Append </s> in pretraining chat template file

Update chat_template.jinja, which Transformers prioritizes over tokenizer_config.json, so the pretraining template emits </s> as EOD.

chat_template.jinja

@@ -6,6 +6,6 @@
 {%- elif message.role == "user" %}
 {{- "\nquestion: " + message.content }}
 {%- elif message.role == "assistant" %}
-{{- "\nanswer: " }}{% generation %}{{- message.content }}{% endgeneration %}
+{{- "\nanswer: " }}{% generation %}{{- message.content }}{{- "</s>" }}{% endgeneration %}
 {%- endif %}
 {%- endfor %}