mosaicml/llm-foundry

Pull Requests Commits

Update text_data.py

jfrankle committed 2 years ago

Verified ad8ba4bd

Update text_data.py

jfrankle committed 2 years ago

Verified 1c0428d9

Fix training integration test (#517)

j316chuck committed 2 years ago

Verified 6d582877

Grouped Query Attention + Refactor Attn (#492)

sashaDoubov committed 2 years ago

Verified d2fbc3b4

Updates GPU test timeout to use mcloud flag (#510)

mvpatel2000 committed 2 years ago

Verified 7ac554da

remove logging (#509)

mvpatel2000 committed 2 years ago

Verified 3072abb3

Add MLFlow as a logger option (#475)

aspfohl committed 2 years ago

Verified ad6545c1

Updated StreamingTextDataloader to pass take in shuffle_block_size (#503)

snarayan21 committed 2 years ago

Verified 358e064f

fix no bias assignment (#502)

vchiley committed 2 years ago

Verified b4e27c0d

Adding pyright to pre-commit (#477)

bcui19 committed 2 years ago

Verified 9250e846

fix boolean for reentrant setting (#500)

dakinggg committed 2 years ago

Verified 05c6055e

Increase lint ci timeout (#498)

dakinggg committed 2 years ago

Verified 2445274a

Fix Local World Size (#495)

rishab-partha committed 2 years ago

Verified bd812725

If Alibi is on, we should turn learned_pos_emb to False (#489)

bcui19 committed 2 years ago

Verified 634597d8

change accelerate (#485)

rishab-partha committed 2 years ago

Verified 776cf0ea

adds early stopping call back (#488)

Cody Blakeney committed 2 years ago

Verified 80286387

Fix ambiguous `throughput` in README (#476)

abhi-mosaic committed 2 years ago

Verified c6cfb428

Support eval loader when finetuning from JSONL files in object stores (#469)

samhavens committed 2 years ago

Verified 66b84bb4

Autoresume default on (#467)

mvpatel2000 committed 2 years ago

Verified ef8d4143

Fix missing import (#470)

bcui19 committed 2 years ago

Verified ca0bf0fc

Adding different device intialization to eval (#466)

bcui19 committed 2 years ago

Verified bb0ad428

Update inference benchmark with HF changes (#461)

sashaDoubov committed 2 years ago

Verified 3e292334

Update mpt-30b-instruct.yaml (#448)

samhavens committed 2 years ago

Verified 3b0b678a

Add linear layer and ffn config to enable TransformerEngine layers (with FP8) (#432)

vchiley committed 2 years ago

Verified 340a5665

setup.py: replace composer with mosaicml (#458)

guoyejun committed 2 years ago

Verified 275fa56d

Add trust remote code for tokenizer in inference conversion script (#446)

margaretqian committed 2 years ago

Verified 496b50bd

Update mcli-hf-generate.yaml (#456)

vchiley committed 2 years ago

Verified 2f49eb7f

Use torch.repeat instead of expand on key & value in Triton MQA to prevent NaNs with certain h_dims (#442)

sashaDoubov committed 2 years ago

Verified 86a99e2a

Remove try catch in eval.py; make model_gauntlet optional in eval.py (#434)

bmosaicml committed 2 years ago

Verified 62e2feac

Update README.md (#429)

abhi-mosaic committed 2 years ago

Verified ef350d9e

Older