Responsible for generating GPT calls

• A_llamarec: llamarec preference and perception generation
• input: Text input of embeding.py
• embeding.py: Performs embedding on the text information from input and maps it to the input dimension of SASRec
• SASRec_emb: Calls GPT to generate text content embedded in SASRec

Training and inference of LlamaRec

• 0_Grocery_and_Gourmet_Food_sft0.py: Constructs the GPT call prompt
• 0_Grocery_and_Gourmet_Food_sft1.py: Builds SFT training data and trains using Llama-Factory
• 0_sft_data.py: Directly constructs SFT without perception
• 2_inference.py: Performs inference on the trained model
• bias.py: Computes bias

LlamaRec training and inference enhanced with SASRec

• Structure is the same as LlamaRec

SASRec plus training and inference

• SASRec_item_emb
  • Adds item embedding
  • pretrain: Stores offline generated item embedding
  • run.py: Performs training
• SASRec_user_emb
  • Adds user embedding
  • pretrain: Stores offline generated user embedding and item embedding
  • run.py: Performs training