huggingface-trends

从 Hugging Face 监控和获取趋势模型，支持按任务、库和流行度指标进行过滤。当用户想要检查趋势 AI 模型、比较模型受欢迎程度或按任务或库探索流行模型时使用。支持导出为 JSON 和格式化输出。

# Hugging Face Trending Models

## Quick Start

Fetch the top trending models:

```bash
scripts/hf_trends.py -n 10 -p http://172.28.96.1:10808
```

## Core Features

### Fetch Trending Models

Basic usage:

```bash
# Get top 10 trending models
scripts/hf_trends.py -n 10 -p http://172.28.96.1:10808

# Get top 5 most liked models
scripts/hf_trends.py -n 5 -s likes -p http://172.28.96.1:10808

# Get most downloaded models
scripts/hf_trends.py -n 10 -s downloads -p http://172.28.96.1:10808
```

### Filter by Task

Filter models by specific AI tasks:

```bash
# Text generation models
scripts/hf_trends.py -n 10 -t text-generation -p http://172.28.96.1:10808

# Image classification models
scripts/hf_trends.py -n 10 -t image-classification -p http://172.28.96.1:10808

# Translation models
scripts/hf_trends.py -n 10 -t translation -p http://172.28.96.1:10808
```

Common task filters:
- `text-generation` - Large language models
- `image-classification` - Vision models
- `image-to-text` - Multimodal models
- `translation` - Machine translation
- `summarization` - Text summarization
- `question-answering` - QA models

### Filter by Library

Filter by ML framework:

```bash
# PyTorch models only
scripts/hf_trends.py -n 10 -l pytorch -p http://172.28.96.1:10808

# TensorFlow models only
scripts/hf_trends.py -n 10 -l tensorflow -p http://172.28.96.1:10808

# JAX models
scripts/hf_trends.py -n 10 -l jax -p http://172.28.96.1:10808
```

### Export to JSON

Save results for further analysis:

```bash
# Export to JSON file
scripts/hf_trends.py -n 10 -j trending_models.json -p http://172.28.96.1:10808

# Export with specific filters
scripts/hf_trends.py -n 20 -t text-generation -j text_models.json -p http://172.28.96.1:10808
```

### Proxy Configuration

The script requires an HTTP proxy to access Hugging Face API (network restrictions).

Use the `-p` flag:

```bash
scripts/hf_trends.py -p http://172.28.96.1:10808
```

For most WSL2 environments with v2rayN:
- Proxy URL: `http://172.28.96.1:10808`
- Or use dynamic IP: `http://$(ip route show | grep default | awk '{print $3}'):10808`

## Command-Line Options

| Flag | Long Form | Description | Default |
|------|-----------|-------------|---------|
| `-n` | `--limit` | Number of models to fetch | 10 |
| `-s` | `--sort` | Sort by: trending, likes, downloads, created | trending |
| `-t` | `--task` | Filter by task/pipeline | None |
| `-l` | `--library` | Filter by library (pytorch, tensorflow, jax) | None |
| `-j` | `--json` | Export results to JSON file | None |
| `-p` | `--proxy` | Proxy URL for HTTP requests | None |

## Output Format

The script displays models in a structured format:

```
🤖 Hugging Face 热门模型 (5 个)
============================================================
1. moonshotai/Kimi-K2.5
   ⭐ 2.0K likes   📥 647.6K downloads
   📊 Task: image-text-to-text   📚 Library: transformers
   📅 Created: 2026-01-01   Updated: N/A
...
```

### Model Information

Each model entry includes:
- **Model ID**: Full Hugging Face model name
- **Likes**: Number of likes (popularity metric)
- **Downloads**: Total download count
- **Task**: Primary task/pipeline (e.g., text-generation)
- **Library**: ML framework (transformers, pytorch, tensorflow)
- **Created/Updated**: Date information

## Use Cases

### Daily Monitoring

Check trending models daily for new releases:

```bash
# Create cron job for daily monitoring
0 9 * * * cd /home/ltx/.openclaw/workspace && \
  /home/ltx/.openclaw/workspace/skills/huggingface-trends/scripts/hf_trends.py \
  -n 20 -p http://172.28.96.1:10808 >> /tmp/hf-trends.log 2>&1
```

### Task-Specific Research

Explore popular models for specific AI tasks:

```bash
# Research trending text generation models
scripts/hf_trends.py -n 15 -t text-generation -s likes -p http://172.28.96.1:10808

# Find popular image-to-text models
scripts/hf_trends.py -n 15 -t image-to-text -s downloads -p http://172.28.96.1:10808
```

### Framework-Specific Analysis

Compare models by ML framework:

```bash
# Compare PyTorch vs TensorFlow popularity
scripts/hf_trends.py -n 20 -l pytorch -j pytorch_models.json -p http://172.28.96.1:10808
scripts/hf_trends.py -n 20 -l tensorflow -j tensorflow_models.json -p http://172.28.96.1:10808
```

## Integration with OpenClaw

Use within OpenClaw sessions:

```python
# Fetch trending models programmatically
from skills.huggingface-trends.scripts import hf_trends

fetcher = hf_trends.HuggingFaceTrends(proxy="http://172.28.96.1:10808")
models = fetcher.fetch_trending_models(limit=10)

# Format for display
output = fetcher.format_models(models)
print(output)
```

## Troubleshooting

### Network Errors

**Problem:** "Network is unreachable" or connection errors

**Solution:** Ensure proxy is specified with `-p` flag:
```bash
scripts/hf_trends.py -p http://172.28.96.1:10808
```

Check if v2rayN proxy is running on Windows.

### Empty Results

**Problem:** "No models found"

**Solution:** Try different filters or increase limit:
```bash
scripts/hf_trends.py -n 50 -p http://172.28.96.1:10808
```

### Dependencies Missing

**Problem:** "requests package not installed"

**Solution:** Install required dependencies:
```bash
pip install requests
```

## Technical Notes

- **API Limitation:** Hugging Face's public API doesn't provide a dedicated trending endpoint without authentication. The script fetches recent models and sorts by popularity metrics.
- **Proxy Requirement:** Due to network restrictions, all requests must go through a proxy. The script supports HTTP proxy configuration.
- **Rate Limits:** The public API has rate limits. Avoid making too many requests in quick succession.
- **Data Freshness:** Models are fetched from the Hugging Face API. Recent changes may take time to reflect.

## Reference

See [Hugging Face API Documentation](https://huggingface.co/docs/huggingface_hub/guides/models) for more details on model metadata and available filters.