def on_train_begin(self, **kwargs):
info = layers_info(self.learner)
try_mlflow_log(mlflow.log_param, "num_layers", len(info))
try_mlflow_log(mlflow.log_param, "opt_func",
self.opt_func.func.__name__)
if hasattr(self.opt, "true_wd"):
try_mlflow_log(mlflow.log_param, "true_wd", self.opt.true_wd)
if hasattr(self.opt, "bn_wd"):
try_mlflow_log(mlflow.log_param, "bn_wd", self.opt.bn_wd)
if hasattr(self.opt, "train_bn"):
try_mlflow_log(mlflow.log_param, "train_bn", self.train_bn)
summary = model_summary(self.learner)
try_mlflow_log(mlflow.set_tag, "model_summary", summary)
tempdir = tempfile.mkdtemp()
try:
summary_file = os.path.join(tempdir, "model_summary.txt")
with open(summary_file, "w") as f:
f.write(summary)
try_mlflow_log(mlflow.log_artifact, local_path=summary_file)
finally:
shutil.rmtree(tempdir)
def on_train_begin(self, **kwargs):
info = layers_info(self.learner)
try_mlflow_log(kiwi.log_param, 'num_layers', len(info))
try_mlflow_log(kiwi.log_param, 'opt_func', self.opt_func.func.__name__)
if hasattr(self.opt, 'true_wd'):
try_mlflow_log(kiwi.log_param, 'true_wd', self.opt.true_wd)
if hasattr(self.opt, 'bn_wd'):
try_mlflow_log(kiwi.log_param, 'bn_wd', self.opt.bn_wd)
if hasattr(self.opt, 'train_bn'):
try_mlflow_log(kiwi.log_param, 'train_bn', self.train_bn)
summary = model_summary(self.learner)
try_mlflow_log(kiwi.set_tag, 'model_summary', summary)
tempdir = tempfile.mkdtemp()
try:
summary_file = os.path.join(tempdir, "model_summary.txt")
with open(summary_file, 'w') as f:
f.write(summary)
try_mlflow_log(kiwi.log_artifact, local_path=summary_file)
finally:
shutil.rmtree(tempdir)
cb_handler=cb_handler)
# %%
# lr_find(learner)
# %%
# learner.recorder.plot()
# %%
fit_one_cycle(learner, 1, max_lr=1e-2)
# %%
from utils.torch_utils import model_info
model_info(model)
learner.model
learner.layer_groups
learner.lr_range(slice(1e-6, 1e-4))
from fastai.callbacks import hooks
hooks.model_summary(learner)
# %%
model_state = model.module.state_dict() if type(
model) is nn.parallel.DistributedDataParallel else model.state_dict()
torch.save({'model': model_state}, "weights/fastai.pt")