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/pretrain_vision_classify.py -------------------------------------------------------------------------------- /pretrain_vision_dino.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/pretrain_vision_dino.py -------------------------------------------------------------------------------- /pretrain_vision_inpaint.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/pretrain_vision_inpaint.py -------------------------------------------------------------------------------- /pyproject.toml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/pyproject.toml -------------------------------------------------------------------------------- /report_theoretical_memory.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/report_theoretical_memory.py -------------------------------------------------------------------------------- /sample_scripts/accuracy/125M/baseline/train.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/accuracy/125M/baseline/train.sh -------------------------------------------------------------------------------- /sample_scripts/accuracy/125M/quantGrad/train.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/accuracy/125M/quantGrad/train.sh -------------------------------------------------------------------------------- /sample_scripts/accuracy/125M/quantGradwithoutHT/train.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/accuracy/125M/quantGradwithoutHT/train.sh -------------------------------------------------------------------------------- /sample_scripts/accuracy/125M/quantWeightDiff/train.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/accuracy/125M/quantWeightDiff/train.sh -------------------------------------------------------------------------------- /sample_scripts/accuracy/125M/quantWeightDiff_Grad/train.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/accuracy/125M/quantWeightDiff_Grad/train.sh -------------------------------------------------------------------------------- /sample_scripts/accuracy/125M/quantWeightDiff_GradwithoutHT/train.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/accuracy/125M/quantWeightDiff_GradwithoutHT/train.sh -------------------------------------------------------------------------------- /sample_scripts/accuracy/1_3B/grad_quant/train.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/accuracy/1_3B/grad_quant/train.sh -------------------------------------------------------------------------------- /sample_scripts/accuracy/1_3B/grad_quant_without_hadamard/train.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/accuracy/1_3B/grad_quant_without_hadamard/train.sh -------------------------------------------------------------------------------- /sample_scripts/accuracy/6_7B/baseline/train.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/accuracy/6_7B/baseline/train.sh -------------------------------------------------------------------------------- /sample_scripts/accuracy/6_7B/quantWeightDiff/train.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/accuracy/6_7B/quantWeightDiff/train.sh -------------------------------------------------------------------------------- /sample_scripts/accuracy/6_7B/quantWeightDiff_Grad/train.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/accuracy/6_7B/quantWeightDiff_Grad/train.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp1/13B_Baseline.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp1/13B_Baseline.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp1/13B_QWG.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp1/13B_QWG.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp1/18B_Baseline.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp1/18B_Baseline.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp1/18B_QWG.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp1/18B_QWG.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp1/1_3B_Baseline.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp1/1_3B_Baseline.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp1/1_3B_QWG.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp1/1_3B_QWG.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp1/2_7B_Baseline.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp1/2_7B_Baseline.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp1/2_7B_QWG.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp1/2_7B_QWG.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp1/6_7B_Baseline.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp1/6_7B_Baseline.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp1/6_7B_QWG.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp1/6_7B_QWG.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp1/ReadME.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp1/ReadME.md -------------------------------------------------------------------------------- /sample_scripts/speed/Exp2/2_7B_Baseline.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp2/2_7B_Baseline.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp2/2_7B_QGrad.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp2/2_7B_QGrad.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp2/2_7B_QWG.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp2/2_7B_QWG.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp2/2_7B_QWeightdiff.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp2/2_7B_QWeightdiff.sh -------------------------------------------------------------------------------- /sample_scripts/speed/Exp2/ReadME.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp2/ReadME.md -------------------------------------------------------------------------------- /sample_scripts/speed/Exp3/ReadME.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/Exp3/ReadME.md -------------------------------------------------------------------------------- /sample_scripts/speed/ReadME.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/ReadME.md -------------------------------------------------------------------------------- /sample_scripts/speed/starter_exp1.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/starter_exp1.sh -------------------------------------------------------------------------------- /sample_scripts/speed/starter_exp2.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/starter_exp2.sh -------------------------------------------------------------------------------- /sample_scripts/speed/starter_exp3.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/sample_scripts/speed/starter_exp3.sh -------------------------------------------------------------------------------- /setup.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/setup.py -------------------------------------------------------------------------------- /tasks/data_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/data_utils.py -------------------------------------------------------------------------------- /tasks/ensemble_classifier.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/ensemble_classifier.py -------------------------------------------------------------------------------- /tasks/eval_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/eval_utils.py -------------------------------------------------------------------------------- /tasks/finetune_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/finetune_utils.py -------------------------------------------------------------------------------- /tasks/glue/data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/glue/data.py -------------------------------------------------------------------------------- /tasks/glue/finetune.