Kai Wu
|
e6f69f84ad
add max_steps_reached to reduce redundancy
|
1 year ago |
Kai Wu
|
fa0a389f74
add max_step feature for training and eval
|
1 year ago |
jpgard
|
6954b16b3b
only save training params on rank 0
|
1 year ago |
Hamid Shojanazeri
|
761b7e6e51
adding wandb_run ro eval
|
1 year ago |
Hamid Shojanazeri
|
ffdc93f00a
Merge branch 'main' into wandb_logging
|
1 year ago |
Matthias Reso
|
c5a382e509
Make tests run on cpu only machines
|
1 year ago |
Hamid Shojanazeri
|
162be4c045
Revert "Flop counter, profiling and GC (#357)"
|
1 year ago |
Hamid Shojanazeri
|
1a09fb5d27
add logging for setting profiler
|
1 year ago |
Hamid Shojanazeri
|
71d137c722
Merge branch 'main' into flop_counter_gc
|
1 year ago |
Hamid Shojanazeri
|
8bf474b455
clean up
|
1 year ago |
Hamid Shojanazeri
|
19089269d3
add gc
|
1 year ago |
Hamid Shojanazeri
|
dbfea484c6
Feature : Enable Intel GPU/XPU finetuning and inference (#116)
|
1 year ago |
Beto
|
1f5b202c18
Adding tests for the save_metrics param in the train function
|
1 year ago |
Beto
|
7474514fe0
Merging with main
|
1 year ago |
gaopengzhi
|
c7d410725b
Merge branch 'main' into grad_clip
|
1 year ago |
Abhilash Majumder
|
4793f0fdf3
Merge branch 'main' into ipex_feature
|
1 year ago |
gaopengzhi
|
e2797abe9b
Add gradient_clipping and gradient_clipping_threshold parameters
|
1 year ago |
kldarek
|
fc5485d916
fixing wandb for fsdp
|
1 year ago |
gaopengzhi
|
bb7c6c1e33
Support FSDP scenario
|
1 year ago |
kldarek
|
cf373529f7
basic wandb logging instrumentation
|
1 year ago |
gaopengzhi
|
b1d9efd155
Refactor gradient clipping feature
|
1 year ago |
Beto
|
17d02c3b44
Adding config to conditionally save stats
|
1 year ago |
Beto
|
b974c87035
Merging latest from main
|
1 year ago |
Jeremy Howard
|
eca8410b32
Use bf16 parameters in bf16 mixed prec
|
1 year ago |
gaopengzhi
|
04befdef69
Add gradient clipping feature
|
1 year ago |
Abhilash Majumder
|
11465d6329
remove duplicate import
|
1 year ago |
Abhilash Majumder
|
6a78b96764
Merge branch 'main' into ipex_feature
|
1 year ago |
Matthias Reso
|
e8bb7fbabc
Merge remote-tracking branch 'origin/main' into feature/length_based_batch_sampling
|
1 year ago |
Matthias Reso
|
33925f71e6
Add missing amp context if use_fp16 is enabled
|
1 year ago |
Hamid Shojanazeri
|
35b394e49f
adding profiler and flop_counter
|
1 year ago |