enh - improve the text explanations

Author: contsili

examples/06_transfer_extend.ipynb

@@ -662,7 +662,9 @@
     "tags": []
    },
    "source": [
-    "These centiles look much better in comparison to the model we fit on the small dataset (especially at the age range 22-45).  "
+    "These centiles look much better in comparison to the model we fit on the small dataset (especially at the age range 22-45).\n",
+    "\n",
+    "This extended model can be extended again to another dataset. Such an example exists [here](https://github.com/predictive-clinical-neuroscience/pu25_code/blob/main/notebooks/additional_worflows/federated_learning_extend.ipynb)."
    ]
   },
   {
@@ -703,7 +705,7 @@
     "For this reason, we can _not_ use a transfered model to make predictions on the original train data and a model is not meant to be transferred more than once.\n",
     "\n",
     "Another consequence is that a transferred model centiles cover the age range present in the data you\n",
-    "transferred to (in our case between 20 and 65 years old), not the age range of the reference model."
+    "transferred to (see x-axis range in the centiles plot below), not the age range of the reference model."
    ]
   },
   {

examples/12_transfer_pretrained.ipynb

@@ -568,9 +568,7 @@
     "The transferred model is an ordinary PCNtoolkit model, so we can plot centiles\n",
     "for it just like any other.\n",
     "\n",
-    "The centiles only span roughly 18-26 years because that is the age range of\n",
-    "Beijing_Zang. A transferred model covers the ages present in the data you\n",
-    "transferred to, not the full lifespan of the reference model."
+    "Note that the centiles only span in the age range of the site that we selected to transfer to and not the the age range of the reference model."
    ]
   },
   {