Fix rnn language model

[ehoelzl]

Fixes the LSTM language model

[codecov]

Codecov Report

Merging #303 (e860a9b) into develop (a656fd3) will decrease coverage by 0.23%.
The diff coverage is 18.81%.

@@             Coverage Diff             @@
##           develop     #303      +/-   ##
===========================================
- Coverage    24.88%   24.65%   -0.24%     
===========================================
  Files          105      108       +3     
  Lines         6204     6292      +88     
===========================================
+ Hits          1544     1551       +7     
- Misses        4660     4741      +81     

Impacted Files	Coverage Δ
mlbench_core/api.py	88.66% <ø> (ø)
mlbench_core/evaluation/pytorch/metrics.py	89.15% <ø> (ø)
mlbench_core/lr_scheduler/pytorch/__init__.py	100.00% <ø> (ø)
...ch_core/models/pytorch/language_models/__init__.py	0.00% <0.00%> (ø)
...lbench_core/models/pytorch/language_models/lstm.py	0.00% <0.00%> (ø)
mlbench_core/models/pytorch/layers/__init__.py	0.00% <0.00%> (ø)
...bench_core/models/pytorch/layers/dropout_layers.py	0.00% <0.00%> (ø)
mlbench_core/utils/task_args.py	0.00% <0.00%> (ø)
mlbench_core/evaluation/goals.py	58.18% <10.00%> (+4.95%)	⬆️
...ench_core/dataset/nlp/pytorch/wikitext2_dataset.py	18.68% <18.68%> (ø)
... and 13 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a656fd3...e860a9b. Read the comment docs.