yandex / YaLM-100B

Pretrained language model with 100B parameters

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Model dataset irregularity

joshlk opened this issue Β· comments

Hi,

Thank you for releasing the model dataset πŸ˜ƒ!

Ive just downloaded the data and done a quick spot check on the sizes of the layers. Most layers have a size of 2.4Gb, however layer 44 is 817Mb - is there an issue with this layers data?

Full layers sizes:

501M layer_00-model_00-model_states.pt
9.0K layer_01-model_00-model_states.pt
2.4G layer_03-model_00-model_states.pt
2.4G layer_04-model_00-model_states.pt
2.4G layer_05-model_00-model_states.pt
2.4G layer_06-model_00-model_states.pt
2.4G layer_07-model_00-model_states.pt
2.4G layer_08-model_00-model_states.pt
2.4G layer_09-model_00-model_states.pt
2.4G layer_10-model_00-model_states.pt
2.4G layer_11-model_00-model_states.pt
2.4G layer_12-model_00-model_states.pt
2.4G layer_13-model_00-model_states.pt
2.4G layer_14-model_00-model_states.pt
2.4G layer_15-model_00-model_states.pt
2.4G layer_16-model_00-model_states.pt
2.4G layer_17-model_00-model_states.pt
2.4G layer_18-model_00-model_states.pt
2.4G layer_19-model_00-model_states.pt
2.4G layer_20-model_00-model_states.pt
2.4G layer_21-model_00-model_states.pt
2.4G layer_22-model_00-model_states.pt
2.4G layer_23-model_00-model_states.pt
2.4G layer_24-model_00-model_states.pt
2.4G layer_25-model_00-model_states.pt
2.4G layer_26-model_00-model_states.pt
2.4G layer_27-model_00-model_states.pt
2.4G layer_28-model_00-model_states.pt
2.4G layer_29-model_00-model_states.pt
2.4G layer_30-model_00-model_states.pt
2.4G layer_31-model_00-model_states.pt
2.4G layer_32-model_00-model_states.pt
2.4G layer_33-model_00-model_states.pt
2.4G layer_34-model_00-model_states.pt
2.4G layer_35-model_00-model_states.pt
2.4G layer_36-model_00-model_states.pt
2.4G layer_37-model_00-model_states.pt
2.4G layer_38-model_00-model_states.pt
2.4G layer_39-model_00-model_states.pt
2.4G layer_40-model_00-model_states.pt
2.4G layer_41-model_00-model_states.pt
2.4G layer_42-model_00-model_states.pt
2.4G layer_43-model_00-model_states.pt
817M layer_44-model_00-model_states.pt
1.8G layer_45-model_00-model_states.pt
2.4G layer_46-model_00-model_states.pt
2.4G layer_47-model_00-model_states.pt
2.4G layer_48-model_00-model_states.pt
2.4G layer_49-model_00-model_states.pt
2.4G layer_50-model_00-model_states.pt
2.4G layer_51-model_00-model_states.pt
2.4G layer_52-model_00-model_states.pt
2.4G layer_53-model_00-model_states.pt
2.4G layer_54-model_00-model_states.pt
2.4G layer_55-model_00-model_states.pt
2.4G layer_56-model_00-model_states.pt
2.4G layer_57-model_00-model_states.pt
2.4G layer_58-model_00-model_states.pt
2.4G layer_59-model_00-model_states.pt
2.4G layer_60-model_00-model_states.pt
2.4G layer_61-model_00-model_states.pt
2.4G layer_62-model_00-model_states.pt
2.4G layer_63-model_00-model_states.pt
2.4G layer_64-model_00-model_states.pt
2.4G layer_65-model_00-model_states.pt
2.4G layer_66-model_00-model_states.pt
2.4G layer_67-model_00-model_states.pt
2.4G layer_68-model_00-model_states.pt
2.4G layer_69-model_00-model_states.pt
2.4G layer_70-model_00-model_states.pt
2.4G layer_71-model_00-model_states.pt
2.4G layer_72-model_00-model_states.pt
2.4G layer_73-model_00-model_states.pt
2.4G layer_74-model_00-model_states.pt
2.4G layer_75-model_00-model_states.pt
2.4G layer_76-model_00-model_states.pt
2.4G layer_77-model_00-model_states.pt
2.4G layer_78-model_00-model_states.pt
2.4G layer_79-model_00-model_states.pt
2.4G layer_80-model_00-model_states.pt
2.4G layer_81-model_00-model_states.pt
2.4G layer_82-model_00-model_states.pt
 41M layer_84-model_00-model_states.pt