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/glue/finetune.py -------------------------------------------------------------------------------- /tasks/glue/mnli.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/glue/mnli.py -------------------------------------------------------------------------------- /tasks/glue/qqp.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/glue/qqp.py -------------------------------------------------------------------------------- /tasks/main.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/main.py -------------------------------------------------------------------------------- /tasks/msdp/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/msdp/README.md -------------------------------------------------------------------------------- /tasks/msdp/evaluate.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/msdp/evaluate.py -------------------------------------------------------------------------------- /tasks/msdp/main.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/msdp/main.py -------------------------------------------------------------------------------- /tasks/msdp/metrics.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/msdp/metrics.py -------------------------------------------------------------------------------- /tasks/msdp/preprocessing.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/msdp/preprocessing.py -------------------------------------------------------------------------------- /tasks/msdp/prompt.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/msdp/prompt.py -------------------------------------------------------------------------------- /tasks/orqa/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/orqa/README.md -------------------------------------------------------------------------------- /tasks/orqa/evaluate_orqa.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/orqa/evaluate_orqa.py -------------------------------------------------------------------------------- /tasks/orqa/evaluate_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/orqa/evaluate_utils.py -------------------------------------------------------------------------------- /tasks/orqa/supervised/data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/orqa/supervised/data.py -------------------------------------------------------------------------------- /tasks/orqa/supervised/eval_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/orqa/supervised/eval_utils.py -------------------------------------------------------------------------------- /tasks/orqa/supervised/finetune.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/orqa/supervised/finetune.py -------------------------------------------------------------------------------- /tasks/orqa/unsupervised/nq.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/orqa/unsupervised/nq.py -------------------------------------------------------------------------------- /tasks/orqa/unsupervised/qa_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/orqa/unsupervised/qa_utils.py -------------------------------------------------------------------------------- /tasks/orqa/unsupervised/tokenizers.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/orqa/unsupervised/tokenizers.py -------------------------------------------------------------------------------- /tasks/race/data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/race/data.py -------------------------------------------------------------------------------- /tasks/race/finetune.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/race/finetune.py -------------------------------------------------------------------------------- /tasks/vision/classification/classification.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/classification/classification.py -------------------------------------------------------------------------------- /tasks/vision/classification/eval_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/classification/eval_utils.py -------------------------------------------------------------------------------- /tasks/vision/finetune_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/finetune_utils.py -------------------------------------------------------------------------------- /tasks/vision/main.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/main.py -------------------------------------------------------------------------------- /tasks/vision/segmentation/cityscapes.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/segmentation/cityscapes.py -------------------------------------------------------------------------------- /tasks/vision/segmentation/data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/segmentation/data.py -------------------------------------------------------------------------------- /tasks/vision/segmentation/finetune_segformer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/segmentation/finetune_segformer.py -------------------------------------------------------------------------------- /tasks/vision/segmentation/finetune_setr.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/segmentation/finetune_setr.py -------------------------------------------------------------------------------- /tasks/vision/segmentation/metrics.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/segmentation/metrics.py -------------------------------------------------------------------------------- /tasks/vision/segmentation/seg_heads.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/segmentation/seg_heads.py -------------------------------------------------------------------------------- /tasks/vision/segmentation/seg_models.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/segmentation/seg_models.py -------------------------------------------------------------------------------- /tasks/vision/segmentation/transforms.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/segmentation/transforms.py -------------------------------------------------------------------------------- /tasks/vision/segmentation/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/vision/segmentation/utils.py -------------------------------------------------------------------------------- /tasks/zeroshot_gpt/datasets.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/zeroshot_gpt/datasets.py -------------------------------------------------------------------------------- /tasks/zeroshot_gpt/detokenizer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/zeroshot_gpt/detokenizer.py -------------------------------------------------------------------------------- /tasks/zeroshot_gpt/evaluate.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tasks/zeroshot_gpt/evaluate.py -------------------------------------------------------------------------------- /tests/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /tests/functional_tests/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /tests/functional_tests/python_test_utils/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /tests/functional_tests/python_test_utils/check_slurm_job_completion.