Enhance dataset card: update metadata and add sample usage

README.md

@@ -1,18 +1,28 @@
 ---
-license: cc-by-nc-sa-4.0
-task_categories:
-- question-answering
 language:
 - zh
 - en
-tags:
-- agent
+license: cc-by-nc-sa-4.0
 size_categories:
 - n<1K
+task_categories:
+- text-generation
+- image-to-image
+- image-to-video
+- automatic-speech-recognition
+- text-retrieval
+tags:
+- agent
+- benchmark
+- code-agent
+- software-engineering
+- multimodal
 ---
 
 # Dataset Card for **GitTaskBench**
 
+The dataset was presented in the paper [GitTaskBench: A Benchmark for Code Agents Solving Real-World Tasks Through Code Repository Leveraging](https://huggingface.co/papers/2508.18993).
+
 ## Dataset Details
 
 ### Dataset Description
@@ -23,7 +33,7 @@ It contains **54 representative tasks** across **7 domains**, carefully curated
 - **Funded by [optional]:** Not specified  
 - **Shared by [optional]:** GitTaskBench Team  
 - **Language(s):** Primarily English (task descriptions, documentation)  
-- **License:** [Specify license chosen, e.g., `cc-by-nc-sa-4.0`]  
+- **License:** `cc-by-nc-sa-4.0`  
 
 ### Dataset Sources
 - **Repository:** [GitTaskBench GitHub](https://github.com/QuantaAlpha/GitTaskBench)  
@@ -67,8 +77,63 @@ Each task specifies:
 
 ## Usage Example
 
-See GitHub Repository:** [GitTaskBench GitHub](https://github.com/QuantaAlpha/GitTaskBench)  
+To get started with GitTaskBench, follow these steps for environment setup and evaluation.
+
+### 1. Set Up ⚙️ 
+First, create a new conda environment:
+```console
+conda create -n gittaskbench python=3.10 -y
+conda activate gittaskbench
+
+pip install torch==1.11.0+cu113 torchvision==0.12.0+cu113 torchaudio==0.11.0 \
+  --extra-index-url https://download.pytorch.org/whl/cu113
+```
+
+Then, you can install `gittaskbench` with pip:
+```console
+git clone https://github.com/QuantaAlpha/GitTaskBench.git
+cd GitTaskBench
+# config
+pip install -e .
+```
+Alternatively:
+```console
+# config
+pip install -r requirements.txt
+```
+
+### 2. Quick Start 💡 
+
+*   #### **Single Task Evaluation:**
+
+If you need to evaluate a single, specific task, you can use the following command. The example below shows how to evaluate the `Trafilatura_01` task:
+
+```console
+cd GitTaskBench
+# The outputs are saved in the DEFAULT "./output" directory, for example: "./output/Trafilatura_01/output.txt"
+```
+
+```console
+gittaskbench grade --taskid Trafilatura_01
+```
+
+Running the command will produce an analysis report (.jsonl) at the DEFAULT path (./test_results/Trafilatura_01). See ```test_results_for_show/``` for a sample. 
 
+The complete commands can be found in the [🤖 Automation Evaluation](#automation-evaluation) section.
+
+*   #### **All Tasks Evaluation**
+When you need to evaluate all tasks, you can use the --all parameter. This command will automatically iterate through and execute the evaluation of all tasks:
+```console
+gittaskbench grade --all
+```
+
+*   #### **Test Results Analysis**
+After completing the evaluation, if you want to analyze & summary the test results, you can use the statistics command. This command will analyze & summary the evaluation results in the specified directory and output an analysis report (.txt):
+
+```console
+gittaskbench eval
+```
+See ```test_reports/``` for a sample.
 
 Each task entry contains:
 - **task_id**: Unique task identifier (e.g., `Trafilatura_01`)  
@@ -145,4 +210,4 @@ If you use GitTaskBench, please cite the paper:
 - Multi-modal tasks (vision, speech, text, signals).  
 - Repository-level evaluation.  
 - Real-world relevance (PDF extraction, video coloring, speech analysis, etc.).  
-- Extensible design for new tasks.  
+- Extensible design for new tasks.