Property MaxCompletionTokens
MaxCompletionTokens
Gets or sets the maximum number of tokens for completion per turn.
public int? MaxCompletionTokens { get; set; }
Property Value
- int?
Examples
Controlling response length:
using LMKit.Model;
using LMKit.Agents;
using var model = new LM("path/to/model.gguf");
var agent = new Agent(model);
// Short summary
var shortOptions = new AgentExecutionOptions
{
MaxCompletionTokens = 100
};
var summary = agent.Run("Summarize machine learning in brief.", shortOptions);
// Detailed explanation
var longOptions = new AgentExecutionOptions
{
MaxCompletionTokens = 2000
};
var detailed = agent.Run("Explain machine learning in detail.", longOptions);
Remarks
Controls the maximum length of the agent's response. Set to null to use the default (2048 tokens). Set to -1 to disable the limit entirely.
Lower values are useful for:
- Summary tasks requiring concise responses.
- Reducing latency in chat applications.
- Controlling output costs.