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/python_test_utils/check_slurm_job_completion.py -------------------------------------------------------------------------------- /tests/functional_tests/python_test_utils/common.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/python_test_utils/common.py -------------------------------------------------------------------------------- /tests/functional_tests/python_test_utils/get_test_results_from_tensorboard_logs.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/python_test_utils/get_test_results_from_tensorboard_logs.py -------------------------------------------------------------------------------- /tests/functional_tests/python_test_utils/jet_test_pipeline.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/python_test_utils/jet_test_pipeline.py -------------------------------------------------------------------------------- /tests/functional_tests/python_test_utils/multitest_ci_pipeline.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/python_test_utils/multitest_ci_pipeline.py -------------------------------------------------------------------------------- /tests/functional_tests/python_test_utils/test_ci_pipeline.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/python_test_utils/test_ci_pipeline.py -------------------------------------------------------------------------------- /tests/functional_tests/python_test_utils/test_resume_checkpoint_pipeline.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/python_test_utils/test_resume_checkpoint_pipeline.py -------------------------------------------------------------------------------- /tests/functional_tests/shell_test_utils/jobwait.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/shell_test_utils/jobwait.sh -------------------------------------------------------------------------------- /tests/functional_tests/shell_test_utils/run_selene_test_launcher_script.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/shell_test_utils/run_selene_test_launcher_script.sh -------------------------------------------------------------------------------- /tests/functional_tests/shell_test_utils/run_selene_test_resume_checkpoint_launcher_script.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/shell_test_utils/run_selene_test_resume_checkpoint_launcher_script.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_results/bert/bert_tp1_pp2_1nodes_50steps.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/bert/bert_tp1_pp2_1nodes_50steps.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/bert/bert_tp1_pp2_1nodes_50steps_core_enabled.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/bert/bert_tp1_pp2_1nodes_50steps_core_enabled.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/bert/bert_tp1_pp2_1nodes_50steps_core_enabled_rope_embeddings.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/bert/bert_tp1_pp2_1nodes_50steps_core_enabled_rope_embeddings.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/bert/bert_tp1_pp2_1nodes_50steps_core_enabled_sequence_parallel.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/bert/bert_tp1_pp2_1nodes_50steps_core_enabled_sequence_parallel.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/bert/bert_tp1_pp4_1nodes_50steps.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/bert/bert_tp1_pp4_1nodes_50steps.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/bert/bert_tp1_pp4_interleaved_1nodes_50steps.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/bert/bert_tp1_pp4_interleaved_1nodes_50steps.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/bert/bert_tp1_pp4_interleaved_1nodes_50steps_core_enabled.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/bert/bert_tp1_pp4_interleaved_1nodes_50steps_core_enabled.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/bert/bert_tp2_pp2_1nodes_50steps.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/bert/bert_tp2_pp2_1nodes_50steps.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/bert/bert_tp2_pp2_1nodes_50steps_core_enabled.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/bert/bert_tp2_pp2_1nodes_50steps_core_enabled.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/bert/bert_tp4_pp1_1nodes_50steps.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/bert/bert_tp4_pp1_1nodes_50steps.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/bert/bert_tp4_pp1_1nodes_50steps_core_enabled.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/bert/bert_tp4_pp1_1nodes_50steps_core_enabled.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp1_1nodes_50steps_dist_optimizer.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp1_1nodes_50steps_dist_optimizer.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp1_1nodes_50steps_dist_optimizer_overlap_grad_reduce.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp1_1nodes_50steps_dist_optimizer_overlap_grad_reduce.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp1_1nodes_50steps_dist_optimizer_overlap_grad_reduce_param_gather.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp1_1nodes_50steps_dist_optimizer_overlap_grad_reduce_param_gather.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp1_1nodes_50steps_overlap_grad_reduce.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp1_1nodes_50steps_overlap_grad_reduce.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp2_1nodes_50steps.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp2_1nodes_50steps.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp2_1nodes_50steps_core_enabled.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp2_1nodes_50steps_core_enabled.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp2_1nodes_50steps_core_enabled_rope_embeddings.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp2_1nodes_50steps_core_enabled_rope_embeddings.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps_core_enabled.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps_core_enabled.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps_core_enabled_disable_bias_linear.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps_core_enabled_disable_bias_linear.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps_core_enabled_sequence_parallel.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps_core_enabled_sequence_parallel.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps_core_enabled_swiglu.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps_core_enabled_swiglu.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps_core_enabled_untie_embeddings_and_outputs.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps_core_enabled_untie_embeddings_and_outputs.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps_overlap_grad_reduce.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_1nodes_50steps_overlap_grad_reduce.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_interleaved_1nodes_50steps.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_interleaved_1nodes_50steps.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_interleaved_1nodes_50steps_core_enabled.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_interleaved_1nodes_50steps_core_enabled.