Commits

Commits on Apr 18, 2024

exclude transformers 4.38 & 4.39
warner-benjamin
committedApr 18, 2024

Commits on Apr 17, 2024

Merge pull request AnswerDotAI#52 from AnswerDotAI/llama-pro
KeremTurgutlu
committedApr 17, 2024
Merge pull request AnswerDotAI#51 from AnswerDotAI/dora
KeremTurgutlu
committedApr 17, 2024

Commits on Apr 16, 2024

quantized llama-pro
KeremTurgutlu
committedApr 16, 2024
update readme
KeremTurgutlu
committedApr 16, 2024
minor fixes
KeremTurgutlu
committedApr 16, 2024
add tests
KeremTurgutlu
committedApr 16, 2024
add bnb_dora and hqq_dora changes
KeremTurgutlu
committedApr 16, 2024

Commits on Mar 28, 2024

Merge pull request AnswerDotAI#39 from AnswerDotAI/reserved_memory
warner-benjamin
committedMar 28, 2024

Commits on Mar 27, 2024

print reserved memory, log allocated & reserved
warner-benjamin
committedMar 27, 2024

Commits on Mar 26, 2024

add additional model instructions, mistral support, & small cleanup
warner-benjamin
committedMar 26, 2024

Commits on Mar 21, 2024

add loading progress bar & fix verbose printing
warner-benjamin
committedMar 21, 2024

Commits on Mar 19, 2024

Merge pull request AnswerDotAI#35 from AnswerDotAI/dataset_samples
johnowhitaker
committedMar 19, 2024
Add dataset_samples arg for alpaca_sample and dummy datasets
warner-benjamin
committedMar 19, 2024

Commits on Mar 15, 2024

Merge pull request AnswerDotAI#30 from geronimi73/fix_ProcessExitedException
johnowhitaker
committedMar 15, 2024

Commits on Mar 14, 2024

changes for technical post & settable loading workers
warner-benjamin
committedMar 14, 2024

Commits on Mar 13, 2024

Merge branch 'AnswerDotAI:main' into fix_ProcessExitedException
geronimi73
committedMar 13, 2024
Revert "Launch world_size processs even when more GPUs are available"
johnowhitaker
committedMar 13, 2024

Commits on Mar 12, 2024

Merge branch 'main' of https://github.com/AnswerDotAI/fsdp_qlora
johnowhitaker
committedMar 12, 2024
Launch world_size processs even when more GPUs are available
johnowhitaker
committedMar 12, 2024
reduce n_workers for hqq
KeremTurgutlu
committedMar 12, 2024
use iterator during weight loading
KeremTurgutlu
committedMar 12, 2024

Commits on Mar 11, 2024

Axolotl integration
warner-benjamin
committedMar 11, 2024
Update README.md to add ntoe on RAM requirement
johnowhitaker
committedMar 11, 2024
move peft imports
geronimi73
committedMar 11, 2024

Commits on Mar 9, 2024

Update README.md
johnowhitaker
committedMar 9, 2024

Commits on Mar 8, 2024

Update bitsandbytes install to 0.43
warner-benjamin
committedMar 8, 2024
Update ReadMe for Release
warner-benjamin
committedMar 8, 2024

Commits on Mar 7, 2024

Update loading to match HQQ source
warner-benjamin
committedMar 7, 2024
adjust n_workers for model loading and reentrant grad ckpt
KeremTurgutlu
committedMar 7, 2024
Update README.md
johnowhitaker
committedMar 7, 2024
Stop recording memory history after profiling
johnowhitaker
committedMar 7, 2024
Merge pull request AnswerDotAI#17 from AnswerDotAI/arguments
johnowhitaker
committedMar 7, 2024
Merge pull request AnswerDotAI#16 from AnswerDotAI/license
johnowhitaker
committedMar 7, 2024

Commits on Mar 6, 2024

Add arguments for reentrant_checkpointing & wandb, and arg string pass
warner-benjamin
committedMar 6, 2024