Served on MixpeekEmbeddingVerified 2026-11-05
OpenCLIP ViT-L/14 LAION-2B s32B b82K
laion/CLIP-ViT-L-14-laion2B-s32B-b82K
OpenCLIP ViT-L/14 LAION-2B s32B b82K is a image embedding model (MIT). Served on Mixpeek, scaling to zero when idle.
A Ray Serve app in Mixpeek's inference registry hosts it. It starts at zero replicas and wakes on first use.
New opens the Designer in Studio. For an existing namespace, open it in Studio and use its Designer tab.
No demo for OpenCLIP ViT-L/14 LAION-2B s32B b82K 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
- Image-to-image similarity from a large web-trained ViT-L/14 image tower.
- Worst at
- Text-to-image search: the Serve app loads only CLIPVisionModel and returns the raw CLS hidden state, which is not in CLIP's shared text-image space. No extractor writes this vector.
- Task
- Embedding
- Input
- Image
- Output
- Vector
- Use cases
- Visual search, Product search
Templates that use it
No template uses this model yet. Browse templates.
Input and output
Input
- image
- resolution
- 224x224
- color
- RGB
- normalize
- CLIP mean/std (CLIPImageProcessor)
Output
- vector
- dims
- 768
- distance
- cosine
- type
- dense
How Mixpeek wires it
- Extractors
- None today
- Feature URIs
- None today
- Vector indexes
- None
- Retriever stages
- None
- Combines with
- Realtime image embedding through POST /v1/inference only; laion_clip_vit_l_14_v1 is a named inference in the vector models but no extractor declares an index for it
How it runs
- Status
- Served on Mixpeek Ray Serve app in the inference registry, image modality only; no extractor uses it.
- Serve apps
laion__CLIP_ViT_L_14_laion2B_s32B_b82K- Runtime
- transformers CLIPVisionModel + CLIPImageProcessor
- Weights
- safetensors, no remote code
- Parameters
- 428 M
- Size on disk
- 1,711 MB
- Preprocessing
- CLIPImageProcessor (resize, center crop 224, CLIP normalization).
- Postprocessing
- last_hidden_state[:, 0] (CLS before post_layernorm and before the visual projection), L2-normalized. The checkpoint config gives vision hidden_size 1024 and projection_dim 768, so this vector is 1024-d while the registry declares 768.
- Serving
- both
- Hardware
- Runs on CPU (slower without a GPU)
Scale to zero
- Minimum replicas
- 0
- Maximum replicas
- 1
- Idle timeout
- 600 s
- Memory
- Not recorded
- CPUs
- 0.5
- GPUs
- Set at deploy time: hardware_config.gpu_fraction if hardware_config.gpu_concurrency > 0 else 0
- Cold start, node warm
- Not recorded
- Cold start, new node
- Not recorded
Measured cost
Unmeasured
No committed throughput or cost measurement for this model.
Our evaluations
Unmeasured
We have not evaluated this model on our own data yet. Publisher results are in the sources below.
Questions
- What is OpenCLIP ViT-L/14 LAION-2B s32B b82K best at?
- Image-to-image similarity from a large web-trained ViT-L/14 image tower.
- What are the limits of OpenCLIP ViT-L/14 LAION-2B s32B b82K?
- Text-to-image search: the Serve app loads only CLIPVisionModel and returns the raw CLS hidden state, which is not in CLIP's shared text-image space. No extractor writes this vector.
- Can OpenCLIP ViT-L/14 LAION-2B s32B b82K be used commercially?
- Yes. MIT on the Hugging Face repo. The model card calls the release research-oriented and LAION-2B is uncurated web data; no license restriction on use.
- How does OpenCLIP ViT-L/14 LAION-2B s32B b82K run on Mixpeek?
- Ray Serve app in the inference registry, image modality only; no extractor uses it.