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_interleaved_1nodes_50steps_dist_optimizer_overlap_grad_reduce.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_interleaved_1nodes_50steps_dist_optimizer_overlap_grad_reduce.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_interleaved_1nodes_50steps_dist_optimizer_overlap_grad_reduce_param_gather.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_interleaved_1nodes_50steps_dist_optimizer_overlap_grad_reduce_param_gather.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_interleaved_1nodes_50steps_overlap_grad_reduce.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp1_pp4_interleaved_1nodes_50steps_overlap_grad_reduce.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp2_pp1_1nodes_50steps_core_enabled_context_parallelism_cp2.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp2_pp1_1nodes_50steps_core_enabled_context_parallelism_cp2.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp2_pp1_1nodes_50steps_core_enabled_te_8experts2parallel.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp2_pp1_1nodes_50steps_core_enabled_te_8experts2parallel.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp2_pp1_1nodes_50steps_core_enabled_te_8experts2parallel_groupedGEMM.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp2_pp1_1nodes_50steps_core_enabled_te_8experts2parallel_groupedGEMM.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_4experts.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_4experts.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_core_enabled.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_core_enabled.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_core_enabled_context_parallelism_cp2.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_core_enabled_context_parallelism_cp2.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_core_enabled_te_2experts.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_core_enabled_te_2experts.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_core_enabled_te_4experts2parallel.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_core_enabled_te_4experts2parallel.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_overlap_grad_reduce.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_overlap_grad_reduce.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_te_enabled.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp2_pp2_1nodes_50steps_te_enabled.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp4_pp1_1nodes_50steps.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp4_pp1_1nodes_50steps.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp4_pp1_1nodes_50steps_core_enabled.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp4_pp1_1nodes_50steps_core_enabled.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp4_pp1_1nodes_50steps_dist_optimizer_overlap_grad_reduce.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp4_pp1_1nodes_50steps_dist_optimizer_overlap_grad_reduce.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp4_pp1_1nodes_50steps_dist_optimizer_overlap_grad_reduce_param_gather.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp4_pp1_1nodes_50steps_dist_optimizer_overlap_grad_reduce_param_gather.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/gpt3/gpt3_tp4_pp1_1nodes_50steps_overlap_grad_reduce.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/gpt3/gpt3_tp4_pp1_1nodes_50steps_overlap_grad_reduce.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-1_args--recompute-granularity-full-recompute-method-uniform-recompute-num-layers-1-.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-1_args--recompute-granularity-full-recompute-method-uniform-recompute-num-layers-1-.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-2.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-2.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-2_args--position-embedding-type-rope-.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-2_args--position-embedding-type-rope-.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4_args--disable-bias-linear.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4_args--disable-bias-linear.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4_args--sequence-parallel.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4_args--sequence-parallel.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4_args--swiglu.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4_args--swiglu.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4_args--untie-embeddings-and-output-weights.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-1_pp-4_args--untie-embeddings-and-output-weights.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-2_pp-2.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-2_pp-2.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-4_pp-1.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/jet/gpt3_345m_mcore-pyt_func-train_bf16_nodes-1_gpus-8_bs-32_steps-50_tp-4_pp-1.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/retro/retro_tp1_pp1_1nodes_50steps_core_enabled.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/retro/retro_tp1_pp1_1nodes_50steps_core_enabled.json -------------------------------------------------------------------------------- /tests/functional_tests/test_results/t5/t5_tp1_pp1_interleaved_1nodes_100steps_te_enabled_core_enabled.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_results/t5/t5_tp1_pp1_interleaved_1nodes_100steps_te_enabled_core_enabled.json -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/bert/pretrain_bert_distributed_resume_checkpoint_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/bert/pretrain_bert_distributed_resume_checkpoint_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/bert/pretrain_bert_distributed_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/bert/pretrain_bert_distributed_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/bert/sbatch_bert_distributed_resume_checkpoint_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/bert/sbatch_bert_distributed_resume_checkpoint_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/bert/sbatch_bert_distributed_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/bert/sbatch_bert_distributed_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/gpt3/pretrain_gpt3_distributed_resume_checkpoint_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/gpt3/pretrain_gpt3_distributed_resume_checkpoint_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/gpt3/pretrain_gpt3_distributed_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/gpt3/pretrain_gpt3_distributed_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/gpt3/sbatch_gpt3_distributed_resume_checkpoint_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/gpt3/sbatch_gpt3_distributed_resume_checkpoint_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/gpt3/sbatch_gpt3_distributed_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/gpt3/sbatch_gpt3_distributed_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/retro/pretrain_retro_distributed_resume_checkpoint_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/retro/pretrain_retro_distributed_resume_checkpoint_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/retro/pretrain_retro_distributed_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/retro/pretrain_retro_distributed_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/retro/sbatch_retro_distributed_resume_checkpoint_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/retro/sbatch_retro_distributed_resume_checkpoint_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/retro/sbatch_retro_distributed_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/retro/sbatch_retro_distributed_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/t5/pretrain_t5_distributed_resume_checkpoint_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/t5/pretrain_t5_distributed_resume_checkpoint_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/t5/pretrain_t5_distributed_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/t5/pretrain_t5_distributed_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/t5/sbatch_t5_distributed_resume_checkpoint_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/t5/sbatch_t5_distributed_resume_checkpoint_test.sh -------------------------------------------------------------------------------- /tests/functional_tests/test_scripts/t5/sbatch_t5_distributed_test.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/functional_tests/test_scripts/t5/sbatch_t5_distributed_test.sh -------------------------------------------------------------------------------- /tests/unit_tests/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /tests/unit_tests/data/test_builder.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/data/test_builder.py -------------------------------------------------------------------------------- /tests/unit_tests/data/test_builder_mock_gpt_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/data/test_builder_mock_gpt_dataset.py -------------------------------------------------------------------------------- /tests/unit_tests/data/test_preprocess_data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/data/test_preprocess_data.py -------------------------------------------------------------------------------- /tests/unit_tests/data/test_preprocess_mmdata.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/data/test_preprocess_mmdata.py -------------------------------------------------------------------------------- /tests/unit_tests/dist_checkpointing/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/dist_checkpointing/__init__.py -------------------------------------------------------------------------------- /tests/unit_tests/dist_checkpointing/conftest.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/dist_checkpointing/conftest.py -------------------------------------------------------------------------------- /tests/unit_tests/dist_checkpointing/models/test_gpt_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/dist_checkpointing/models/test_gpt_model.py -------------------------------------------------------------------------------- /tests/unit_tests/dist_checkpointing/models/test_mlp_glu.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/dist_checkpointing/models/test_mlp_glu.py -------------------------------------------------------------------------------- /tests/unit_tests/dist_checkpointing/test_mapping.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/dist_checkpointing/test_mapping.py -------------------------------------------------------------------------------- /tests/unit_tests/dist_checkpointing/test_optimizer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/dist_checkpointing/test_optimizer.py -------------------------------------------------------------------------------- /tests/unit_tests/dist_checkpointing/test_serialization.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/dist_checkpointing/test_serialization.py -------------------------------------------------------------------------------- /tests/unit_tests/models/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /tests/unit_tests/models/test_base_embedding.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/models/test_base_embedding.py -------------------------------------------------------------------------------- /tests/unit_tests/models/test_bert_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/models/test_bert_model.py -------------------------------------------------------------------------------- /tests/unit_tests/models/test_gpt_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/models/test_gpt_model.py -------------------------------------------------------------------------------- /tests/unit_tests/models/test_t5_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/models/test_t5_model.py -------------------------------------------------------------------------------- /tests/unit_tests/pipeline_parallel/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /tests/unit_tests/pipeline_parallel/test_schedules.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/pipeline_parallel/test_schedules.py -------------------------------------------------------------------------------- /tests/unit_tests/tensor_parallel/test_cross_entropy.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/tensor_parallel/test_cross_entropy.py -------------------------------------------------------------------------------- /tests/unit_tests/tensor_parallel/test_data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/tensor_parallel/test_data.py -------------------------------------------------------------------------------- /tests/unit_tests/tensor_parallel/test_mappings.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/tensor_parallel/test_mappings.py -------------------------------------------------------------------------------- /tests/unit_tests/tensor_parallel/test_random.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/tensor_parallel/test_random.py -------------------------------------------------------------------------------- /tests/unit_tests/tensor_parallel/test_tensor_parallel_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/tensor_parallel/test_tensor_parallel_utils.py -------------------------------------------------------------------------------- /tests/unit_tests/test_basic.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/test_basic.py -------------------------------------------------------------------------------- /tests/unit_tests/test_imports.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/test_imports.py -------------------------------------------------------------------------------- /tests/unit_tests/test_parallel_state.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/test_parallel_state.py -------------------------------------------------------------------------------- /tests/unit_tests/test_utilities.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/test_utilities.py -------------------------------------------------------------------------------- /tests/unit_tests/test_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/test_utils.py -------------------------------------------------------------------------------- /tests/unit_tests/transformer/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /tests/unit_tests/transformer/moe/test_grouped_mlp.