Simplify model card

README.md

@@ -16,23 +16,21 @@ tags:
 # XQDT E2E verifier: qwen3 4B
 
 This repository contains the LoRA adapter for the **qwen3 4B**
-XQDT verifier reported in *XQDT: eXplainable and Quantitative Data-Text Alignment
-Metric with Feedback Signals*. The base model is
-[`Qwen/Qwen3-4B`](https://huggingface.co/Qwen/Qwen3-4B); base-model
-weights are not included here.
+XQDT verifier from *XQDT: eXplainable and Quantitative Data-Text Alignment Metric
+with Feedback Signals*. It is used with
+[`Qwen/Qwen3-4B`](https://huggingface.co/Qwen/Qwen3-4B).
 
 This E2E checkpoint was trained on the joint WebNLG--E2E synthetic training set.
 
-## Intended use
+## Overview
 
 XQDT verifies alignment between English text and structured triples. It returns
-`missing`, `extra`, and `incorrect` units, or `All correct`. In the paper's
-terminology, `missing` identifies an input unit omitted from the text; `extra`
-identifies text content unsupported by the input; and `incorrect` identifies an
-input unit realised with incorrect information.
+`missing`, `extra`, and `incorrect` units, or `All correct`. `missing` identifies
+an input unit omitted from the text, `extra` identifies text content unsupported
+by the input, and `incorrect` identifies an input unit realised with incorrect
+information.
 
-The expected prompt and four example inputs and outputs are provided in
-`smoke_test.json`.
+Example inputs and outputs are provided in `smoke_test.json`.
 
 ## Prompt format
 
@@ -44,10 +42,6 @@ TRIPLES:
 Output as markdown table with Type and Triple columns.
 ```
 
-The canonical implementation uses **ms-swift PtEngine**. Transformers and vLLM
-examples are also provided. Outputs may vary slightly across runtimes; evaluation
-uses the parsed error units.
-
 ## ms-swift PtEngine
 
 ```python
@@ -120,8 +114,6 @@ print(tokenizer.decode(generated, skip_special_tokens=True))
 
 ## vLLM
 
-Install the versions listed in `requirements.txt` before running this example.
-
 ```python
 from huggingface_hub import snapshot_download
 from vllm import LLM, SamplingParams
@@ -148,12 +140,6 @@ outputs = llm.chat(
 print(outputs[0].outputs[0].text)
 ```
 
-## Reproducibility
-
-- Training configuration: `training_manifest.json`
-- Example inputs and outputs: `smoke_test.json`
-- Code: [https://github.com/guihuzhang/xqdt](https://github.com/guihuzhang/xqdt)
-
 ## Citation
 
 ```bibtex