mosaicml/llm-foundry

Pull Requests Commits

mattyding committed 1 year ago

0ae048cb

mattyding committed 1 year ago

ee8e274c

mattyding committed 1 year ago

5120abdc

i love debugging

mattyding committed 1 year ago

dda4ec7a

mattyding committed 1 year ago

084a2e48

mattyding committed 1 year ago

01c953b7

print out dtype for debugging

mattyding committed 1 year ago

2d7dbde1

mattyding committed 1 year ago

d87886ec

first try specifying bytes type

mattyding committed 1 year ago

d10e89e8

mattyding committed 1 year ago

f47093f8

convert ndarray to bytes

mattyding committed 1 year ago

512a165c

mattyding committed 1 year ago

92c58861

mattyding committed 1 year ago

f8640ea0

mattyding committed 1 year ago

38652da7

Merge branch 'main' into matt/split-mds-script

mattyding committed 1 year ago

5b918c42

mattyding committed 1 year ago

654d084f

Change RMSNorm to use PyTorch native implementation (#1809)

josejg committed 1 year ago

Verified 7993aebe

Allow subselecting the appropriate config for llama4 (#1815)

dakinggg committed 1 year ago

Verified 5c47350d

Upgrade yapf version (#1814)

dakinggg committed 1 year ago

Verified 502bf051

Update foundry version in MCLI yamls (#1813)

dakinggg committed 1 year ago

Verified ea773086

Adding temperature tuning in attention (#1793)

ShashankMosaicML committed 1 year ago

Verified 67875360

Bump foundry version to 0.21.0.dev0 (#1812)

dakinggg committed 1 year ago

Verified faa7f913

Adding support for nope positional encoding in block overrides. (#1794)

ShashankMosaicML committed 1 year ago

Verified 1a8776cf

Remove some old deprecated code/comments (#1811)

dakinggg committed 1 year ago

Verified 86741ead

Update grouped-gemm version (#1810)

dakinggg committed 1 year ago

Verified f91d8f69

Update datasets requirement from <3.4,>=3.3.2 to >=3.3.2,<3.6 (#1807)

dependabot[bot] committed 1 year ago

Verified 77f89674

Update setuptools requirement from <80.0.0 to <81.0.0 (#1803)

dependabot[bot] committed 1 year ago

Verified 71cd3a6b

added attn_implementation for hf_base.py (#1801)

ethantang-db committed 1 year ago

Verified 272dbd6c

Load model only on global rank 0 for mixed init (#1795)

dakinggg committed 1 year ago

Verified 0f6d9ba6

extended hf_checkpointer for any additional content saving (#1792)

ethantang-db committed 1 year ago

Verified ec9de523

Older