Skip to main content

​
Models

This guide covers the model management capabilities of the Ollama service.

​
Models Endpoint

Manage and query available AI models.

​
List Models Endpoint

GET /api/models

curl "https://ollama.moodmnky.com/api/models" \
  -H "x-api-key: your_api_key"
Response: 
{
  "models": [
    {
      "name": "llama2",
      "size": 3791730718,
      "modified_at": "2024-03-20T15:23:44Z",
      "details": {
        "format": "gguf",
        "family": "llama",
        "parameter_size": "7B",
        "quantization_level": "Q4_K_M"
      }
    }
  ]
}

​
Show Model Info Endpoint

GET /api/models/{model}

curl "https://ollama.moodmnky.com/api/models/llama2" \
  -H "x-api-key: your_api_key"
Response: 
{
  "name": "llama2",
  "size": 3791730718,
  "modified_at": "2024-03-20T15:23:44Z",
  "details": {
    "format": "gguf",
    "family": "llama",
    "parameter_size": "7B",
    "quantization_level": "Q4_K_M",
    "license": "llama2",
    "system_prompt": "You are a helpful AI assistant.",
    "context_window": 4096,
    "default_generation_params": {
      "temperature": 0.7,
      "top_p": 0.9,
      "top_k": 40
    }
  }
}

​
Pull Model Endpoint

POST /api/models/pull

curl -X POST "https://ollama.moodmnky.com/api/models/pull" \
  -H "x-api-key: your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "llama2",
    "insecure": false
  }'
Response (streaming): 
{"status":"pulling manifest"}
{"status":"pulling layers","digest":"sha256:..."}
{"status":"verifying sha256 digest"}
{"status":"writing manifest"}
{"status":"removing any unused layers"}
{"status":"success"}

​
Delete Model Endpoint

DELETE /api/models/{model}

curl -X DELETE "https://ollama.moodmnky.com/api/models/llama2" \
  -H "x-api-key: your_api_key"

​
Available Models

ModelSizeDescriptionUse Cases
llama27BGeneral purpose modelText generation, chat, Q&A
codellama7BCode-specialized modelCode completion, explanation
mistral7BInstruction-tuned modelTask completion, reasoning
orca-mini3BLightweight modelQuick responses, testing

​
Usage Examples

​
List Available Models

const response = await client.ollama.listModels();
console.log("Available models:", response.models);

​
Get Model Information

const response = await client.ollama.getModel("llama2");
console.log("Model details:", response.details);

​
Pull New Model

const response = await client.ollama.pullModel({
  name: "codellama",
  insecure: false
});

// Handle streaming response
for await (const chunk of response) {
  console.log("Pull status:", chunk.status);
}

​
Delete Model

await client.ollama.deleteModel("unused-model");

​
Model Management Best Practices

  1. Model Selection
    • Choose appropriate model size for task
    • Consider memory and performance requirements
    • Test models before production use
    • Monitor model performance metrics
  2. Resource Management
    • Clean up unused models regularly
    • Monitor disk space usage
    • Implement model rotation policies
    • Cache frequently used models
  3. Security Considerations
    • Verify model checksums
    • Use secure connections for pulls
    • Implement access controls
    • Monitor model usage patterns
  4. Performance Optimization
    • Pre-load frequently used models
    • Use appropriate quantization levels
    • Monitor memory usage
    • Implement model warm-up strategies

​
Error Handling

try {
  const response = await client.ollama.pullModel({
    name: "llama2"
  });
} catch (error) {
  switch (error.code) {
    case "MODEL_NOT_FOUND":
      console.error("Model not found in registry");
      break;
    case "INSUFFICIENT_SPACE":
      console.error("Not enough disk space");
      break;
    case "PULL_ERROR":
      console.error("Error pulling model:", error.message);
      break;
  }
}

​
Model Lifecycle Management

  1. Initial Setup
    • Model selection and testing
    • Performance benchmarking
    • Resource allocation
    • Access control configuration
  2. Maintenance
    • Regular updates
    • Performance monitoring
    • Usage analytics
    • Health checks
  3. Retirement
    • Usage evaluation
    • Replacement planning
    • Data migration
    • Clean removal

​
Support & Resources