GemmaModelConfig.ixx File Reference

Deployment configuration for Gemma language models. More...

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

Classes
struct	Mila::Dnn::GemmaModelConfig
	Deployment configuration for Gemma language models. More...

Namespaces
namespace	Mila
	Mila main API namespace.

Detailed Description

Deployment configuration for Gemma language models.

GemmaModelConfig is the concrete configuration type passed to GemmaModel::fromPretrained(). It inherits all universal language model deployment concerns from LanguageModelConfig<GemmaModelConfig>:

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

Every Gemma architectural parameter (num_layers, head_dim, the global-layer geometry, the sliding window, dual RoPE bases, logit softcap, etc.) is read from checkpoint metadata at load time and is not a deployment concern, so GemmaModelConfig carries no architecture-specific fields beyond the base.