SparkBeyond/roberta-large-sts-b

The model is a fine tuned RoBERTA model over STS-B . It was trained with these parameters: --learning_rate 2e-5 and --num_train_epochs 3.0 . The model was trained using these params: --max_seq_length 128, --per_gpu_eval_eval batch_size=8, --learning rate 2-5, --num train epoch epooms 3.0, --nose_data/STS-b// --output_dir /content/roberta-sts-b; --data_dir/glue_data; --Data_,