Models

The .models accessor provides methods for listing and downloading models.

Model class

QuantizedFile class

Downloading models

Download a quantized model file using one of the following approaches:

Using client.models.download()

client.models.download('OpenHermes-2.5-Mistral-7B/Q4_K_M')

client.models.download() accepts either a string reference ('model/quantization') or a QuantizedFile object.

---

If the model has already been downloaded, the function exits immediately. Otherwise, a progress bar displays.

Using Model.download()

model = client.models.get('OpenHermes-2.5-Mistral-7B')
model.download('Q4_K_M')

Using QuantizedFile.download()

model = client.models.get('OpenHermes-2.5-Mistral-7B')
quantization = model.get_quantization('Q4_K_M')
quantization.download()

Example usage

from anaconda_ai import AnacondaAIClient

client = AnacondaAIClient()

# List all models
models = client.models.list()
for model in models:
    print(f"{model.name}: {model.num_parameters / 1e9:.1f}B parameters")

# Get specific model details
model = client.models.get('OpenHermes-2.5-Mistral-7B')
print(f"Description: {model.description}")
print(f"Trained for: {model.trained_for}")

# Check available quantizations
for quant in model.quantized_files:
    status = "downloaded" if quant.is_downloaded else "available"
    print(f"  {quant.quant_method}: {status} ({quant.size_bytes / 1e9:.1f}GB)")

# Download and use a quantization
if not model.get_quantization('Q4_K_M').is_downloaded:
    model.download('Q4_K_M')

quantization = model.get_quantization('Q4_K_M') 
print(f"Local path: {quantization.local_path}")