| .. |
|
common.py
|
Dist testing backend fixes, etc. (#708)
|
2021-01-29 13:08:37 -08:00 |
|
ds_batch_config.json
|
Support fp32 grad clipping and fix max_grad_norm confusion (#232)
|
2020-05-26 19:40:01 -07:00 |
|
gpt2-merges.txt
|
Quantization + inference release (#1091)
|
2021-05-24 01:10:39 -07:00 |
|
gpt2-vocab.json
|
Quantization + inference release (#1091)
|
2021-05-24 01:10:39 -07:00 |
|
megatron_model.py
|
Quantization + inference release (#1091)
|
2021-05-24 01:10:39 -07:00 |
|
modeling.py
|
Quantization + inference release (#1091)
|
2021-05-24 01:10:39 -07:00 |
|
modelingpreln.py
|
Quantization + inference release (#1091)
|
2021-05-24 01:10:39 -07:00 |
|
multi_output_model.py
|
Support multi-output models (#170)
|
2020-03-27 13:57:29 -07:00 |
|
simple_model.py
|
Add find_unused_parameters option to DeepSpeedEngine (#945)
|
2021-04-25 04:45:27 -07:00 |
|
test_activation_checkpointing.py
|
Activation checkpointing for non-tensor arguments and return values (#741)
|
2021-02-12 14:01:52 -08:00 |
|
test_adamw.py
|
set adamw_mode default true (follows FusedAdam and < 0.3.11 logic) (#844)
|
2021-03-10 18:02:08 -08:00 |
|
test_aio.py
|
Add nvme unit/perf tests (#993)
|
2021-04-22 10:02:30 -07:00 |
|
test_autocast.py
|
Use amp autocast in ZeRO3 linear (#990)
|
2021-04-23 10:53:55 -07:00 |
|
test_checkpointing.py
|
Quantization + inference release (#1091)
|
2021-05-24 01:10:39 -07:00 |
|
test_config.py
|
Refactor param_dict to config (#1008)
|
2021-04-28 17:05:03 -07:00 |
|
test_configurable_parallel.py
|
Quantization + inference release (#1091)
|
2021-05-24 01:10:39 -07:00 |
|
test_cpu_adam.py
|
DeepSpeed JIT op + PyPI support (#496)
|
2020-11-12 11:51:38 -08:00 |
|
test_csr.py
|
Sparse attn + ops/runtime refactor + v0.3.0 (#343)
|
2020-09-01 18:06:15 -07:00 |
|
test_cuda_backward.py
|
Quantization + inference release (#1091)
|
2021-05-24 01:10:39 -07:00 |
|
test_cuda_forward.py
|
Quantization + inference release (#1091)
|
2021-05-24 01:10:39 -07:00 |
|
test_data.py
|
Pipeline parallel training engine. (#392)
|
2020-09-09 23:14:55 -07:00 |
|
test_dist.py
|
Enables NCCL backend in @distributed_test (#13)
|
2020-02-05 11:57:02 -08:00 |
|
test_ds_arguments.py
|
Improve doc string for add_XXX_arguments (#32)
|
2020-02-06 13:14:22 -08:00 |
|
test_ds_config.py
|
Sparse attn + ops/runtime refactor + v0.3.0 (#343)
|
2020-09-01 18:06:15 -07:00 |
|
test_dynamic_loss_scale.py
|
Dist testing backend fixes, etc. (#708)
|
2021-01-29 13:08:37 -08:00 |
|
test_elastic.py
|
add additional validation checks in elastic config (#646)
|
2021-01-08 11:02:33 -08:00 |
|
test_flops_profiler.py
|
Add flops profiler tutorial (#682)
|
2021-02-10 18:03:55 -08:00 |
|
test_fp16.py
|
Quantization + inference release (#1091)
|
2021-05-24 01:10:39 -07:00 |
|
test_ignore_unused_parameters.py
|
Avoid unused parameters assert by default (#1039)
|
2021-05-07 11:15:09 -07:00 |
|
test_lr_schedulers.py
|
LR scheduler unit tests (#429)
|
2021-01-08 15:32:05 -08:00 |
|
test_multi_output_model.py
|
ZeRO-2 (#217)
|
2020-05-19 01:00:53 -07:00 |
|
test_onebit.py
|
1-bit LAMB optimizer (#970)
|
2021-04-20 18:28:22 -07:00 |
|
test_partition.py
|
Pipeline parallel training engine. (#392)
|
2020-09-09 23:14:55 -07:00 |
|
test_pipe.py
|
disable pipe test (#915)
|
2021-04-02 13:20:21 -07:00 |
|
test_pipe_module.py
|
ZeRO-Infinity (#976)
|
2021-04-18 23:45:37 -07:00 |
|
test_pipe_schedule.py
|
Pipeline parallel training engine. (#392)
|
2020-09-09 23:14:55 -07:00 |
|
test_pld.py
|
PLD release (#513)
|
2020-11-10 12:53:50 -08:00 |
|
test_run.py
|
Sparse attn + ops/runtime refactor + v0.3.0 (#343)
|
2020-09-01 18:06:15 -07:00 |
|
test_runtime_utils.py
|
Pipeline parallel training engine. (#392)
|
2020-09-09 23:14:55 -07:00 |
|
test_sparse_attention.py
|
Change the sparse attention API to be compatible with latest changes of triton (#902)
|
2021-06-02 12:42:53 -07:00 |
|
test_topology.py
|
Pipeline parallel training engine. (#392)
|
2020-09-09 23:14:55 -07:00 |
|
test_zero.py
|
Fix unbalanced gradients bug in ZeRO-2 gradient accumulation (#545)
|
2020-11-20 15:39:01 -08:00 |
|
test_zero_context.py
|
[Stage][Fix] Add additional conditions when checking types of output from the model (#1026)
|
2021-05-01 08:46:46 -07:00 |
|
test_zero_tiled.py
|
ZeRO-Infinity (#976)
|
2021-04-18 23:45:37 -07:00 |
8def3cb3a2