untitled.py

#!/usr/bin/env python
# coding: utf-8

# In[1]:



from fastai.vision import *
from fastai.metrics import *
import glob
from shutil import copyfile, copy, move
import pandas as pd
import numpy as np

get_ipython().run_line_magic('matplotlib', 'inline')
get_ipython().run_line_magic('reload_ext', 'autoreload')
get_ipython().run_line_magic('autoreload', '2')


# In[3]:


path = Path('whales')
train = path/'train'
test = path/'test'
valid = path/'valid'


# In[4]:


path.ls()


# In[5]:


pd.read_csv(path/"train.csv").head()


# In[ ]:





# In[4]:


data = (ImageList.from_csv(path, 'train.csv', folder='train')
        #Where to find the data? -> in planet 'train' folder
        .split_by_rand_pct()
        #How to split in train/valid? -> randomly with the default 20% in valid
        .label_from_df(label_delim=' ')
        #How to label? -> use the second column of the csv file and split the tags by ' '
        .transform(tfms = get_transforms(), size=128)
        .add_test_folder()
        #Data augmentation? -> use tfms with a size of 128
        .databunch())                          
        #Finally -> use the defaults for conversion to databunch


# In[5]:


data


# In[6]:


data.show_batch(rows=3, figsize=(7,6))


# In[9]:



data.c, data.classes


# In[16]:

#Read this would fix but it didnt
# def error_rate(input:Tensor, targs:Tensor)->Rank0Tensor:
#     "1 - `accuracy`"
#     targs = targs.view(-1).long()
#     return 1 - accuracy(input, targs)


# In[7]:


learn = cnn_learner(data, models.resnet34, metrics = error_rate)
learn.model


# In[8]:


learn.fit_one_cycle(2)


# In[ ]: