Commit History
Ax art (#405) 29241cf unverified
add utils.data.prepare_dataset 2e22404
use context manager to run things on rank0 before others (#397) fc2d6be unverified
Feat(config): add max steps (#387) 3c2ad00 unverified
save tokenizer before training starts (#380) 86a91e2 unverified
simplify `load_tokenizer` efb3b2c
improve GPU logging to break out pytorch cache and system mem 7b55fe6
extract module for working with cfg 8cec513
Attention mask and position id fixes for packing (#285) 2bb0b78 unverified
Fix(save): Save as safetensors (#363) a276c9c unverified
feat(merge): save tokenizer on merge (#362) 289d5c4 unverified
Merge pull request #356 from tmm1/load_model-args 11ddccb unverified
simplify load_model signature 7181022
log GPU memory usage e303d64
fix FSDP save of final model (#329) 894cba0 unverified
add runpod envs to .bashrc, fix bnb env (#316) cf62cfd unverified
misc fixes d75adb9
Fixed pre-commit problems, fixed small bug in logging_config to handle LOG_LEVEL env var b1f4f7a
Adding logging enhancement 553a86b
Merge pull request #92 from OpenAccess-AI-Collective/flash-optimum 16bb627 unverified
chore: Refactor inf_kwargs out dc77c8e
Merge branch 'main' into flash-optimum fd2c981 unverified
Merge pull request #177 from NanoCode012/fix/landmark-patch 8002ffb unverified
Merge pull request #159 from AngainorDev/patch-1 8e568bb unverified
Fix strict and Lint b565ecf
Fix set mem_id for inference and refactor 974dc00
Set mem cache args on inference 572d114
fix formatting 958da70
pass a prompt in from stdin for inference c4e4f81
address PR feedback 0c6f928
add streaming dataset support for pretraining datasets eea2731
more tweaks to do pre-training with bettertransformers 1210dc8
experimental expansion of ctx len 488a67d
add flash attn context for efficient training and attempt setting model to train mode: 8792199
add support for opimum bettertransformers 1edc30c
Merge branch 'main' into patch-1 79e2a6f unverified
Angainor Development commited on
Remove explicit definition of cfg.inference c250898 unverified
Angainor Development commited on
formatting for linter f36e227 unverified
Add streaming inference & fix stopping at EOS fec6bcc
Feed cfg.inference bd3b537 unverified
Angainor Development commited on