File size: 3,103 Bytes
c881b77 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 | from collections import OrderedDict
from typing import Any, Tuple
import os, torch
class Dict(OrderedDict):
"""Base ModelOutput class fixing the output type from the models. This class is inspired from
the ``ModelOutput`` class from hugginface transformers library"""
def __getitem__(self, k):
if isinstance(k, str):
self_dict = {k: v for (k, v) in self.items()}
return self_dict[k]
else:
return self.to_tuple()[k]
def __setattr__(self, name, value):
super().__setitem__(name, value)
super().__setattr__(name, value)
def __setitem__(self, key, value):
if isinstance(value, dict) and not isinstance(value, Dict):
value = Dict(value)
elif isinstance(value, list):
value = [Dict(v) if isinstance(v, dict) and not isinstance(v, Dict) else v for v in value]
super().__setitem__(key, value)
super().__setattr__(key, value)
def to_tuple(self) -> Tuple[Any]:
"""
Convert self to a tuple containing all the attributes/keys that are not ``None``.
"""
return tuple(self[k] for k in self.keys())
def to_dict(self) -> dict:
result = {}
for k, v in self.items():
if isinstance(v, Dict):
result[k] = v.to_dict()
elif isinstance(v, list):
result[k] = [i.to_dict() if isinstance(i, Dict) else i for i in v]
else:
result[k] = v
return result
def __repr__(self):
return repr(self.to_dict())
def __str__(self):
return str(self.to_dict())
def get_ckpt_path(config, ckpt_steps=None):
parts = [config.ckpt_dir, config.task_name, config.phase]
if config.phase == 'novel':
parts.append(f'run-{config.run}' if config.run is not None else "run-default")
parts.append(f'{config.dataset.novel_settings.k_shot}-shot')
parts.append(f'shuffle_seed-{config.dataset.novel_settings.shuffle_seed}')
if ckpt_steps is not None:
parts.append(f'checkpoint-{ckpt_steps}')
os.makedirs(os.path.join(*parts), exist_ok=True)
return os.path.join(*parts)
def get_output_path(config, ckpt_steps=None):
parts = [config.inference.output_dir, config.task_name, config.phase]
if config.phase == 'novel':
parts.append(f'run-{config.run}' if config.run is not None else "run-default")
parts.append(f'{config.dataset.novel_settings.k_shot}-shot')
parts.append(f'shuffle_seed-{config.dataset.novel_settings.shuffle_seed}')
if ckpt_steps is not None:
parts.append(f'checkpoint-{ckpt_steps}')
os.makedirs(os.path.join(*parts), exist_ok=True)
return os.path.join(*parts)
def manual_average_gradients(model, accelerator):
if accelerator.num_processes > 1:
for param in model.parameters():
if param.grad is not None and param.requires_grad:
torch.distributed.all_reduce(param.grad.data, op=torch.distributed.ReduceOp.SUM)
param.grad.data /= accelerator.num_processes |