microsoft/DeepSpeed

Pull Requests Commits

Remove GatheredParameters context from replace_with_policy (#2591)

lekurile committed 2 years ago

Verified 503706ac

call empty_cache to really free up GPU memory as described in comment (#2620)

guoyejun committed 2 years ago

Verified d0dbc95a

skip torch.zeros and tensor.copy_ when model parallel is not used (#2479)

guoyejun committed 2 years ago

Verified b08cf416

Use rocm/pytorch:latest (#2613)

jithunnair-amd committed 2 years ago

Verified 7e2103f8

add fix for older pydantic versions (#2611)

mrwyattii committed 2 years ago

Verified 2e596e67

[deepspeed/autotuner] Bug fix for binary search for batch size (#2162)

rahilbathwal5 committed 2 years ago

Verified 384f17b0

Move layer norm to new schedule (#2590)

lokoppakmsft committed 2 years ago

Verified 3a3dfe66

Migrate ops tests to new inference_ops marker (#2599)

cmikeh2 committed 2 years ago

Verified 7425a365

fix blog link (#2600)

conglongli committed 2 years ago

Verified acde873c

DeepSpeed Data Efficiency Library (#2585)

conglongli committed 2 years ago

Verified ef869377

jeffra committed 2 years ago

2600db54

get mask token from tokenizer (#2592)

jeffra committed 3 years ago

Verified 2076bf23

Fix issues w. python 3.6 + add py-version checks to CI (#2589)

jeffra committed 3 years ago

Verified 35eabb0a

Updating API docs (#2586)

jomayeri committed 3 years ago

Verified 18713c68

Updating docs README (#2587)

jomayeri committed 3 years ago

Verified 377c770a

Add checkpoint sharding unit tests (#2561)

mrwyattii committed 3 years ago

Verified ccb8eb81

Support N-dimension input in quantization kernel (#2575)

lokoppakmsft committed 3 years ago

Verified 591744eb

Update barrier and reduce_scatter_base to conform to PyTorch signatures (#2570)

Quentin-Anthony committed 3 years ago

Verified 18d55e54

Fix MegatronLayerPolicy to have megatron_v2=True (#2579)

lekurile committed 3 years ago

Verified 731965db

Fix quantized-inference & Add generic support of checkpoint loading (#2547)

RezaYazdaniAminabadi committed 3 years ago

Verified 35b350b2

Drop Maxwell Support (#2574)

cmikeh2 committed 3 years ago

Verified b8416282

Support fp32 gradaccum for bf16 model (#2566)

delock committed 3 years ago

Verified 06938835

Add Determined to open-source DL frameworks (#2573)

sirredbeard committed 3 years ago

Verified 2d8f3f56

use get_global_rank if available (#2567)

jeffra committed 3 years ago

Verified d4cab2ce

docs: Update the recent url for Megatron-LM (#2564)

rapsealk committed 3 years ago

Verified bbe030c5

jeffra committed 3 years ago

c77d42dc

Fix invalid check of recorded parameter orders in zero stage3. (#2550)

inkcherry committed 3 years ago

Verified aeda7f9f

Abstract accelerator (step 1) (#2504)

delock committed 3 years ago

Verified ffcf3846

add missing moe deprecated fields to inference config (#2556)

mrwyattii committed 3 years ago

Verified c5f85858

encoded ds config into command line argument when launching child processes in autotuning (#2524)

cli99 committed 3 years ago

Verified abe4fc6b

Newer Older