Remove text-gen-custom-gpt2 model as it's quite large and not necessary for testing purposes (+ remove references in docs) (#6998)

Author: tyndria

docs-gb/servers/huggingface.md

@@ -92,8 +92,4 @@ spec:
         value: text-generation
     name: default
     replicas: 1
-```
-
-{% hint style="note" %}
-As a next step, why not try running a larger-scale model? You can find one in gs://seldon-models/v1.18.0/huggingface/text-gen-custom-gpt2. However, you may need to request more memory!
-{% endhint %}
+```

servers/Makefile

@@ -11,6 +11,5 @@ sklearnserver-models:
 
 
 huggingface-models:
-	make -C huggingface/models/text-gen-gpt2
 	make -C huggingface/models/text-gen-tiny-stories