Upload folder using huggingface_hub

.gitattributes

@@ -57,3 +57,16 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+Dataset/api_sequence_split.7z.001 filter=lfs diff=lfs merge=lfs -text
+Dataset/api_sequence_split.7z.002 filter=lfs diff=lfs merge=lfs -text
+Dataset/api_sequence_split.7z.003 filter=lfs diff=lfs merge=lfs -text
+Dataset/asm_split.7z.001 filter=lfs diff=lfs merge=lfs -text
+Dataset/asm_split.7z.002 filter=lfs diff=lfs merge=lfs -text
+Dataset/asm_split.7z.003 filter=lfs diff=lfs merge=lfs -text
+Dataset/asm_split.7z.004 filter=lfs diff=lfs merge=lfs -text
+Dataset/asm_split.7z.005 filter=lfs diff=lfs merge=lfs -text
+Dataset/asm_split.7z.006 filter=lfs diff=lfs merge=lfs -text
+Dataset/asm_split.7z.007 filter=lfs diff=lfs merge=lfs -text
+Dataset/cape_reports_split.7z.001 filter=lfs diff=lfs merge=lfs -text
+Dataset/cape_reports_split.7z.002 filter=lfs diff=lfs merge=lfs -text
+Dataset/cape_reports_split.7z.003 filter=lfs diff=lfs merge=lfs -text

Dataset/api_sequence_split.7z.001

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:81605346e38866b327f5c9987b313d9ead50a97fa5188a3e4fdfdf2c7581ec9e
+size 4294967296

Dataset/api_sequence_split.7z.002

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92e116a399e7d409e9142f440769c54e8e51deca7a758ccd94e4a72995100790
+size 4294967296

Dataset/api_sequence_split.7z.003

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99508df99bafbfd2fe43769589324b16a32a973be890e1495342667e8d5123b8
+size 863996510

Dataset/asm_split.7z.001

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d138392dbee178501599ff550a6569df7604e39f63493de77064401b11f02deb
+size 4294967296

Dataset/asm_split.7z.002

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f40bdb44b07b268f5ac0a6dc9ba49c522cdaadccf9c207b95d1a63618afee5ad
+size 4294967296

Dataset/asm_split.7z.003

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f4c9695800ed358ef492c49f983559445a884d7ab8a3c0c0f974e8928f6c79c
+size 4294967296

Dataset/asm_split.7z.004

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:51666043349a769e8692ea988e9ea518f1c53513812f14c7c01c9557f7a3822b
+size 4294967296

Dataset/asm_split.7z.005

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d64ea8ee1cb3d475d0964c96655f2b71f2f0466809d92182dd78291a2ff6bde
+size 4294967296

Dataset/asm_split.7z.006

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14824755b257dc3f330f33c6f9c17f5596677d6fb0586a2e2b0f18de1cf1a951
+size 4294967296

Dataset/asm_split.7z.007

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e56316805f7718694010d96a604f6be84ee95a5c24d9a7bd9f51a57fc1548df1
+size 2536312314

Dataset/cape_reports_split.7z.001

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:421b78923b46f41b0390d8651616dcf65dadc9481ff4b23a197221b50e789193
+size 4294967296

Dataset/cape_reports_split.7z.002

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:efe6bddb55850279e17e5abcf1753e304f754a5d8af94f32aab59e6c05869217
+size 4294967296

Dataset/cape_reports_split.7z.003

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bdd59617c927b1cf3171de73d105647bf113621f1ec6cd31a0cbb9faf6300cfb
+size 473645877

Dataset/cape_reports_to_md.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5d8833036a580cbbcb3d3459903c307aff818a6f4a2a98eb3c822060b925254
+size 69799879

