microsoft/DeepSpeed

Pull Requests Commits

fix table syntax (#2204)

kamalkraj committed 3 years ago

Verified 87b20133

Refactor Distributed Tests (#2180)

mrwyattii committed 3 years ago

Verified ac995198

Fix the tensor-slicing copy for qkv parameters (#2198)

RezaYazdaniAminabadi committed 3 years ago

Verified 8920308c

Log user config exactly (#2201)

tjruwase committed 3 years ago

Verified 28dfca8a

Adding additional instructiosn in the compression tutorial on pre-training distillation and quantization for GPT (#2197)

minjiaz committed 3 years ago

Verified f82846d7

Add gradient_average flag support for sparse grads (#2188)

Dipet committed 3 years ago

Verified b62e0cc5

Fix the layer-past for GPT based models (#2196)

RezaYazdaniAminabadi committed 3 years ago

Verified 54a9e1b9

Update README.md (#2192)

gasparitiago committed 3 years ago

Verified 0f5c2012

Update nv-lightning-v100.yml (#2190)

jeffra committed 3 years ago

Verified 1223b13c

[docs] add more models to adoption (#2189)

jeffra committed 3 years ago

Verified 33667e0e

fix missing import (#2175)

rahilbathwal5 committed 3 years ago

Verified ee5ce524

Update README to latest Composer version (#2177)

hanlint committed 3 years ago

Verified 80b5b925

Enable fused_lamb_cuda_kernel on ROCm (#2148)

rraminen committed 3 years ago

Verified 2e3769a1

Match compute and reduce dtype (#2145)

tjruwase committed 3 years ago

Verified e419f7cb

fixing model partitioning without injection (#2179)

RezaYazdaniAminabadi committed 3 years ago

Verified e7d99595

update offload docs to include stage 1 (#2178)

jeffra committed 3 years ago

Verified fad0a410

Update for AMD CI workflow (#2172)

mrwyattii committed 3 years ago

Verified d1cd18e5

[docs] adoption updates (#2173)

jeffra committed 3 years ago

Verified bb49dc73

Fix wrong unit of latency in flops-profiler (#2090) (#2095)

zionwu committed 3 years ago

Verified 6bfcf3c6

delay torch import for inference compatability check (#2167)

jeffra committed 3 years ago

Verified 776e3698

Fix for distributed tests on pytorch>=1.12 (#2141)

mrwyattii committed 3 years ago

Verified 1a71e77d

jeffra committed 3 years ago

b005db86

Tensor parallelism for Mixture of Experts (#2074)

siddharth9820 committed 3 years ago

Verified 5fe9d610

Release swap buffers for persisted params (#2089)

tjruwase committed 3 years ago

Verified 2210ebe7

enable fp16 input autocasting (#2158)

jeffra committed 3 years ago

Verified a039e226

[zero-3] shutdown zero.Init from within ds.init (#2150)

jeffra committed 3 years ago

Verified 46401b38

prevent cuda 10 builds of inference kernels on ampere (#2157)

jeffra committed 3 years ago

Verified 63f470ee

Elastic Training support in DeepSpeed (#2153) (#2156)

aj-prime committed 3 years ago

Verified 1ed5aa96

Added retain_graph as a kwarg to the main engine backward function (#1149)

ncilfone committed 3 years ago

Verified ba67bd9a

Fix random token-generation issue + MP-checkpoint loading/saving (#2132)

RezaYazdaniAminabadi committed 3 years ago

Verified 556f0051

Newer Older