README.md

metadata

language: 'no'
license: CC-BY 4.0
tags:
  - seq2seq
datasets:
  - Norwegian Nynorsk/Bokmål

🇳🇴 Norwegian T5 Base model 🇳🇴

This T5-base model is trained from scratch on a 19GB Balanced Bokmål-Nynorsk Corpus.

Update: Due to disk space errors, the model had to be restarted July 20. It is currently still running.

Parameters used in training:

python3 ./run_t5_mlm_flax_streaming.py 
    --model_name_or_path="./norwegian-t5-base"
    --output_dir="./norwegian-t5-base" 
    --config_name="./norwegian-t5-base" 
    --tokenizer_name="./norwegian-t5-base" 
    --dataset_name="pere/nb_nn_balanced_shuffled"  
    --max_seq_length="512" 
    --per_device_train_batch_size="32" 
    --per_device_eval_batch_size="32" 
    --learning_rate="0.005" 
    --weight_decay="0.001" 
    --warmup_steps="2000" 
    --overwrite_output_dir  
    --logging_steps="100" 
    --save_steps="500" 
    --eval_steps="500"
    --push_to_hub 
    --preprocessing_num_workers 96 
    --adafactor