question/problem about the training and the evaluation results

[DarlineFiedler]

I have no idea if the training and the evaluation are going to be done correctly. And how can I look at the values of the evaluation. Below I have listed the results I get and which data is generated. I hope someone can help me.

When I start the training it runs through without any errors, but I don't know if the console output is correct. I get this as last output:

when I run the time: !python automatische-generierung-von-titeln-und-abstracts/BertSum/src/train.py -mode train -encoder transformer -dropout 0.1 -bert_data_path bert_data/test/test -model_path automatische-generierung-von-titeln-und-abstracts/BertSum/models/test -lr 2e-3 -visible_gpus 0 -gpu_ranks 0 -world_size 1 -report_every 50 -save_checkpoint_steps 1000 -batch_size 3000 -decay_method noam -train_steps 000 -accum_count 2 -log_file automatische-generierung-von-titeln-und-abstracts/BertSum/logs/bert_transformer -use_interval true -warmup_steps 10000 -ff_size 2048 -inter_layers 2 -heads 8 and how many steps is the right number?

When I start the evaluation with the saved data I get this as last output:

I run this: !python automatische-generierung-von-titeln-und-abstracts/BertSum/src/train.py -mode validate -bert_data_path bert_data/test/test -model_path automatische-generierung-von-titeln-und-abstracts/BertSum/models/test -visible_gpus 0 -gpu_ranks 0 -batch_size 30000 -log_file automatische-generierung-von-titeln-und-abstracts/BertSum/logs/bert_transformer_vali -result_path automatische-generierung-von-titeln-und-abstracts/BertSum/results/test -bert_config_path automatische-generierung-von-titeln-und-abstracts/BertSum/bert_config_uncased_base.json -test_all -block_trigram true

But my results/test folder is simply empty. But there are some data in results. This one:

[DarlineFiedler]

And I have a second Question.

How can I generate and display specific example outputs? So that I give a text as input and the model then gives me the generated counterpart?

[Dejavu2019]

I donot know how to train the dates, could you help me? thank U very much!