Hugging Face's logo

StepLaw
community
https://step-law.github.io/
step-law
Activity Feed Request to join this org

AI & ML interests

None defined yet.

Recent Activity

brain-zhang  authored a paper 5 days ago
Step-3 is Large yet Affordable: Model-system Co-design for Cost-effective Decoding
zhengwenzhen  authored a paper about 2 months ago
Farseer: A Refined Scaling Law in Large Language Models
brain-zhang  updated a model 4 months ago
StepLaw/StepLaw-N_59M-D_7.0B-LR7.812e-03-BS1572864
View all activity

StepLaw 's collections 41

StepLaw-N_59M-D_7.0B
Models with 59M parameters trained with 7.0B tokens.
StepLaw-N_59M-D_1.0B
Models with 59M parameters trained with 1.0B tokens.
StepLaw-N_536M-D_7.0B
Models with 536M parameters trained with 7.0B tokens.
StepLaw-N_536M-D_3.0B
Models with 536M parameters trained with 3.0B tokens.
StepLaw-N_536M-D_19.0B
Models with 536M parameters trained with 19.0B tokens.
StepLaw-N_429M-D_99.0B
Models with 429M parameters trained with 99.0B tokens.
StepLaw-N_429M-D_49.0B
Models with 429M parameters trained with 49.0B tokens.
StepLaw-N_429M-D_3.0B
Models with 429M parameters trained with 3.0B tokens.
StepLaw-N_429M-D_19.0B
Models with 429M parameters trained with 19.0B tokens.
StepLaw-N_268M-D_99.0B
Models with 268M parameters trained with 99.0B tokens.
StepLaw-N_268M-D_7.0B
Models with 268M parameters trained with 7.0B tokens.
StepLaw-N_268M-D_3.0B
Models with 268M parameters trained with 3.0B tokens.
StepLaw-N_268M-D_19.0B
Models with 268M parameters trained with 19.0B tokens.
StepLaw-N_268M-D_1.0B
Models with 268M parameters trained with 1.0B tokens.
StepLaw-N_214M-D_7.0B
Models with 214M parameters trained with 7.0B tokens.
StepLaw-N_214M-D_3.0B
Models with 214M parameters trained with 3.0B tokens.
StepLaw-N_214M-D_1.0B
Models with 214M parameters trained with 1.0B tokens.
StepLaw-N_119M-D_3.0B
Models with 119M parameters trained with 3.0B tokens.
StepLaw-N_1.0B-D_7.0B
Models with 1.0B parameters trained with 7.0B tokens.
StepLaw-N_1.0B-D_3.0B
Models with 1.0B parameters trained with 3.0B tokens.
StepLaw-N_1.0B-D_1.0B
Models with 1.0B parameters trained with 1.0B tokens. Architecture: H=2048, FFN=8192, Heads=16, Layers=16.
StepLaw-N_59M-D_3.0B
Models with 59M parameters trained with 3.0B tokens.
StepLaw-N_536M-D_9.0B
Models with 536M parameters trained with 9.0B tokens.
StepLaw-N_536M-D_49.0B
Models with 536M parameters trained with 49.0B tokens.
StepLaw-N_536M-D_28.0B
Models with 536M parameters trained with 28.0B tokens.
StepLaw-N_536M-D_1.0B
Models with 536M parameters trained with 1.0B tokens.
StepLaw-N_429M-D_7.0B
Models with 429M parameters trained with 7.0B tokens.
StepLaw-N_429M-D_39.0B
Models with 429M parameters trained with 39.0B tokens.
StepLaw-N_429M-D_22.0B
Models with 429M parameters trained with 22.0B tokens.
StepLaw-N_429M-D_1.0B
Models with 429M parameters trained with 1.0B tokens.
StepLaw-N_268M-D_79.0B
Models with 268M parameters trained with 79.0B tokens.
StepLaw-N_268M-D_4.0B
Models with 268M parameters trained with 4.0B tokens.
StepLaw-N_268M-D_24.0B
Models with 268M parameters trained with 24.0B tokens.
StepLaw-N_268M-D_14.0B
Models with 268M parameters trained with 14.0B tokens.
StepLaw-N_214M-D_99.0B
Models with 214M parameters trained with 99.0B tokens.
StepLaw-N_214M-D_19.0B
Models with 214M parameters trained with 19.0B tokens.
StepLaw-N_214M-D_11.0B
Models with 214M parameters trained with 11.0B tokens.
StepLaw-N_119M-D_7.0B
