RefalMachine's picture
Create README.md
9ba05c7 verified
|
raw
history blame
379 Bytes
metadata
base_model: >-
  RefalMachine/llama3_extended_darulm_20_05_24_part1-2_64000_bpe_full_lr2e4_bs256
datasets:
  - IlyaGusev/saiga_scored
language:
  - ru
  - en

Model description

LoRa tuned version of RefalMachine/llama3_extended_darulm_20_05_24_part1-2_64000_bpe_full_lr2e4_bs256 after LEP (Learned Embedding Propagation, paper will be soon) procedure on saiga_scored d7 dataset.