Int4llamaDecoder Class Reference

Public Member Functions
	Int4llamaDecoder (std::string param_path, const struct model_config config)
Matrix3D< float >	prepare_decoder_attention_mask (int length, int past_length)
struct Int4llamaDecoder_output	forward (std::string param_path, const struct Int4llamaDecoder_input &input)

Public Attributes
int	voc_size
int	embed_dim
int	padding_idx
int	hidden_dim
int	num_heads
float	rms_norm_eps
std::vector< Int4llamaDecoderLayer >	layers
std::string	profile_name = "Int4llamaDecoder"
Embedding	embed_tokens
LlamaRMSNorm	norm
float *	attention_mask_buf
float *	pos_embeds_buf
float *	last_hidden_states_buf
float *	hidden_states_buf
float *	inputs_embeds_buf
float *	first_input_ids_buf
float *	image_embed_buf
float *	second_input_ids_buf
float *	norm_weight_ptr = nullptr

---

The documentation for this class was generated from the following files:

• llm/include/nn_modules/Int4llamaDecoder.h
• llm/src/nn_modules/non_cuda/Int4llamaDecoder.cc