[Engine] Apply the STS task to bge models

[Zhenzhong1]

Type of Change

• Example add
• API extend
• Doc update

Description

• Add the new directory for the text-embedding task and Massive Text Embedding Benchmark (mteb).
• Verifyed the BAAI/bge-small-en-v1.5 accuracy with the Neural Engine backend.
• Verifyed the BAAI/bge-base-en-v1.5 accuracy with the Neural Engine backend.

Accuracy
int8 bge-base with the Neural Engine on the STS: 82.191

int8 bge-small with the Neural Engine on the STS: 81.43

Usage:

from transformers import AutoTokenizer
from intel_extension_for_transformers.transformers import AutoModel

sentences_batch = ['sentence-1', 'sentence-2', 'sentence-3', 'sentence-4']

tokenizer = AutoTokenizer.from_pretrained('BAAI/bge-base-en-v1.5')
encoded_input = tokenizer(sentences_batch,
                            padding=True,
                            truncation=True,
                            max_length=512,
                            return_tensors="np")

engine_input = [encoded_input['input_ids'], encoded_input['token_type_ids'], encoded_input['attention_mask']]

model = AutoModel.from_pretrained('./model_and_tokenizer/int8-model.onnx', use_embedding_runtime=True)
sentence_embeddings = model.generate(engine_input)['last_hidden_state:0']

print("Sentence embeddings:", sentence_embeddings)

Expected Behavior & Potential Risk

N/A

How has this PR been tested?

bge benchmark

Dependency Change?

remove C_MTEB

[a32543254]

LGTM

[VincyZhang]

Will test LLM deprecated python mode in release tests.