(model, test_loader, device, verbose=0, lossFn=None,
post_proc = lambda args: args)
| 227 | |
| 228 | |
| 229 | def test(model, test_loader, device, verbose=0, lossFn=None, |
| 230 | post_proc = lambda args: args): |
| 231 | |
| 232 | model.eval() |
| 233 | if lossFn is None: |
| 234 | lossFn = nn.MSELoss() |
| 235 | |
| 236 | |
| 237 | total_loss = 0. |
| 238 | predictions = [] |
| 239 | |
| 240 | with torch.no_grad(): |
| 241 | for data, target in test_loader: |
| 242 | bs = len(data) |
| 243 | |
| 244 | data, target = data.to(device), target.to(device) |
| 245 | output = model(data) |
| 246 | output = post_proc(output) |
| 247 | |
| 248 | loss = lossFn(output.view(bs, -1), target.view(bs, -1)) |
| 249 | total_loss += loss.sum().item() |
| 250 | |
| 251 | return total_loss/len(test_loader.dataset) |
| 252 | |
| 253 | |
| 254 | # Till EoF |
nothing calls this directly
no outgoing calls
no test coverage detected