lance-format/openvid-lance

---
license: cc-by-4.0
task_categories:
- text-to-video
- video-classification
language:
- en
tags:
- text-to-video
- video-search
pretty_name: openvid-lance
size_categories:
- 100K<n<1M
---


# OpenVid Dataset (Lance Format)

Lance format version of the [OpenVid dataset](https://huggingface.co/datasets/nkp37/OpenVid-1M) with **937,957 high-quality videos** stored with inline video blobs, embeddings, and rich metadata.

## Why Lance?

Lance is an open-source format designed for multimodal AI data, offering significant advantages over traditional formats for modern AI workloads.

- **Blazing Fast Random Access**: Optimized for fetching scattered rows, making it ideal for random sampling, real-time ML serving, and interactive applications without performance degradation.
- **Native Multimodal Support**: Store text, embeddings, and other data types together in a single file. Large binary objects are loaded lazily, and vectors are optimized for fast similarity search.
- **Efficient Data Evolution**: Add new columns and backfill data without rewriting the entire dataset. This is perfect for evolving ML features, adding new embeddings, or introducing moderation tags over time.
- **Versatile Querying**: Supports combining vector similarity search, full-text search, and SQL-style filtering in a single query, accelerated by on-disk indexes.

## Key Features

The OpenVid dataset is stored in Lance format with inline video blobs, video embeddings, and rich metadata.

- **Videos stored inline as blobs**: No external files to manage
- **Efficient column access**: Load metadata without touching video data
- **Prebuilt indices available**: IVF_PQ index for similarity search, FTS index on captions
- **Fast random access**: Read any video instantly by index
- **HuggingFace integration**: Load directly from the Hub

## Quick Start

### Load with `datasets.load_dataset`

```python
import datasets

hf_ds = datasets.load_dataset(
"lance-format/openvid-lance",
split="train",
streaming=True,
)
# Take first three rows and print captions
for row in hf_ds.take(3):
print(row["caption"])
```

### Load with Lance

Use Lance for ANN search, retrieving specific blob bytes or advanced indexing, while still pointing at the dataset on the Hub:

```python
import lance

lance_ds = lance.dataset("hf://datasets/lance-format/openvid-lance/data/train.lance")
blob_file = lance_ds.take_blobs("video_blob", ids=[0])[0]
video_bytes = blob_file.read()
```

### Load with LanceDB

These tables can also be consumed by [LanceDB](https://docs.lancedb.com/), the multimodal lakehouse for AI (built on top of Lance).
LanceDB provides several convenience APIs for search, index creation and data updates on top of the Lance format.

```python
import lancedb

db = lancedb.connect("hf://datasets/lance-format/openvid-lance/data")
tbl = db.open_table("train")
print(f"LanceDB table opened with {len(tbl)} videos")
```

## Blob API

Lance stores videos as **inline blobs** - binary data embedded directly in the dataset. This provides:

- **Single source of truth** - Videos and metadata together in one dataset
- **Lazy loading** - Videos only loaded when you explicitly request them
- **Efficient storage** - Optimized encoding for large binary data

```python
import lance

ds = lance.dataset("hf://datasets/lance-format/openvid-lance")

# 1. Browse metadata without loading video data
metadata = ds.scanner(
columns=["caption", "aesthetic_score"], # No video_blob column!
filter="aesthetic_score >= 4.5",
limit=10
).to_table().to_pylist()

# 2. User selects video to watch
selected_index = 3

# 3. Load only that video blob
blob_file = ds.take_blobs("video_blob", ids=[selected_index])[0]
video_bytes = blob_file.read()

# 4. Save to disk
with open("video.mp4", "wb") as f:
f.write(video_bytes)
```

> **⚠️ HuggingFace Streaming Note**
>
> When streaming from HuggingFace (as shown above), some operations use minimal parameters to avoid rate limits:
> - `nprobes=1` for vector search (lowest value)
> - Column selection to reduce I/O
>
> **You may still hit rate limits on HuggingFace's free tier.** For best performance and to avoid rate limits, **download the dataset locally**:
>
> ```bash
> # Download once
> huggingface-cli download lance-format/openvid-lance --repo-type dataset --local-dir ./openvid
>
> # Then load locally
> ds = lance.dataset("./openvid")
> ```
>
> Streaming is recommended only for quick exploration and testing.

## Usage Examples

### 1. Browse Metadata quickly (fast, no video loading)

```python
# Load only metadata without heavy video blobs
scanner = ds.scanner(
columns=["caption", "aesthetic_score", "motion_score"],
limit=10
)
videos = scanner.to_table().to_pylist()

for video in videos:
print(f"{video['caption']} - Quality: {video['aesthetic_score']:.2f}")
```

### 2. Export videos from blobs

Retrieve specific video files if you want to work with subsets of the data. This is done by exporting them to files on your local machine.

```python
# Load specific videos by index
indices = [0, 100, 500]
blob_files = ds.take_blobs("video_blob", ids=indices)

# Save to disk
for i, blob_file in enumerate(blob_files):
with open(f"video_{i}.mp4", "wb") as f:
f.write(blob_file.read())
```