Models with 119M parameters trained with 7.0B tokens.
StepLaw-N_119M-D_1.0B
Models with 119M parameters trained with 1.0B tokens.
StepLaw-N_1.0B-D_56.0B
Models with 1.0B parameters trained with 56.0B tokens.
StepLaw-N_1.0B-D_19.0B
Models with 1.0B parameters trained with 19.0B tokens.
StepLaw-N_59M-D_7.0B
Models with 59M parameters trained with 7.0B tokens.
StepLaw-N_59M-D_3.0B
Models with 59M parameters trained with 3.0B tokens.
StepLaw-N_59M-D_1.0B
Models with 59M parameters trained with 1.0B tokens.
StepLaw-N_536M-D_9.0B
Models with 536M parameters trained with 9.0B tokens.
StepLaw-N_536M-D_7.0B
Models with 536M parameters trained with 7.0B tokens.
StepLaw-N_536M-D_49.0B
Models with 536M parameters trained with 49.0B tokens.
StepLaw-N_536M-D_3.0B
Models with 536M parameters trained with 3.0B tokens.
StepLaw-N_536M-D_28.0B
Models with 536M parameters trained with 28.0B tokens.
StepLaw-N_536M-D_19.0B
Models with 536M parameters trained with 19.0B tokens.
StepLaw-N_536M-D_1.0B
Models with 536M parameters trained with 1.0B tokens.
StepLaw-N_429M-D_99.0B
Models with 429M parameters trained with 99.0B tokens.
StepLaw-N_429M-D_7.0B
Models with 429M parameters trained with 7.0B tokens.
StepLaw-N_429M-D_49.0B
Models with 429M parameters trained with 49.0B tokens.
StepLaw-N_429M-D_39.0B
Models with 429M parameters trained with 39.0B tokens.
StepLaw-N_429M-D_3.0B
Models with 429M parameters trained with 3.0B tokens.
StepLaw-N_429M-D_22.0B
Models with 429M parameters trained with 22.0B tokens.
StepLaw-N_429M-D_19.0B
Models with 429M parameters trained with 19.0B tokens.
StepLaw-N_429M-D_1.0B
Models with 429M parameters trained with 1.0B tokens.
StepLaw-N_268M-D_99.0B
Models with 268M parameters trained with 99.0B tokens.
StepLaw-N_268M-D_79.0B
Models with 268M parameters trained with 79.0B tokens.
StepLaw-N_268M-D_7.0B
Models with 268M parameters trained with 7.0B tokens.
StepLaw-N_268M-D_4.0B
Models with 268M parameters trained with 4.0B tokens.
StepLaw-N_268M-D_3.0B
Models with 268M parameters trained with 3.0B tokens.
StepLaw-N_268M-D_24.0B
Models with 268M parameters trained with 24.0B tokens.
StepLaw-N_268M-D_19.0B
Models with 268M parameters trained with 19.0B tokens.
StepLaw-N_268M-D_14.0B
Models with 268M parameters trained with 14.0B tokens.
StepLaw-N_268M-D_1.0B
Models with 268M parameters trained with 1.0B tokens.
StepLaw-N_214M-D_99.0B
Models with 214M parameters trained with 99.0B tokens.
StepLaw-N_214M-D_7.0B
Models with 214M parameters trained with 7.0B tokens.
StepLaw-N_214M-D_19.0B
Models with 214M parameters trained with 19.0B tokens.
StepLaw-N_214M-D_3.0B
Models with 214M parameters trained with 3.0B tokens.
StepLaw-N_214M-D_11.0B
Models with 214M parameters trained with 11.0B tokens.
StepLaw-N_214M-D_1.0B
Models with 214M parameters trained with 1.0B tokens.
StepLaw-N_119M-D_7.0B
Models with 119M parameters trained with 7.0B tokens.
StepLaw-N_119M-D_3.0B
Models with 119M parameters trained with 3.0B tokens.
StepLaw-N_119M-D_1.0B
Models with 119M parameters trained with 1.0B tokens.
StepLaw-N_1.0B-D_7.0B
Models with 1.0B parameters trained with 7.0B tokens.
StepLaw-N_1.0B-D_56.0B
Models with 1.0B parameters trained with 56.0B tokens.
StepLaw-N_1.0B-D_3.0B
Models with 1.0B parameters trained with 3.0B tokens.
StepLaw-N_1.0B-D_19.0B
Models with 1.0B parameters trained with 19.0B tokens.
StepLaw-N_1.0B-D_1.0B
Models with 1.0B parameters trained with 1.0B tokens. Architecture: H=2048, FFN=8192, Heads=16, Layers=16.