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/transformer/moe/test_grouped_mlp.py -------------------------------------------------------------------------------- /tests/unit_tests/transformer/moe/test_switch_mlp.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/transformer/moe/test_switch_mlp.py -------------------------------------------------------------------------------- /tests/unit_tests/transformer/test_attention.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/transformer/test_attention.py -------------------------------------------------------------------------------- /tests/unit_tests/transformer/test_attention_packed_seq.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/transformer/test_attention_packed_seq.py -------------------------------------------------------------------------------- /tests/unit_tests/transformer/test_core_attention.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/transformer/test_core_attention.py -------------------------------------------------------------------------------- /tests/unit_tests/transformer/test_mlp.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/transformer/test_mlp.py -------------------------------------------------------------------------------- /tests/unit_tests/transformer/test_module.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/transformer/test_module.py -------------------------------------------------------------------------------- /tests/unit_tests/transformer/test_retro_attention.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/transformer/test_retro_attention.py -------------------------------------------------------------------------------- /tests/unit_tests/transformer/test_spec_customization.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/transformer/test_spec_customization.py -------------------------------------------------------------------------------- /tests/unit_tests/transformer/test_transformer_block.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/transformer/test_transformer_block.py -------------------------------------------------------------------------------- /tests/unit_tests/transformer/test_transformer_layer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tests/unit_tests/transformer/test_transformer_layer.py -------------------------------------------------------------------------------- /tools/autoformat.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/autoformat.sh -------------------------------------------------------------------------------- /tools/bert_embedding/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/bert_embedding/__init__.py -------------------------------------------------------------------------------- /tools/bert_embedding/dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/bert_embedding/dataset.py -------------------------------------------------------------------------------- /tools/bert_embedding/embed.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/bert_embedding/embed.py -------------------------------------------------------------------------------- /tools/bert_embedding/external_libs.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/bert_embedding/external_libs.py -------------------------------------------------------------------------------- /tools/bert_embedding/huggingface.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/bert_embedding/huggingface.py -------------------------------------------------------------------------------- /tools/bert_embedding/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/bert_embedding/utils.py -------------------------------------------------------------------------------- /tools/checkpoint/loader_llama2_hf.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/checkpoint/loader_llama2_hf.py -------------------------------------------------------------------------------- /tools/checkpoint/loader_megatron.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/checkpoint/loader_megatron.py -------------------------------------------------------------------------------- /tools/checkpoint/saver_megatron.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/checkpoint/saver_megatron.py -------------------------------------------------------------------------------- /tools/checkpoint/util.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/checkpoint/util.py -------------------------------------------------------------------------------- /tools/jet_quant_cuda/hadamard/fast_hadamard_transform_cuda.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/hadamard/fast_hadamard_transform_cuda.cu -------------------------------------------------------------------------------- /tools/jet_quant_cuda/hadamard/hadamard_binding.cpp: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/hadamard/hadamard_binding.cpp -------------------------------------------------------------------------------- /tools/jet_quant_cuda/includes/conversion_utils.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/includes/conversion_utils.h -------------------------------------------------------------------------------- /tools/jet_quant_cuda/includes/dequantization_utils.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/includes/dequantization_utils.h -------------------------------------------------------------------------------- /tools/jet_quant_cuda/includes/ds_kernel_utils.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/includes/ds_kernel_utils.h -------------------------------------------------------------------------------- /tools/jet_quant_cuda/includes/fast_hadamard_transform.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/includes/fast_hadamard_transform.h -------------------------------------------------------------------------------- /tools/jet_quant_cuda/includes/fast_hadamard_transform_common.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/includes/fast_hadamard_transform_common.h -------------------------------------------------------------------------------- /tools/jet_quant_cuda/includes/fast_hadamard_transform_special.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/includes/fast_hadamard_transform_special.h -------------------------------------------------------------------------------- /tools/jet_quant_cuda/includes/hadamard_binding.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/includes/hadamard_binding.h -------------------------------------------------------------------------------- /tools/jet_quant_cuda/includes/memory_access_utils.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/includes/memory_access_utils.h -------------------------------------------------------------------------------- /tools/jet_quant_cuda/includes/quantization.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/includes/quantization.h -------------------------------------------------------------------------------- /tools/jet_quant_cuda/includes/quantization_utils.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/includes/quantization_utils.h -------------------------------------------------------------------------------- /tools/jet_quant_cuda/includes/reduction_utils.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/includes/reduction_utils.h -------------------------------------------------------------------------------- /tools/jet_quant_cuda/includes/static_switch.h: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/includes/static_switch.h -------------------------------------------------------------------------------- /tools/jet_quant_cuda/quantization/dequant_reduce_quant.