How to display full training table in ipywidgets

Hi, i am trying to build a simple full training process using ipywidgets and voila,
I have been able to show dropbox to select model and metric. When click button start training, i want to display the full training process out but seem like it only show an empty bar. Is anyone have the same problem? My code is below:

import fastai
from import *
from import *
import torchvision

path = untar_data(URLs.MNIST_TINY)

model_dropdown = widgets.Dropdown(options=‘resnet34 resnet50 resnet101 resnet152’.split(" “), description=“Select model”)
metric_dropdown = widgets.Dropdown(options=‘accuracy error_rate’.split(” "), description=“Select metric”)

train_btn = train_btn = widgets.Button(description=“Start training”)
out = widgets.Output()
def train_btn_click(b):
dls = ImageDataLoaders.from_folder(path, item_tfms=Resize(128, method=ResizeMethod.Squish))
model = model_dropdown.value
metric = metric_dropdown.value
learner = cnn_learner(dls, getattr(torchvision.models,model), metrics=getattr(fastai.metrics,metric))

widgets.VBox([model_dropdown, metric_dropdown, train_btn])