README.md

@@ -1,3 +1,102 @@
----
-license: cc-by-nc-sa-4.0
----
+---
+pretty_name: MELD-DS-451
+license: cc-by-nc-sa-4.0
+language:
+- en
+tags:
+- malware
+- cybersecurity
+- CAPE
+- Windows
+size_categories:
+- 10K<n<100K
+task_categories:
+- other
+---
+
+# Appendix: MELD-DS-451 Dataset Overview
+
+## Dataset Overview
+
+MELD-DS-451 contains **25,722 malicious samples** spanning **454 distinct malware families** collected from April 2020 to August 2025. All samples are uniquely identified by SHA-256 hashes and include precise "First Seen" timestamps.
+
+**Family Distribution Characteristics**: The dataset exhibits a typical long-tail distribution, with 35.7% singleton families (only 1 sample) and 65.0% small-scale families (≤5 samples). Head concentration is significant, with the top 5 families covering 44.6% of samples and the top 10 families covering 57.1% of samples. Major families include unknown (4,517 samples, 17.6%), LummaStealer (2,903 samples, 11.3%), Formbook (2,057 samples, 8.0%), and others.
+
+**Temporal Evolution Patterns**: Samples are primarily concentrated in 2024-2025 (96.8%), with an average family lifespan of 93 days and only 4 families remaining active for more than one year. The number of active families surged to 413 in 2025, demonstrating the rapid evolution characteristics of contemporary malware ecosystems. Some families exhibit burst growth patterns, such as CryptOne with a monthly burst ratio of 9.8x.
+
+### Dataset Statistical Overview
+
+| Metric | Value | Description |
+|--------|-------|-------------|
+| **Total Families** | 454 | Total number of distinct malware families |
+| **Total Samples** | 25,722 | Total number of malicious samples |
+| **Avg Samples/Family** | 56.7 | Average samples per family |
+| **Sample Count Median** | 3 | Median of family sample counts |
+| **Singleton Families** | 162 (35.7%) | Families with only 1 sample |
+| **Small Families** | 295 (65.0%) | Families with ≤5 samples |
+| **Large Families** | 37 (8.1%) | Families with ≥100 samples |
+| **Top 5 Coverage** | 44.6% | Sample coverage by top 5 families |
+| **Top 10 Coverage** | 57.1% | Sample coverage by top 10 families |
+
+### Annual Evolution Statistics
+
+| Year | Active Families | Samples | Percentage |
+|------|----------------|---------|------------|
+| 2020 | 1 | 5 | 0.0% |
+| 2021 | 2 | 257 | 1.0% |
+| 2022 | 3 | 343 | 1.3% |
+| 2023 | 4 | 227 | 0.9% |
+| 2024 | 181 | 5,088 | 19.8% |
+| 2025 | 413 | 19,802 | 77.0% |
+
+## Standardized Analysis Artifacts
+
+Each sample in MELD-DS-451 provides four types of standardized analysis data generated through unified CAPE Sandbox analysis in virtualized Windows 10 x64 (22H2) environments:
+
+**1. CAPE JSON Reports** - Complete structured analysis results containing behavioral indicators, network activities, file system operations, registry modifications, and process execution traces, as the original analysis reports from CAPEv2.
+
+**2. Markdown Structured Reports** - Converting CAPE JSON reports into LLM-friendly structured Markdown format containing complete behavioral events, API call patterns, process tree information, and temporal analysis. These reports are specifically designed for large language model processing and understanding.
+
+**3. API Call Sequences** - Chronologically ordered sequences of Windows API function calls captured during dynamic execution, including parameters and return values, converted from CAPEv2's JSON reports. These sequences enable fine-grained behavioral modeling and sequence-based machine learning approaches.
+
+**4. ASM Disassembly Files** - Static disassembly output providing low-level instruction sequences and control flow information. These artifacts support static analysis techniques and hybrid approaches combining static and dynamic features.
+
+## Data Quality and Coverage
+
+All 25,722 samples (100% coverage) include complete metadata and all four analysis artifact types. The dataset ensures sample uniqueness through SHA-256 deduplication and maintains temporal consistency with verified timestamps. File sizes range from 87.3 KB to 301.3 MB (median: 3.5 MB), with the complete dataset totaling 479 GB of analysis artifacts and metadata.
+
+## Dataset File Structure
+
+The dataset files are organized in the `Dataset/` directory with large files split into volumes for easier download and Git LFS compatibility:
+
+### File Restoration Instructions
+
+Due to file size limitations, large dataset files have been split into 4GB volumes. To restore the original files, use the following commands:
+
+**1. ASM Disassembly Files (27GB total)**
+```bash
+7z x asm_split.7z.001
+```
+
+**2. API Call Sequences (8.9GB total)**  
+```bash
+7z x api_sequence_split.7z.001
+```
+
+**3. CAPE JSON Reports (8.5GB total)**
+```bash
+7z x cape_reports_split.7z.001
+```
+
+**4. Markdown Reports (67MB - no splitting needed)**
+- File: `cape_reports_to_md.zip`
+- Can be extracted directly: `unzip cape_reports_to_md.zip`
+
+### Requirements
+- **7-Zip**: Required for extracting split archives
+- **Disk Space**: Ensure at least 50GB free space for extraction
+- **Memory**: Recommended 8GB+ RAM for processing large files
+
+## License
+
+This project is licensed under the [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/) (Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International) license.