### 3. Open inline videos with PyAV and run seeks directly on the blob file

Using seeks, you can open a specific set of frames within a blob. The example below shows this.

```python
import av

selected_index = 123
blob_file = ds.take_blobs("video_blob", ids=[selected_index])[0]

with av.open(blob_file) as container:
stream = container.streams.video[0]

for seconds in (0.0, 1.0, 2.5):
target_pts = int(seconds / stream.time_base)
container.seek(target_pts, stream=stream)

frame = None
for candidate in container.decode(stream):
if candidate.time is None:
continue
frame = candidate
if frame.time >= seconds:
break

print(
f"Seek {seconds:.1f}s -> {frame.width}x{frame.height} "
f"(pts={frame.pts}, time={frame.time:.2f}s)"
)
```

### 4. Inspecting Existing Indices

You can inspect the prebuilt indices on the dataset:

```python
import lance

# Open the dataset
dataset = lance.dataset("hf://datasets/lance-format/openvid-lance/data/train.lance")

# List all indices
indices = dataset.list_indices()
print(indices)
```

### 5. Create New Index

While this dataset comes with pre-built indices, you can also create your own custom indices if needed.
The example below creates a vector index on the `embedding` column.

```python
# ds is a local Lance dataset
ds.create_index(
"embedding",
index_type="IVF_PQ",
num_partitions=256,
num_sub_vectors=96,
replace=True,
)
```

### 6. Vector Similarity Search

```python
import pyarrow as pa

# Find similar videos
ref_video = ds.take([0], columns=["embedding"]).to_pylist()[0]
query_vector = pa.array([ref_video['embedding']], type=pa.list_(pa.float32(), 1024))

results = ds.scanner(
nearest={
"column": "embedding",
"q": query_vector[0],
"k": 5,
"nprobes": 1,
"refine_factor": 1
}
).to_table().to_pylist()

for video in results[1:]: # Skip first (query itself)
print(video['caption'])
```

### 7. Full-Text Search

```python
# Search captions using FTS index
results = ds.scanner(
full_text_query="sunset beach",
columns=["caption", "aesthetic_score"],
limit=10,
fast_search=True
).to_table().to_pylist()

for video in results:
print(f"{video['caption']} - {video['aesthetic_score']:.2f}")
```

## Dataset Evolution

Lance supports flexible schema and data evolution ([docs](https://lance.org/guide/data_evolution/?h=evol)). You can add/drop columns, backfill with SQL or Python, rename fields, or change data types without rewriting the whole dataset. In practice this lets you:
- Introduce fresh metadata (moderation labels, embeddings, quality scores) as new signals become available.
- Add new columns to existing datasets without re-exporting terabytes of video.
- Adjust column names or shrink storage (e.g., cast embeddings to float16) while keeping previous snapshots queryable for reproducibility.

```python
import lance
import pyarrow as pa
import numpy as np

base = pa.table({"id": pa.array([1, 2, 3])})
dataset = lance.write_dataset(base, "openvid_evolution", mode="overwrite")

# 1. Grow the schema instantly (metadata-only)
dataset.add_columns(pa.field("quality_bucket", pa.string()))

# 2. Backfill with SQL expressions or constants
dataset.add_columns({"status": "'active'"})

# 3. Generate rich columns via Python batch UDFs
@lance.batch_udf()
def random_embedding(batch):
arr = np.random.rand(batch.num_rows, 128).astype("float32")
return pa.RecordBatch.from_arrays(
[pa.FixedSizeListArray.from_arrays(arr.ravel(), 128)],
names=["embedding"],
)

dataset.add_columns(random_embedding)

# 4. Bring in offline annotations with merge
labels = pa.table({
"id": pa.array([1, 2, 3]),
"label": pa.array(["horse", "rabbit", "cat"]),
})
dataset.merge(labels, "id")

# 5. Rename or cast columns as needs change
dataset.alter_columns({"path": "quality_bucket", "name": "quality_tier"})
dataset.alter_columns({"path": "embedding", "data_type": pa.list_(pa.float16(), 128)})
```

These operations are automatically versioned, so prior experiments can still point to earlier versions while OpenVid keeps evolving.

## LanceDB

LanceDB users can follow the following examples to run search queries on the dataset.

### LanceDB Vector Similarity Search

```python
import lancedb

db = lancedb.connect("hf://datasets/lance-format/openvid-lance/data")
tbl = db.open_table("train")

# Get a video to use as a query
ref_video = tbl.limit(1).select(["embedding", "caption"]).to_pandas().to_dict('records')[0]
query_embedding = ref_video["embedding"]

results = tbl.search(query_embedding, vector_column_name="embedding") \
.metric("L2") \
.nprobes(1) \
.limit(5) \
.to_list()

for video in results[1:]: # Skip first (query itself)
print(f"{video['caption'][:60]}...")
```

### LanceDB Full-Text Search

```python
import lancedb

db = lancedb.connect("hf://datasets/lance-format/openvid-lance/data")
tbl = db.open_table("train")

results = tbl.search("sunset beach") \
.select(["caption", "aesthetic_score"]) \
.limit(10) \
.to_list()

for video in results:
print(f"{video['caption']} - {video['aesthetic_score']:.2f}")
```

## Citation

@article{nan2024openvid,
title={OpenVid-1M: A Large-Scale High-Quality Dataset for Text-to-video Generation},
author={Nan, Kepan and Xie, Rui and Zhou, Penghao and Fan, Tiehan and Yang, Zhenheng and Chen, Zhijie and Li, Xiang and Yang, Jian and Tai, Ying},
journal={arXiv preprint arXiv:2407.02371},
year={2024}
}

## License

Please check the original OpenVid dataset license for usage terms.