Here is a list of all variables with links to the classes they belong to:

- q -

• q_ : Mila::Dnn::Compute::CpuAttentionOp, Mila::Dnn::Compute::Cuda::Gqa::CudaGqaOp< TPrecision >, Mila::Dnn::Compute::Cuda::MultiHeadAttention::CudaMultiHeadAttentionOp< TPrecision >, Mila::Dnn::LlamaBlock< TDeviceType, TPrecision, TWeightQuant, TKvPolicy >
• q_offset_ : Mila::Dnn::LlamaBlock< TDeviceType, TPrecision, TWeightQuant, TKvPolicy >
• q_permute : Mila::Dnn::Compute::GqaState
• q_permute_opt_ : Mila::Dnn::Compute::Cuda::Gqa::CudaGqaOp< TPrecision >
• q_permute_tensor_optimized_ : Mila::Dnn::Compute::Cuda::Gqa::CudaGqaOp< TPrecision >
• q_prefill_offset_ : Mila::Dnn::LlamaBlock< TDeviceType, TPrecision, TWeightQuant, TKvPolicy >
• q_prefill_shape_ : Mila::Dnn::LlamaBlock< TDeviceType, TPrecision, TWeightQuant, TKvPolicy >
• q_proj : Mila::Dnn::Visualization::VisualizerContext
• q_shape_ : Mila::Dnn::LlamaBlock< TDeviceType, TPrecision, TWeightQuant, TKvPolicy >, Mila::Dnn::Rope< TDeviceType, TPrecision >
• q_tensor_ : Mila::Dnn::Compute::CpuAttentionOp, Mila::Dnn::Compute::Cuda::Gqa::CudaGqaOp< TPrecision >, Mila::Dnn::Compute::Cuda::MultiHeadAttention::CudaMultiHeadAttentionOp< TPrecision >
• qk_decode_plan_ : Mila::Dnn::Compute::Cuda::Gqa::CudaGqaOp< TPrecision >, Mila::Dnn::Compute::Cuda::MultiHeadAttention::CudaMultiHeadAttentionOp< TPrecision >
• qk_decode_plan_optimized_ : Mila::Dnn::Compute::Cuda::Gqa::CudaGqaOp< TPrecision >
• qk_partial_prefill_plan_cache_ : Mila::Dnn::Compute::Cuda::Gqa::CudaGqaOp< TPrecision >
• qk_partial_prefill_plan_cache_optimized_ : Mila::Dnn::Compute::Cuda::Gqa::CudaGqaOp< TPrecision >
• qk_prefill_plan_ : Mila::Dnn::Compute::Cuda::Gqa::CudaGqaOp< TPrecision >
• qk_prefill_plan_optimized_ : Mila::Dnn::Compute::Cuda::Gqa::CudaGqaOp< TPrecision >
• qk_score_plan_ : Mila::Dnn::Compute::Cuda::Gqa::CudaGqaOp< TPrecision >, Mila::Dnn::Compute::Cuda::MultiHeadAttention::CudaMultiHeadAttentionOp< TPrecision >
• qkv_dim_ : Mila::Dnn::Compute::CpuAttentionOp, Mila::Dnn::Compute::Cuda::MultiHeadAttention::CudaMultiHeadAttentionOp< TPrecision >
• qkv_proj_ : Mila::Dnn::GptBlock< TDeviceType, TPrecision >, Mila::Dnn::LlamaBlock< TDeviceType, TPrecision, TWeightQuant, TKvPolicy >
• queue_ : Mila::Dnn::Compute::ExecutionContext< DeviceType::Vulkan >
• queue_created_ : Mila::Dnn::Compute::ExecutionContext< DeviceType::Metal >, Mila::Dnn::Compute::ExecutionContext< DeviceType::Vulkan >