Add vLLMEmbeddings

[plaguss]

Description

Adds a new embeddings provider:

from distilabel.embeddings import vLLMEmbeddings

embeddings = vLLMEmbeddings(model="intfloat/e5-mistral-7b-instruct")

embeddings.load()

results = embeddings.encode(inputs=["distilabel is awesome!", "and Argilla!"])
# [
#   [-0.05447685346007347, -0.01623094454407692, ...],
#   [4.4889533455716446e-05, 0.044016145169734955, ...],
# ]

Tested on the cluster and works nicely outside of a Pipeline, on a Pipeline with ray fails, not sure if that's due to ray or the CudaDevicePlacement. Still need to investigate that, but this should be ready

[github-actions]

Documentation for this PR has been built. You can view it at: https://distilabel.argilla.io/pr-920/

[codspeed-hq]

CodSpeed Performance Report

Merging #920 will not alter performance

_{Comparing vllm-embeddings (98a337a) with develop (516909e)}

Summary

✅ 1 untouched benchmarks

[gabrielmbmb]

LGTM!