Groq¶

欢迎来到 Groq！🚀 在 Groq，我们研发了全球首款语言处理单元™（LPU）。Groq LPU 采用确定性单核流式架构，为生成式 AI 推理速度树立了新标杆，在任何工作负载下都能提供可预测且可重复的性能表现。

除了硬件架构，我们的软件同样致力于为开发者赋能，提供打造创新强大 AI 应用所需的工具。选择 Groq 作为引擎，您将能够：

• 为实时 AI 和高性能计算推理实现无可妥协的低延迟与高性能 🔥
• 精准掌握任意工作负载下的确切性能表现与计算时间 🔮
• 运用尖端技术保持竞争优势 💪

想了解更多？访问我们的官网获取更多资源，并加入我们的Discord 社区与开发者们交流！

安装¶

如果您在 Colab 上打开此 Notebook，可能需要安装 LlamaIndex 🦙。

In [ ]:

% pip install llama-index-llms-groq

% pip install llama-index-llms-groq

In [ ]:

!pip install llama-index

!pip install llama-index

In [ ]:

from llama_index.llms.groq import Groq

from llama_index.llms.groq import Groq

None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.

在 Groq 控制台 创建 API 密钥，然后将其设置为环境变量 GROQ_API_KEY。

export GROQ_API_KEY=<your api key>

或者，您可以在初始化 LLM 时直接传入 API 密钥：

In [ ]:

llm = Groq(model="llama3-70b-8192", api_key="your_api_key")

llm = Groq(model="llama3-70b-8192", api_key="your_api_key")

可用的 LLM 模型列表详见此处。

In [ ]:

response = llm.complete("Explain the importance of low latency LLMs")

response = llm.complete("Explain the importance of low latency LLMs")

In [ ]:

print(response)

print(response)

Low latency Large Language Models (LLMs) are important in certain applications due to their ability to process and respond to inputs quickly. Latency refers to the time delay between a user's request and the system's response. In some real-time or time-sensitive applications, low latency is critical to ensure a smooth user experience and prevent delays or lag.

For example, in conversational agents or chatbots, users expect quick and responsive interactions. If the system takes too long to process and respond to user inputs, it can negatively impact the user experience and lead to frustration. Similarly, in applications such as real-time language translation or speech recognition, low latency is essential to provide accurate and timely feedback to the user.

Furthermore, low latency LLMs can enable new use cases and applications that require real-time or near real-time processing of language inputs. For instance, in the field of autonomous vehicles, low latency LLMs can be used for real-time speech recognition and natural language understanding, enabling voice-controlled interfaces that allow drivers to keep their hands on the wheel and eyes on the road.

In summary, low latency LLMs are important for providing a smooth and responsive user experience, enabling real-time or near real-time processing of language inputs, and unlocking new use cases and applications that require real-time or near real-time processing of language inputs.

使用消息列表调用 chat 方法¶

In [ ]:

from llama_index.core.llms import ChatMessage

messages = [
    ChatMessage(
        role="system", content="You are a pirate with a colorful personality"
    ),
    ChatMessage(role="user", content="What is your name"),
]
resp = llm.chat(messages)

from llama_index.core.llms import ChatMessage messages = [ ChatMessage( role="system", content="You are a pirate with a colorful personality" ), ChatMessage(role="user", content="What is your name"), ] resp = llm.chat(messages)

In [ ]:

print(resp)

print(resp)

assistant: Arr, I be known as Captain Redbeard, the fiercest pirate on the seven seas! But ye can call me Cap'n Redbeard for short. I'm a fearsome pirate with a love for treasure and adventure, and I'm always ready for a good time! Whether I'm swabbin' the deck or swiggin' grog, I'm always up for a bit of fun. So hoist the Jolly Roger and let's set sail for adventure, me hearties!

流式传输¶

使用 stream_complete 端点

In [ ]:

response = llm.stream_complete("Explain the importance of low latency LLMs")

response = llm.stream_complete("Explain the importance of low latency LLMs")

In [ ]:

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

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

Low latency Large Language Models (LLMs) are important in the field of artificial intelligence and natural language processing (NLP) due to several reasons:

1. Real-time applications: Low latency LLMs are essential for real-time applications such as chatbots, voice assistants, and real-time translation services. These applications require immediate responses, and high latency can result in a poor user experience.
2. Improved user experience: Low latency LLMs can provide a more seamless and responsive user experience. Users are more likely to continue using a service that provides quick and accurate responses, leading to higher user engagement and satisfaction.
3. Better decision-making: In some applications, such as financial trading or autonomous vehicles, low latency LLMs can provide critical information in real-time, enabling better decision-making and reducing the risk of accidents.
4. Scalability: Low latency LLMs can handle a higher volume of requests, making them more scalable and suitable for large-scale applications.
5. Competitive advantage: Low latency LLMs can provide a competitive advantage in industries where real-time decision-making and responsiveness are critical. For example, in online gaming or e-commerce, low latency LLMs can provide a more immersive and engaging user experience, leading to higher customer loyalty and revenue.

In summary, low latency LLMs are essential for real-time applications, providing a better user experience, enabling better decision-making, improving scalability, and providing a competitive advantage. As LLMs continue to play an increasingly important role in various industries, low latency will become even more critical for their success.

使用 stream_chat 端点

In [ ]:

from llama_index.core.llms import ChatMessage

messages = [
    ChatMessage(
        role="system", content="You are a pirate with a colorful personality"
    ),
    ChatMessage(role="user", content="What is your name"),
]
resp = llm.stream_chat(messages)

from llama_index.core.llms import ChatMessage messages = [ ChatMessage( role="system", content="You are a pirate with a colorful personality" ), ChatMessage(role="user", content="What is your name"), ] resp = llm.stream_chat(messages)

In [ ]:

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

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

Arr, I be known as Captain Candybeard! A more colorful and swashbuckling pirate, ye will never find!