Optimum Intel LLMs optimized with IPEX backend

Optimum Intel accelerates Hugging Face pipelines on Intel architectures leveraging Intel Extension for Pytorch, (IPEX) optimizations

Optimum Intel models can be run locally through OptimumIntelLLM entitiy wrapped by LlamaIndex :

In the below line, we install the packages necessary for this demo:

%pip install llama-index-llms-optimum-intel

Now that we’re set up, let’s play around:

If you’re opening this Notebook on colab, you will probably need to install LlamaIndex 🦙.

!pip install llama-index

from llama_index.llms.optimum_intel import OptimumIntelLLM

def messages_to_prompt(messages):
    prompt = ""
    for message in messages:
        if message.role == "system":
            prompt += f"<|system|>\n{message.content}</s>\n"
        elif message.role == "user":
            prompt += f"<|user|>\n{message.content}</s>\n"
        elif message.role == "assistant":
            prompt += f"<|assistant|>\n{message.content}</s>\n"

    # ensure we start with a system prompt, insert blank if needed
    if not prompt.startswith("<|system|>\n"):
        prompt = "<|system|>\n</s>\n" + prompt

    # add final assistant prompt
    prompt = prompt + "<|assistant|>\n"

    return prompt


def completion_to_prompt(completion):
    return f"<|system|>\n</s>\n<|user|>\n{completion}</s>\n<|assistant|>\n"

Model Loading

Models can be loaded by specifying the model parameters using the OptimumIntelLLM method.

oi_llm = OptimumIntelLLM(
    model_name="Intel/neural-chat-7b-v3-3",
    tokenizer_name="Intel/neural-chat-7b-v3-3",
    context_window=3900,
    max_new_tokens=256,
    generate_kwargs={"temperature": 0.7, "top_k": 50, "top_p": 0.95},
    messages_to_prompt=messages_to_prompt,
    completion_to_prompt=completion_to_prompt,
    device_map="cpu",
)

response = oi_llm.complete("What is the meaning of life?")
print(str(response))

Streaming

Using stream_complete endpoint

response = oi_llm.stream_complete("Who is Mother Teresa?")
for r in response:
    print(r.delta, end="")

Using stream_chat endpoint

from llama_index.core.llms import ChatMessage

messages = [
    ChatMessage(
        role="system",
        content="You are an American chef in a small restaurant in New Orleans",
    ),
    ChatMessage(role="user", content="What is your dish of the day?"),
]
resp = oi_llm.stream_chat(messages)

for r in resp:
    print(r.delta, end="")