Skip to content
Search models or paste a model link. Press Enter to search.

ViTamin Large2 (224, CLIP)

timm/vitamin_large2_224.datacomp1b_clip
Left and right arrow keys switch the vendor.
Embedl-optimized
timm vitamin_large2_224.datacomp1b_clip architecture graph
timm vitamin_large2_224.datacomp1b_clip architecture graph
  • timm/vitamin_large2_224.datacomp1b_clip
  • INT8
  • L4
Labels
  • Image
Build
Embedl-optimized
Precision
INT8
Runtime
TensorRT 11.3.0.99 with Embedl optimizations
Input size
224×224
Run ViTamin Large2 (224, CLIP) on L4

vitamin_large2_224.datacomp1b_clip is an image encoder from timm: it turns an image into a 1,024-dimensional embedding, for search, retrieval and as features for other models.

Methodology

  • Latency: trtexec on the NVIDIA L4, batch 1, 201 iterations after warm-up, during the build: GPU compute time per inference, mean and p99, and throughput.
  • Baseline TensorRT FP16: the same Embedl-optimized model before quantization, converted to FP16 with ModelOpt AutoCast (float32 inputs and outputs), compiled with TensorRT 11.3.0.99 and timed with the same trtexec settings on an NVIDIA L4.
  • Quantization: Embedl Deploy, activations INT8, weights INT8 per-channel; calibrated on Imagenette (1,000 images). Exported as ONNX opset 20, compiled with TensorRT 11.3.0.99, built on NVIDIA L4 (sm_89).
  • Build: config vitamin_large2_224/datacomp1b_clip-timm_int8_embedl_tensorrt_l4, on an NVIDIA L4.

Files

Image Size

Model

docker.io/embedl/model-artifacts:vitamin_large2_224.datacomp1b_clip-timm_int8_embedl_tensorrt_l4.5c1a3ff9436a6073

3.74 GiB

Runtime

docker.io/embedl/models:runtimes.tensorrt.amd64.2026102005300fa3

0.52 GiB