microsoft/DeepSpeed

Pull Requests Commits

Fix path in workflow sed

lekurile committed 1 year ago

f3a2813e

Merge branch 'master' into lekurile/infv2_lm_eval

lekurile committed 1 year ago

Verified 9a44a860

lekurile committed 1 year ago

5567b90f

Fix formatting and update sed command in GH workflow

lekurile committed 1 year ago

15606059

Use get_accelerator().empty_cache() to clear GPU cache

lekurile committed 1 year ago

33ba898d

Update human eval unit test

lekurile committed 1 year ago

289d9010

Fix typos in blogs/ (#5172)

jinyouzhi committed 1 year ago

Verified 8d0150d9

Write multiple items to output file at once, in distributed data analyzer. (#5169)

bm-synth committed 1 year ago

Verified d5fa87ff

Merge branch 'master' into lekurile/infv2_lm_eval

lekurile committed 1 year ago

Verified a7e41053

Fix allreduce for BF16 and ZeRO0 (#5170)

tohtana committed 1 year ago

Verified dd3690c5

Update flops profiler to handle attn and __matmul__ (#4724)

KimmiShi committed 1 year ago

Verified 98c96e79

Use ninja to speed up build (#5088)

jinzhen-lin committed 1 year ago

Verified b00533e4

Fix gradient clipping (#5150)

tohtana committed 1 year ago

Verified 005afe12

MOE: Fix save checkpoint when TP > 1 (#5157)

mosheisland committed 1 year ago

Verified a84d07c5

DeepSpeedZeroOptimizer_Stage3: remove cuda specific optimizer (#5138)

nelyahu committed 1 year ago

Verified 7f0950f8

Distributed in-memory map-reduce for data analyzer (#5129)

bm-synth committed 1 year ago

Verified e977c7d6

get_grad_norm_direct: fix a case of empty norm group (#5148)

nelyahu committed 1 year ago

Verified f062a1b2

Pin to PyTest 8.0.0 (#5163)

loadams committed 1 year ago

Verified fa3662f7

Update deprecated HuggingFace function (#5144)

mrwyattii committed 1 year ago

Verified a37e59b5

Fix UserWarning: The torch.cuda.*DtypeTensor constructors are no long… (#5018)

ShukantPal committed 1 year ago

Verified 177dc143

TestEmptyParameterGroup: replace fusedAdam with torch.optim.AdamW (#5139)

nelyahu committed 1 year ago

Verified a3be0d44

Added missing `isinstance` check in PR 5112 (#5142)

bm-synth committed 1 year ago

Verified 9bd62e0b

Method `run_map_reduce` to fix errors when running `run_map` followed by `run_reduce` (#5131)

bm-synth committed 1 year ago

Verified 95be499a

Update nv-inference torch version (#5128)

loadams committed 1 year ago

Verified 39f6ee59

Check overflow on device without host synchronization for each tensor (#5115)

BacharL committed 1 year ago

Verified 1c9e5ef2

tensorboard logging: avoid item() outside gas to improve performance (#5135)

nelyahu committed 1 year ago

Verified fac1df99

remove mandatory `index` key from output of `metric_function` in `DataAnalysis` map operation (#5112)

bm-synth committed 1 year ago

Verified 2d0a6bc2

Avoid zero-sized microbatches for incomplete minibatches when doing curriculum learning (#5118)

bm-synth committed 1 year ago

Verified a7864846

Fix broadcast deadlock for incomplete batches in data sample for data analysis (#5117)

bm-synth committed 1 year ago

Verified 2b411103

Solve tensor vs numpy dtype conflicts in data efficiency map-reduce. (#5108)

bm-synth committed 1 year ago

Verified 3e5d4004

Older