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/quantization/dequant_reduce_quant.cu -------------------------------------------------------------------------------- /tools/jet_quant_cuda/quantization/dequantize.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/quantization/dequantize.cu -------------------------------------------------------------------------------- /tools/jet_quant_cuda/quantization/dequantize_add.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/quantization/dequantize_add.cu -------------------------------------------------------------------------------- /tools/jet_quant_cuda/quantization/pt_binding.cpp: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/quantization/pt_binding.cpp -------------------------------------------------------------------------------- /tools/jet_quant_cuda/quantization/quant_reduce.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/quantization/quant_reduce.cu -------------------------------------------------------------------------------- /tools/jet_quant_cuda/quantization/quant_reduce_ht.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/quantization/quant_reduce_ht.cu -------------------------------------------------------------------------------- /tools/jet_quant_cuda/quantization/stochastic_quantize.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/quantization/stochastic_quantize.cu -------------------------------------------------------------------------------- /tools/jet_quant_cuda/quantization/stochastic_quantize_ht.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/quantization/stochastic_quantize_ht.cu -------------------------------------------------------------------------------- /tools/jet_quant_cuda/quantization/sub_quantize.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/quantization/sub_quantize.cu -------------------------------------------------------------------------------- /tools/jet_quant_cuda/quantization/swizzled_quantize.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/quantization/swizzled_quantize.cu -------------------------------------------------------------------------------- /tools/jet_quant_cuda/quantization/swizzled_quantize_ht.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/quantization/swizzled_quantize_ht.cu -------------------------------------------------------------------------------- /tools/jet_quant_cuda/setup.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/setup.py -------------------------------------------------------------------------------- /tools/jet_quant_cuda/unit_test/2hop_quant_test.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/unit_test/2hop_quant_test.py -------------------------------------------------------------------------------- /tools/jet_quant_cuda/unit_test/dequant_function.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/unit_test/dequant_function.py -------------------------------------------------------------------------------- /tools/jet_quant_cuda/unit_test/fused_dequantize_add.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/unit_test/fused_dequantize_add.ipynb -------------------------------------------------------------------------------- /tools/jet_quant_cuda/unit_test/fused_sto_test.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/unit_test/fused_sto_test.ipynb -------------------------------------------------------------------------------- /tools/jet_quant_cuda/unit_test/sto_quant_test.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/unit_test/sto_quant_test.py -------------------------------------------------------------------------------- /tools/jet_quant_cuda/unit_test/tool_function.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/jet_quant_cuda/unit_test/tool_function.py -------------------------------------------------------------------------------- /tools/linter.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/linter.py -------------------------------------------------------------------------------- /tools/merge_datasets.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/merge_datasets.py -------------------------------------------------------------------------------- /tools/openwebtext/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/openwebtext/README.md -------------------------------------------------------------------------------- /tools/openwebtext/add_id.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/openwebtext/add_id.py -------------------------------------------------------------------------------- /tools/openwebtext/blacklist_urls.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/openwebtext/blacklist_urls.py -------------------------------------------------------------------------------- /tools/openwebtext/cleanup_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/openwebtext/cleanup_dataset.py -------------------------------------------------------------------------------- /tools/openwebtext/cleanup_fix_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/openwebtext/cleanup_fix_dataset.py -------------------------------------------------------------------------------- /tools/openwebtext/filter_ngrams.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/openwebtext/filter_ngrams.py -------------------------------------------------------------------------------- /tools/openwebtext/find_duplicates.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/openwebtext/find_duplicates.py -------------------------------------------------------------------------------- /tools/openwebtext/group_duplicate_url.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/openwebtext/group_duplicate_url.py -------------------------------------------------------------------------------- /tools/openwebtext/merge_jsons.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/openwebtext/merge_jsons.py -------------------------------------------------------------------------------- /tools/openwebtext/remove_group_duplicates.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/openwebtext/remove_group_duplicates.py -------------------------------------------------------------------------------- /tools/preprocess_data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/preprocess_data.py -------------------------------------------------------------------------------- /tools/preprocess_data_nmt.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/preprocess_data_nmt.py -------------------------------------------------------------------------------- /tools/preprocess_mmdata.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/preprocess_mmdata.py -------------------------------------------------------------------------------- /tools/retro/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/README.md -------------------------------------------------------------------------------- /tools/retro/build_db.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/build_db.md -------------------------------------------------------------------------------- /tools/retro/cli/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/cli/__init__.py -------------------------------------------------------------------------------- /tools/retro/cli/__main__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/cli/__main__.py -------------------------------------------------------------------------------- /tools/retro/cli/cli.