Mustafa Alahmid

nn-pytorch

pytorch-bricks-classification

image-classification-with-logestic-regression-pytorch

pytorch-1

tensor-operation-and-gradients

# Building a linear regression fro scratch:

## 1- get the data set:
### __  A- create a tensor data using TensorData.
### __  B- load it with DataLoader.
## 2- import the model nn.Linear.
## 3- create prediction.
## 4- create the loss function (mse).
## 5- create the optimizer (SGD).
## 6- create the fit fun:
### __  A- generate preds.
### __  B- calculate loss.
### __  C- calculate grad.
### __  D- optimizer.
### __  E- reset opt.grad to zero.
 

linear-regression-pytorch

# Building a linear regression from scratch using PyTorch
## 1- get the dataset and change it into torch dtype.
## 2- generate weights and bias.
## 3- create the model ( x @ w.t() +b)  @ = multiplication, t=transpose, b= bias 
## 4- get the prediction 
## 5- build the loss function (mse)
## 6- compute gradient = loss.backward() 
## 7- put gradiend of w & b to zero
## 8- adjust w & b
## 9-  build the fit function or loop 

PyTorch course