cl-waffe

cl-waffe copied to clipboard

TODO

8

1. ~~Optimize the overhead of generic functions `call-forward`/`call-backward` (Less Important)~~ It's done. 3. ~~make print-object rich~~(Done)

hikettei

Add LSTM/GRU

1

hikettei

Almost features are based on mgl-mat so it is not going to be a rocky road. If possible, I want backends to support FP16

hikettei

To update the computation nodes, (setf !aref) needed to be called like: ```lisp (setq tensor (setf (!aref tensor ~) x)) ``` This should be rewritten with macro, being easy to...

hikettei

Save and restore the weights of model

1

hikettei

!expands behaves like PyTorch's one

hikettei

Tensorデータ型やBackendは全てdefparameterを参照しているから ```lisp (with-config (:dtype :float) (:backends (:simd :numcl :mgl)) ; Backendの優先順位 ...)``` みたいなマクロでまとめて設定できると便利そう。FP16対応する方法思いついたら作ろっかな〜

hikettei

```lisp (!beta `(10 10) 1.0 1.0) ;arithmetic error FLOATING-POINT-INVALID-OPERATION signalled ;Operation was (/ 0.0 0.0). ``` 後でなおす

hikettei

In the function trainer: ```lisp (defun valid (trainer dataset batch-size) "Valid trainer" (let ((count 0) (correct 0)) (loop for index fixnum upfrom 0 below (get-dataset-length dataset) by batch-size do (let*...

hikettei

hikettei