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/cli/cli.py -------------------------------------------------------------------------------- /tools/retro/db/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/db/__init__.py -------------------------------------------------------------------------------- /tools/retro/db/build.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/db/build.py -------------------------------------------------------------------------------- /tools/retro/db/dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/db/dataset.py -------------------------------------------------------------------------------- /tools/retro/db/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/db/utils.py -------------------------------------------------------------------------------- /tools/retro/examples/Dockerfile: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/examples/Dockerfile -------------------------------------------------------------------------------- /tools/retro/examples/preprocess_data.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/examples/preprocess_data.sh -------------------------------------------------------------------------------- /tools/retro/examples/pretrain_model.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/examples/pretrain_model.sh -------------------------------------------------------------------------------- /tools/retro/external_libs.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/external_libs.py -------------------------------------------------------------------------------- /tools/retro/index/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/index/__init__.py -------------------------------------------------------------------------------- /tools/retro/index/build.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/index/build.py -------------------------------------------------------------------------------- /tools/retro/index/factory.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/index/factory.py -------------------------------------------------------------------------------- /tools/retro/index/index.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/index/index.py -------------------------------------------------------------------------------- /tools/retro/index/indexes/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/index/indexes/__init__.py -------------------------------------------------------------------------------- /tools/retro/index/indexes/faiss_base.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/index/indexes/faiss_base.py -------------------------------------------------------------------------------- /tools/retro/index/indexes/faiss_par_add.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/index/indexes/faiss_par_add.py -------------------------------------------------------------------------------- /tools/retro/index/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/index/utils.py -------------------------------------------------------------------------------- /tools/retro/main.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/main.py -------------------------------------------------------------------------------- /tools/retro/query/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/query/__init__.py -------------------------------------------------------------------------------- /tools/retro/query/chunk_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/query/chunk_dataset.py -------------------------------------------------------------------------------- /tools/retro/query/multi_split_gpt_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/query/multi_split_gpt_dataset.py -------------------------------------------------------------------------------- /tools/retro/query/query.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/query/query.py -------------------------------------------------------------------------------- /tools/retro/query/retro_dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/query/retro_dataset.py -------------------------------------------------------------------------------- /tools/retro/query/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/query/utils.py -------------------------------------------------------------------------------- /tools/retro/sft/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/sft/README.md -------------------------------------------------------------------------------- /tools/retro/sft/dataset_conv.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/sft/dataset_conv.py -------------------------------------------------------------------------------- /tools/retro/sft/open_inst.sh: -------------------------------------------------------------------------------- 1 | DATA_BLEND="1.0 open_inst" 2 | -------------------------------------------------------------------------------- /tools/retro/sft/sft_retro.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/sft/sft_retro.py -------------------------------------------------------------------------------- /tools/retro/sft/sft_retro_lm.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/sft/sft_retro_lm.sh -------------------------------------------------------------------------------- /tools/retro/text_generation/evaluate.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/text_generation/evaluate.py -------------------------------------------------------------------------------- /tools/retro/text_generation/metrics.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/text_generation/metrics.py -------------------------------------------------------------------------------- /tools/retro/text_generation/retro_api.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/text_generation/retro_api.py -------------------------------------------------------------------------------- /tools/retro/text_generation/retro_generate.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/text_generation/retro_generate.sh -------------------------------------------------------------------------------- /tools/retro/text_generation/retro_generation.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/text_generation/retro_generation.py -------------------------------------------------------------------------------- /tools/retro/text_generation/retro_text_generation.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/text_generation/retro_text_generation.py -------------------------------------------------------------------------------- /tools/retro/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/retro/utils.py -------------------------------------------------------------------------------- /tools/run_text_generation_server.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/run_text_generation_server.py -------------------------------------------------------------------------------- /tools/text_generation_cli.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/SDP4Bit/HEAD/tools/text_generation_cli.py --------------------------------------------------------------------------------