README_ZH_CH.md

@@ -0,0 +1,87 @@
+
+# 附录：MELD-DS-451数据集概述
+
+## 数据集概览
+
+MELD-DS-451包含**25,722个恶意样本**，涵盖**454个不同恶意软件家族**，收集时间跨度为2020年4月至2025年8月。所有样本均通过SHA-256哈希值唯一标识，并包含精确的"首次发现"时间戳。
+
+**家族分布特征**：数据集呈现典型的长尾分布，其中35.7%为单例家族（仅1个样本），65.0%为小规模家族（≤5个样本）。头部集中度显著，前5个家族覆盖44.6%的样本，前10个家族覆盖57.1%的样本。主要家族包括unknown（4,517个样本，17.6%）、LummaStealer（2,903个样本，11.3%）、Formbook（2,057个样本，8.0%）等。
+
+**时间演化模式**：样本主要集中在2024-2025年（96.8%），平均家族生命周期为93天，仅4个家族持续活跃超过1年。2025年活跃家族数量激增至413个，显示了当代恶意软件生态的快速演化特征。部分家族表现出突发增长模式，如CryptOne的月度突发比率达9.8倍。
+
+### 数据集统计概览
+
+| 指标 | 数值 | 说明 |
+|------|------|------|
+| **总家族数** | 454 | 不同恶意软件家族总数 |
+| **总样本数** | 25,722 | 恶意样本总数 |
+| **平均样本/家族** | 56.7 | 每个家族平均样本数 |
+| **样本数中位数** | 3 | 家族样本数的中位数 |
+| **单例家族** | 162 (35.7%) | 仅有1个样本的家族 |
+| **小规模家族** | 295 (65.0%) | ≤5个样本的家族 |
+| **大型家族** | 37 (8.1%) | ≥100个样本的家族 |
+| **前5家族覆盖率** | 44.6% | 前5大家族的样本占比 |
+| **前10家族覆盖率** | 57.1% | 前10大家族的样本占比 |
+
+### 年度演化统计
+
+| 年份 | 活跃家族数 | 样本数 | 占比 |
+|------|-----------|--------|------|
+| 2020 | 1 | 5 | 0.0% |
+| 2021 | 2 | 257 | 1.0% |
+| 2022 | 3 | 343 | 1.3% |
+| 2023 | 4 | 227 | 0.9% |
+| 2024 | 181 | 5,088 | 19.8% |
+| 2025 | 413 | 19,802 | 77.0% |
+
+## 标准化分析数据
+
+MELD-DS-451中的每个样本都提供四种类型的标准化分析数据，这些数据通过在虚拟化的Windows 10 x64 (22H2)环境中进行统一的CAPE沙箱分析生成：
+
+**1. CAPE JSON报告** - 包含行为指标、网络活动、文件系统操作、注册表修改和进程执行轨迹的完整结构化分析结果，为CAPEv2的原始分析报告。
+
+**2. Markdown结构化报告** - 将CAPE JSON报告转换为LLM友好的结构化Markdown格式，包含完整的行为事件、API调用模式、进程树信息和时序分析。这些报告专为大语言模型处理和理解而优化。
+
+**3. API调用序列** - 动态执行期间捕获的Windows API函数调用的时间顺序序列，包括参数和返回值，由CAPEv2的JSON报告转换。这些序列支持细粒度行为建模和基于序列的机器学习方法。
+
+**4. ASM反汇编文件** - 提供低级指令序列和控制流信息的静态反汇编输出。这些数据支持静态分析技术以及结合静态和动态特征的混合方法。
+
+## 数据质量与覆盖率
+
+所有25,722个样本（100%覆盖率）都包含完整的元数据和全部四种分析数据类型。数据集通过SHA-256去重确保样本唯一性，并通过验证时间戳维护时间一致性。文件大小范围从87.3 KB到301.3 MB（中位数：3.5 MB），完整数据集的分析数据和元数据总计479 GB。
+
+## 数据集文件结构
+
+数据集文件存放在 `Dataset/` 目录中，大文件已分卷处理以便下载和Git LFS兼容：
+
+### 文件还原说明
+
+由于文件大小限制，大型数据集文件已分割为4GB卷。要还原原始文件，请使用以下命令：
+
+**1. ASM反汇编文件（总计27GB）**
+```bash
+7z x asm_split.7z.001
+```
+
+**2. API调用序列（总计8.9GB）**  
+```bash
+7z x api_sequence_split.7z.001
+```
+
+**3. CAPE JSON报告（总计8.5GB）**
+```bash
+7z x cape_reports_split.7z.001
+```
+
+**4. Markdown报告（67MB - 无需分割）**
+- 文件：`cape_reports_to_md.zip`
+- 可直接解压：`unzip cape_reports_to_md.zip`
+
+### 系统要求
+- **7-Zip**：解压分卷文件所需
+- **磁盘空间**：确保至少50GB可用空间用于解压
+- **内存**：建议8GB+内存处理大文件
+
+## 许可证
+
+本项目使用 [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/)（署名-非商业性使用-相同方式共享 4.0 国际）许可协议。