Redirect vLLM runtime guide to Hugging Face runtime overview (#408)

* Redirect vLLM runtime guide to Hugging Face runtime overview

Signed-off-by: Yuan Tang <terrytangyuan@gmail.com>

* Update README.md

Signed-off-by: Yuan Tang <terrytangyuan@gmail.com>

---------

Signed-off-by: Yuan Tang <terrytangyuan@gmail.com>

docs/modelserving/v1beta1/llm/vllm/README.md

@@ -1,83 +1,3 @@
-## Deploy the LLaMA model with vLLM Runtime
-Serving LLM models can be surprisingly slow even on high end GPUs, [vLLM](https://github.com/vllm-project/vllm) is a fast and easy-to-use LLM inference engine. It can achieve 10x-20x higher throughput than Huggingface transformers.
-It supports [continuous batching](https://www.anyscale.com/blog/continuous-batching-llm-inference) for increased throughput and GPU utilization,
-[paged attention](https://vllm.ai) to address the memory bottleneck where in the autoregressive decoding process all the attention key value tensors(KV Cache) are kept in the GPU memory to generate next tokens.
+## vLLM Runtime
 
-You can deploy the LLaMA model with built vLLM inference server container image using the `InferenceService` yaml API spec. 
-We have work in progress integrating `vLLM` with `Open Inference Protocol` and KServe observability stack.
-
-The LLaMA model can be downloaded from [huggingface](https://huggingface.co/meta-llama/Llama-2-7b) and upload to your cloud storage.
-
-=== "Yaml"
-
-
-    ```yaml
-    kubectl apply -n kserve-test -f - <<EOF
-    apiVersion: serving.kserve.io/v1beta1
-    kind: InferenceService
-    metadata:
-      name: llama-2-7b
-    spec:
-      predictor:
-        containers:
-        - args:
-            - --port
-            - "8080"
-            - --model
-            - /mnt/models
-          command:
-            - python3
-            - -m
-            - vllm.entrypoints.openai.api_server
-          env:
-            - name: STORAGE_URI
-              value: gs://kfserving-examples/llm/huggingface/llama
-          image: kserve/vllmserver:latest
-          name: kserve-container
-          resources:
-            limits:
-              cpu: "4"
-              memory: 50Gi
-              nvidia.com/gpu: "1"
-            requests:
-              cpu: "1"
-              memory: 50Gi
-              nvidia.com/gpu: "1"
-    ```
-
-!!! Warning
-    vLLM runtime is still experimental, please expect API changes and further integration in the next KServe release.
-
-=== "kubectl"
-```bash
-kubectl apply -f ./vllm.yaml
-```
-
-## Benchmarking vLLM Runtime
-
-You can download the benchmark testing data set by running
-```bash
-wget https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/resolve/main/ShareGPT_V3_unfiltered_cleaned_split.json
-```
-
-The tokenizer can be found from the downloaded llama model.
-
-Now, assuming that your ingress can be accessed at
-`${INGRESS_HOST}:${INGRESS_PORT}` or you can follow [this instruction](../../../../get_started/first_isvc.md#4-determine-the-ingress-ip-and-ports)
-to find out your ingress IP and port.
-
-You can run the [benchmarking script](./benchmark.py) and send the inference request to the exposed URL.
-
-```bash
-python benchmark_serving.py --backend openai --port ${INGRESS_PORT} --host ${INGRESS_HOST} --dataset ./ShareGPT_V3_unfiltered_cleaned_split.json --tokenizer ./tokenizer --request-rate 5
-```
-
-!!! success "Expected Output"
-
-    ```{ .json .no-copy }
-       Total time: 216.81 s
-       Throughput: 4.61 requests/s
-       Average latency: 7.96 s
-       Average latency per token: 0.02 s
-       Average latency per output token: 0.04 s
-    ```
+The official vLLM support is available through [Hugging Face Serving Runtime](https://kserve.github.io/website/master/modelserving/v1beta1/llm/huggingface/).