abstract_accelerator.py
Commit
2 years ago
Author
Reza Yazdani
- Files81
- accelerator
- cuda_accelerator.py
- csrc
- aio/py_test
- aio_bench_perf_sweep.py
- ds_aio_basic.py
- ds_aio_handle.py
- validate_async_io.py
- transformer
softmax_kernels.cu
- deepspeed
- autotuning
- autotuner.py
- inference
- engine.py
- launcher
- multinode_runner.py
- runner.py
- model_implementations/transformers
- clip_encoder.py
- ds_transformer.py
- module_inject
- auto_tp.py
- containers
- vae.py
- layers.py
- load_checkpoint.py
- replace_module.py
- monitor
- config.py
- constants.py
- csv_monitor.py
- monitor.py
- tensorboard.py
- wandb.py
- ops
- adagrad
- cpu_adagrad.py
- adam
- cpu_adam.py
- fused_adam.py
- lamb
- fused_lamb.py
- quantizer
- quantizer.py
- random_ltd
- dropping_utils.py
- transformer
- inference
- bias_add.py
- diffusers_attention.py
- diffusers_transformer_block.py
- moe_inference.py
- op_binding
- base.py
- transformer.py
- profiling/flops_profiler
- profiler.py
- runtime
- activation_checkpointing
- checkpointing.py
- comm
- nccl.py
- config.py
- data_pipeline/data_sampling
- data_sampler.py
- dataloader.py
- engine.py
- fp16
- fused_optimizer.py
- onebit
- adam.py
- lamb.py
- zoadam.py
- unfused_optimizer.py
- pipe
- engine.py
- module.py
- p2p.py
- swap_tensor
- partitioned_optimizer_swapper.py
- partitioned_param_swapper.py
- pipelined_optimizer_swapper.py
- utils.py
- utils.py
- weight_quantizer.py
- zero
- linear.py
- parameter_offload.py
- partition_parameters.py
- partitioned_param_coordinator.py
- stage3.py
- stage_1_and_2.py
- utils.py
- utils
- nvtx.py
- timer.py
- docs
- _tutorials
autotuning.md- azure.md
- code-docs/source
- index.rst
- monitor.rst
- op_builder
- __init__.py
- all_ops.py
- builder.py
- builder_names.py
- requirements
requirements.txt
- tests/unit
- inference
- test_inference.py
- monitor
- test_monitor.py
- ops/transformer/inference
- test_bias_gelu.py
- runtime
- test_ds_initialize.py
- zero
- test_zero.py
- test_zero_context.py