QwenModelConfig.ixx File Reference

Deployment configuration for Qwen 3.8 language models. More...

#include <string>
import Dnn.TensorTypes;
import Dnn.LanguageModelConfig;

Classes
struct	Mila::Dnn::QwenModelConfig
	Deployment configuration for Qwen 3.8 language models. More...

Namespaces
namespace	Mila
	Mila main API namespace.

Detailed Description

Deployment configuration for Qwen 3.8 language models.

QwenModelConfig is the concrete configuration type passed to QwenModel::load(). It inherits every universal deployment concern from LanguageModelConfig<QwenModelConfig>:

• context_length – maximum sequence length
• weight_quantization – Linear weight storage strategy
• kv_cache_compression – GroupedQueryAttention cache strategy

Every architectural parameter (the 3:1 mixer interleave, the DeltaNet head geometry, the attention output gate, the partial rotary width) is read from checkpoint metadata at load time and is not a deployment concern, so this carries no architecture fields beyond the base.