Runs as a custom pluginEmbeddingVerified 2026-11-05
X-CLIP base-patch32
microsoft/xclip-base-patch32
X-CLIP base-patch32 is a video and text embedding model (MIT). Can run on Mixpeek as a custom plugin.
Mixpeek does not run it today. It can run as a custom plugin: the weights load without remote code.
No demo for X-CLIP base-patch32 yet
A video, image, slideshow or playground goes here once we record one on our own data.
What it is good and bad at
- Best at
- Cheap video-text matching on short clips with a CLIP backbone; zero-shot action labels (Kinetics-trained).
- Worst at
- Trained on Kinetics-400 action clips: weak on general scenes, logos, people and speech; 8 frames per clip.
- Task
- Embedding
- Input
- Video, Text
- Output
- Vector
- Use cases
- Video understanding, Action recognition
Templates that use it
No template uses this model yet. Browse templates.
Input and output
Input
- video
- frames
- 8
- resolution
- 224x224
- text
- max_tokens
- 77
- prompts
- action phrases
Output
- vector
- dims
- 512
- distance
- cosine
- type
- dense
How Mixpeek wires it
- Extractors
- None today
- Feature URIs
- None today
- Vector indexes
- None
- Retriever stages
- feature_search
- Combines with
- scene splits (transcoding/multimodal extractor) -> clip-level vectors -> feature_search with a text query
How it runs
- Status
- Runs as a custom plugin Not served. Safetensors with transformers XCLIPModel (no remote code); would run as a batch plugin over scene clips, with realtime text query embedding.
- Serve apps
- None
- Runtime
- transformers XCLIPModel + XCLIPProcessor
- Weights
- safetensors, no remote code
- Parameters
- 197 M
- Size on disk
- 786 MB
- Preprocessing
- Sample 8 frames uniformly, resize and center-crop to 224, CLIP mean/std.
- Postprocessing
- L2-normalize video and text embeddings.
- Serving
- both
- Hardware
- Runs on CPU (slower without a GPU)
Scale to zero
- Minimum replicas
- 0
- Maximum replicas
- 1
- Idle timeout
- 600 s
- Memory
- 2 GB
- CPUs
- 2
- GPUs
- 0
- Cold start, node warm
- Not recorded
- Cold start, new node
- Not recorded
Measured cost
Unmeasured
No committed measurement; plan row 3 measures it in the standalone image.
Our evaluations
Unmeasured
We have not evaluated this model on our own data yet. Publisher results are in the sources below.
Questions
- What is X-CLIP base-patch32 best at?
- Cheap video-text matching on short clips with a CLIP backbone; zero-shot action labels (Kinetics-trained).
- What are the limits of X-CLIP base-patch32?
- Trained on Kinetics-400 action clips: weak on general scenes, logos, people and speech; 8 frames per clip.
- Can X-CLIP base-patch32 be used commercially?
- Yes. MIT per the hub tag and repo.
- How does X-CLIP base-patch32 run on Mixpeek?
- Not served. Safetensors with transformers XCLIPModel (no remote code); would run as a batch plugin over scene clips, with realtime text query embedding.
- What does X-CLIP base-patch32 output?
- A 512-dimensional dense vector, compared by cosine similarity.