skewondr/FoLiBi

Forgetting-aware Linear Bias for Attentive Knowledge Tracing

Forgetting-aware Linear Bias for Attentive Knowledge Tracing

[paper link] (https://arxiv.org/abs/2309.14796)

### pip install accelerate einops yacs iopath

### execute baselines 
CUDA_VISIBLE_DEVICES=0 python main.py --seed 12405 --model_name sakt --data_name algebra05 --de_type none_0 --gpu_num 0 --server_num 0 --describe baselines

### execute rc 
CUDA_VISIBLE_DEVICES=0 python main.py --seed 12405 --model_name sakt --data_name algebra05 --de_type relative_0 --gpu_num 0 --server_num 0 --describe rc

### execute mono
CUDA_VISIBLE_DEVICES=0 python main.py --seed 12405 --model_name sakt --data_name algebra05 --de_type monotonic_0 --gpu_num 0 --server_num 0 --describe mono

### execute alibi
CUDA_VISIBLE_DEVICES=0 python main.py --seed 12405 --model_name sakt --data_name algebra05 --de_type alibi1_0 --gpu_num 0 --server_num 0 --describe alibi