The request names a specific tiny test model on Hugging Face, so the best fit is the Hugging Face hub 6, which hosts models and offers an image-classification task and a hub client library. For running ResNet image classification directly, Replicate 3 exposes a ResNet-50 classifier through its API, and Hugging Face's own image-classification tooling 6 covers the same task. Other sources in the list (NVIDIA 1, OpenAI 4, fast.ai 7) are only loosely related and are not recommended for t

1. 1

   Hugging Facehuggingface.coverified and signed

   Hosts ML models such as tiny-random-ResNetForImageClassification and provides image-classification tasks and a hub client (huggingface-hub SDK) to load and run them 6.

   WebsiteDocsPricingGet started

   official:MCP server · @huggingface/mcp-clientSDK · huggingface-hubCLI · text-embeddings-inferencerepository · pwc-cliskill · hf-cli+23 community tools built for it

   for agents: image-classification · image_classification · huggingface-vision-trainer

2. 2

   Replicatereplicate.comverified and signed

   Runs open-source models through a cloud API, including a ResNet-50 image classification model (replicate/resnet) 3.

   WebsiteDocsPricingGet started

   +14 community tools built for it

   for agents: replicate/resnet · black-forest-labs/flux-redux-schnell · replicate/dreambooth

3. 3

   NVIDIAnvidia.comverified and signed

   Offers a TAO PyTorch-based image classification action supporting ResNet backbones, plus TensorRT for optimized inference 1.

   WebsiteDocsPricingGet started

   official:MCP server · NVIDIA CUDA Docs+84 community tools built for it

   for agents: tao-train-image-classification · tensorrt · computer-vision