|
|
|
|
|
|
| import os |
| import torch |
| try: |
| from torch._six import inf |
| except: |
| from torch import inf |
|
|
|
|
| class NativeScalerWithGradNormCount: |
| state_dict_key = "amp_scaler" |
|
|
| def __init__(self): |
| |
| self._scaler = torch.cuda.amp.GradScaler(init_scale=1) |
|
|
| def __call__(self, loss, optimizer, clip_grad=None, parameters=None, create_graph=False, update_grad=True): |
| self._scaler.scale(loss).backward(create_graph=create_graph) |
| if update_grad: |
| if clip_grad is not None: |
| assert parameters is not None |
| self._scaler.unscale_(optimizer) |
| try: |
| norm = torch.nn.utils.clip_grad_norm_(parameters, clip_grad, error_if_nonfinite=True) |
| except: |
| print('NAN gradient ....') |
| else: |
| raise NotImplementedError |
| self._scaler.unscale_(optimizer) |
| norm = get_grad_norm_(parameters) |
| self._scaler.step(optimizer) |
| self._scaler.update() |
| else: |
| norm = None |
| return True |
| |
|
|
| def state_dict(self): |
| return self._scaler.state_dict() |
|
|
| def load_state_dict(self, state_dict): |
| self._scaler.load_state_dict(state_dict) |
|
|
| def get_grad_norm_(parameters, norm_type: float = 2.0) -> torch.Tensor: |
| if isinstance(parameters, torch.Tensor): |
| parameters = [parameters] |
| parameters = [p for p in parameters if p.grad is not None] |
| norm_type = float(norm_type) |
| if len(parameters) == 0: |
| return torch.tensor(0.) |
| device = parameters[0].grad.device |
| if norm_type == inf: |
| total_norm = max(p.grad.detach().abs().max().to(device) for p in parameters) |
| else: |
| total_norm = torch.norm(torch.stack([torch.norm(p.grad.detach(), norm_type).to(device) for p in parameters]), norm_type) |
| return total_norm |
|
|
| def is_bf16_supported(): |
| """Returns a bool indicating if the current CUDA device supports dtype bfloat16""" |
| cu_vers = torch.version.cuda |
| if cu_vers is not None: |
| cuda_maj_decide = int(cu_vers.split('.')[0]) >= 11 |
| else: |
| cuda_maj_decide = False |
| return torch.cuda.get_device_properties(torch.cuda.current_device()).major >= 8 and cuda_maj_decide |
