mosaicml/llm-foundry

Pull Requests Commits

ignore ConsecutiveRepeatedChatRolesError

eitanturok committed 1 year ago

Verified 790ac9fd

change dataset path

eitanturok committed 1 year ago

Verified 3f82ff8f

.json is a safe extension

eitanturok committed 1 year ago

Verified a3e2da2b

register lilacai glaive fc dataset

eitanturok committed 1 year ago

Verified 19e9f51a

Merge branch 'main' into sharegpt-format

samhavens committed 1 year ago

Verified 2ecfcbaa

fixing evaluator microbatch size (#1100)

ShashankMosaicML committed 1 year ago

Verified d0d94342

Norms registry (#1080)

dakinggg committed 1 year ago

Verified b81897ac

fix type signatures and add hints

samhavens committed 1 year ago

1e0b3dcd

dakinggg committed 1 year ago

Verified 60a1ab4c

guess at linting

samhavens committed 1 year ago

f06b9c6b

support ShareGPT format

samhavens committed 1 year ago

a4a654e9

cleaned up HF/MPT conversion test (#1048)

milocress committed 1 year ago

Verified 94301cd1

Allow overrides for nested PretrainedConfig (#1089)

dakinggg committed 1 year ago

Verified 96b27c59

fix yamls typo (#1092)

dakinggg committed 1 year ago

Verified f18768b4

Bump transformers to 4.39.3 (#1086)

dakinggg committed 1 year ago

Verified 607b982d

fix typo in expandable_segments (#1088)

mammothb committed 1 year ago

Verified 06ff95f6

Bump HF version (#1091)

b-chu committed 1 year ago

Verified 26f0619d

log packing ratio progress (#1070)

milocress committed 1 year ago

Verified e70e4241

decrease mlflow hf model shard size (#1087)

dakinggg committed 1 year ago

Verified c766cf95

Update README.md to include DBRX blog under "Latest News" (#1085)

lupesko committed 1 year ago

Verified 5455b404

Background mlflow model registration (#1078)

irenedea committed 1 year ago

Verified 23c31734

Fixes for streaming and auto packing (#1083)

dakinggg committed 1 year ago

Verified d452c60d

Upgrade hf chat (#1061)

j316chuck committed 1 year ago

Verified 394735b1

mvpatel2000 committed 1 year ago

Verified 580a4b00

Doing the loss reduction in foundry instead of in the loss functions. (#1079)

ShashankMosaicML committed 1 year ago

Verified 632cb739

Fix tiktoken vocab size (#1081)

dakinggg committed 1 year ago

Verified caf7fda9

Triton RMSNorm (#1050)

josejg committed 1 year ago

Verified b765b470

Check the user provided eos / bos token id against the tokenizer eos / bos token id (#1039)

ShashankMosaicML committed 1 year ago

Verified d8ea2c56

Add expandeable segments flag (#1075)

dakinggg committed 1 year ago

Verified 349c2ff0

Output eval logging batch (#961)

maxisawesome committed 1 year ago

Verified 7a8a1564

Older