diff --git a/README.md b/README.md
index 8446494..e8a6f6e 100644
--- a/README.md
+++ b/README.md
@@ -693,8 +693,8 @@ Before creating your own setup, have a look into the
 familiar with the Pytorch-IE core concepts like the `document`, `model`, and `taskmodule`.
 
 1. Write your PyTorch-IE dataset loader (see [dataset_builders/pie/conll2003/conll2003.py](dataset_builders/pie/conll2003/conll2003.py) for an example) or try out one of the PIE datasets hosted at [huggingface.co/pie](https://huggingface.co/pie).
-2. Write your PyTorch-IE model (see [src/models/transformer_token_classification.py](src/models/transformer_token_classification.py) for an example) or use one of the implementations from [pytorch-ie](https://github.com/ChristophAlt/pytorch-ie) or [pie-models](https://github.com/ArneBinder/pie-models).
-3. Write your PyTorch-IE taskmodule (see [src/taskmodules/transformer_token_classification.py](src/taskmodules/transformer_token_classification.py) for example) or use one of the implementations from [pytorch-ie](https://github.com/ChristophAlt/pytorch-ie) or [pie-models](https://github.com/ArneBinder/pie-models).
+2. Write your PyTorch-IE model (see [src/models/transformer_token_classification.py](src/models/transformer_token_classification.py) for an example) or use one of the implementations from [pytorch-ie](https://github.com/ChristophAlt/pytorch-ie) or [pie-modules](https://github.com/ArneBinder/pie-modules).
+3. Write your PyTorch-IE taskmodule (see [src/taskmodules/transformer_token_classification.py](src/taskmodules/transformer_token_classification.py) for example) or use one of the implementations from [pytorch-ie](https://github.com/ChristophAlt/pytorch-ie) or [pie-modules](https://github.com/ArneBinder/pie-modules).
 4. Write your experiment config, containing paths to your model, taskmodule and dataset (see [configs/experiment/conll2003.yaml](configs/experiment/conll2003.yaml) for example). You may need to also write configs for your model, taskmodule and dataset, if you do not want to use the default ones.
 5. If necessary, define `additional_model_kwargs` for your model class in the [train.py](src/train.py) (see line with `# NOTE: MODIFY THE additional_model_kwargs IF YOUR MODEL REQUIRES ...`").
 6. Execute a dev run for your setup to ensure that everything works as expected (assuming that `configs/experiments/experiment_name.yaml` is your experiment config file): `python src/train.py experiment=experiment_name +trainer.fast_dev_run=true`
diff --git a/requirements.txt b/requirements.txt
index 4cc30e6..9371e59 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,6 +1,6 @@
 # --------- pytorch-ie --------- #
 pytorch-ie>=0.28.0,<0.30.0
-pie-datasets>=0.3.3,<0.5.0
+pie-datasets>=0.6.0,<0.7.0
 
 # --------- hydra --------- #
 hydra-core>=1.3.0