Spaces:

lucalp
/

blt-entropy-patcher

Running on Zero

App Files Files Community

blt-entropy-patcher / bytelatent /train.py

Commit History

Get evals working again. (#46)

7517ac2
unverified

par-meta commited on Mar 11

Reduce per file resources arrow uses (#77)

63913e4
unverified

par-meta commited on Mar 5

Add approximate state persistence (#73)

ea1fc75
unverified

par-meta commited on Mar 5

Fix: Correct model_args usage in parallelize_model call (#69)

a6ed14f
unverified

Bocheng Li commited on Feb 24

Make it possible to specify multiple config files (#54)

82ab593
unverified

par-meta commited on Feb 18

Fix multiprocessing dataloader checkpointing and use it in the train script (#50)

8c61ab5
unverified

par-meta commited on Feb 13

make sure max_encoder_seq_length matches (#55)

48e4ad0
unverified

Srinivasan Iyer

sviyer commited on Feb 13

Add bpb and n_bytes to metric logging (#41)

fe45f69
unverified

par-meta commited on Feb 7

Update checkpointing to use fsspec (#39)

afedb16
unverified

par-meta commited on Feb 6

Add rope fp32 (#43)

739dc71
unverified

Srinivasan Iyer

sviyer commited on Feb 6

Fix distributed all reduce grad norm (#40)

c79b1fd
unverified

par-meta commited on Feb 5

This includes fixes that make checkpointing and reloading work correctly. (#35)

7044771
unverified

par-meta commited on Jan 28

Initial codes and scripts for training entropy model (#34)

7622d28
unverified

par-meta commited on Jan 27

Changes for training entropy model and correcting attention in local models (#25)

6ffeb66
unverified

par-meta commited on Jan 17

Initial commit

bcc039b

par-meta commited on Dec 12, 2024

Commit History

Get evals working again. (#46) 7517ac2 unverified

Reduce per file resources arrow uses (#77) 63913e4 unverified

Add approximate state persistence (#73) ea1fc75 unverified

Fix: Correct model_args usage in parallelize_model call (#69) a6ed14f unverified

Make it possible to specify multiple config files (#54) 82ab593 unverified

Fix multiprocessing dataloader checkpointing and use it in the train script (#50) 8c61ab5 unverified

make sure max_encoder_seq_length matches (#55) 48e4ad0 unverified

Add bpb and n_bytes to metric logging (#41) fe45f69 unverified

Update checkpointing to use fsspec (#39) afedb16 unverified

Add rope fp32 (#43) 739dc71 unverified

Fix distributed all reduce grad norm (#40) c79b1fd unverified

This includes fixes that make checkpointing and reloading work correctly. (#35) 7044771 unverified

Initial codes and scripts for training entropy model (#34) 7622d28 unverified

Changes for training entropy model and correcting attention in local models (#25) 6ffeb66 unverified

Initial commit bcc039b