Using Language Models (LLMs)

Learn how to leverage Synexa AI’s powerful language models for text generation and processing.

Basic Usage

Our API provides easy access to state-of-the-art language models for various text generation tasks.

Python Example

import synexa
 
# Generate text using an LLM
output = synexa.run(
    "meta/meta-llama-3-8b-instruct",
    input={
        "prompt": "What is the meaning of life?"
    }
)
 
# Get the generated text
response = output[0]["text"]
print(response)

Node.js Example

import Synexa from 'synexa';
 
const synexa = new Synexa.default({
  auth: process.env.SYNEXA_API_TOKEN
});
 
// Generate text using an LLM
const [output] = await synexa.run("meta/meta-llama-3-8b-instruct", {
  input: {
    prompt: "What is the meaning of life?"
  }
});
 
// Get the generated text
console.log(output.text);

Advanced Options

You can customize the text generation with additional parameters:

# Python example with advanced options
output = synexa.run(
    "meta/meta-llama-3-8b-instruct",
    input={
        "top_k": 0,
        "top_p": 0.95,
        "prompt": "Johnny has 8 billion parameters. His friend Tommy has 70 billion parameters. What does this mean when it comes to speed?",
        "max_tokens": 512,
        "temperature": 0.7,
        "system_prompt": "You are a helpful assistant",
        "length_penalty": 1,
        "max_new_tokens": 512,
        "stop_sequences": "<|end_of_text|>,<|eot_id|>",
        "prompt_template": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\n{system_prompt}<|eot_id|><|start_header_id|>user<|end_header_id|>\n\n{prompt}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
        "presence_penalty": 0,
        "log_performance_metrics": False
    }
)

// Node.js example with advanced options
const [output] = await synexa.run("meta/meta-llama-3-8b-instruct", {
  input: {
    top_k: 0,
    top_p: 0.95,
    prompt: "Johnny has 8 billion parameters. His friend Tommy has 70 billion parameters. What does this mean when it comes to speed?",
    max_tokens: 512,
    temperature: 0.7,
    system_prompt: "You are a helpful assistant",
    length_penalty: 1,
    max_new_tokens: 512,
    stop_sequences: "<|end_of_text|>,<|eot_id|>",
    prompt_template: "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\n{system_prompt}<|eot_id|><|start_header_id|>user<|end_header_id|>\n\n{prompt}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
    presence_penalty: 0,
    log_performance_metrics: false
  }
});

System Messages

You can use system messages to control the model’s behavior:

# Python example with system message
output = synexa.run(
    "meta/meta-llama-3-8b-instruct",
    input={
        "system": "You are a helpful AI assistant that speaks like Shakespeare",
        "prompt": "Tell me about artificial intelligence",
        "max_tokens": 200
    }
)

// Node.js example with system message
const [output] = await synexa.run("meta/meta-llama-3-8b-instruct", {
  input: {
    system: "You are a helpful AI assistant that speaks like Shakespeare",
    prompt: "Tell me about artificial intelligence",
    max_tokens: 200
  }
});

Error Handling

Always handle potential errors when using language models:

# Python error handling
try:
    output = synexa.run(
        "meta/meta-llama-3-8b-instruct",
        input={"prompt": "Tell me a joke"}
    )
except Exception as e:
    print(f"Error generating text: {e}")

// Node.js error handling
try {
  const [output] = await synexa.run("meta/meta-llama-3-8b-instruct", {
    input: { prompt: "Tell me a joke" }
  });
} catch (error) {
  console.error("Error generating text:", error);
}

Speech to Text