Fix local models being treated as S3 paths

[RobertLucian]

@vishalbollu when the model's path is not an S3 path, the model cacher and validator still treats it as an S3 path instead of just a local path. This prevents the deployment of APIs with locally-provided models (ONNX and TensorFlow).

How to reproduce:

# cortex.yaml
- name: api-tf-cd7
  predictor:
    type: tensorflow
    path: predictor.py 
    model: ./inferentia-models/tensorflow/resnet50

With the above config, the resulting error would be
error: /home/robert/sandbox/feature/multi-model-endpoint/cortex_tf_sample.yaml: api-tf: predictor: model: "./inferentia-models/tensorflow/resnet50" is not a valid s3 path (e.g. s3://cortex-examples/iris-classifier/tensorflow is a valid s3 path)

---

checklist:

• run make test and make lint
• test manually (i.e. build/push all images, restart operator, and re-deploy APIs)