Simplify model card tags and clarify model type

Changes:

1. **Removed unnecessary tags**

* `english`, `chat`, `chatbot`, `355m`, `gpt`, `gpt2`, `gpt2-medium` were removed.
* Reason: Some of these tags (e.g., `gpt2`) are automatically inferred from `config.json` because it contains `"model_type": "gpt2"`, and others are redundant or overly generic.

2. **Added/kept relevant tags**

* `conversational` -> indicates this model is a chatbot and supports conversational tasks.
* `anwgpt`, `anw`-> retained for brand/model recognition.

3. **Updated model type**

* Changed from `chatbot/conversational` to `text generator / question answering`
* Reason: better reflects the model’s capabilities, especially for tasks beyond just chat.

### Rationale

* Reduces clutter in the model card.
* Improves discoverability by focusing on functional and distinctive tags.
* Aligns model description with intended usage scenarios.

README.md

@@ -10,15 +10,9 @@ base_model:
 pipeline_tag: question-answering
 library_name: transformers
 tags:
-- english
-- chat
-- chatbot
+- conversational
 - anwgpt
 - anw
-- 355m
-- gpt
-- gpt2
-- gpt2-medium
 ---
 
 # ANWGPT3 (anwgpt3-355m)
@@ -28,7 +22,7 @@ Conversational version of ANWGPT2 (anwgpt2-355m).
 - **Developed by:** Subhrajit Sain, aka, ANW
 - **Funded by:** no one
 - **Shared by:** no one
-- **Model type:** chatbot/conversational
+- **Model type:** text generator / question answering
 - **Language (NLP):** English
 - **License:** MIT
 - **Finetuned from model:** SubhrajitSain/anwgpt2-355m