lyrics[i] += row['lyric'] + '.\n'
#When it's done joining a song's lyrics lines , go to the next song :
else :
lyrics,songID,songName = processFirstLine(lyrics,songID,songName,row)
songNumber = row['track_n']
i+=1
# Define a new pandas DataFrame to save songID , songName , Lyrics in it to use them later
lyrics_data = pd.DataFrame({'songID':songID, 'songName':songName, 'lyrics':lyrics })
lyrics_data.head(3)
with open('/content/drive/My Drive/lyrics.txt', 'w') as f:
for item in lyrics_data.lyrics.to_list():
f.write("%s\n" % item)
gpt2.copy_file_from_gdrive('lyrics.txt')
sess = gpt2.start_tf_sess()
gpt2.finetune(sess,
dataset='lyrics.txt',
model_name='124M',
steps=1000,
restore_from='fresh',
run_name='run1',
print_every=10,
sample_every=200,
save_every=500
)
gpt2.generate(sess,
gpt2.mount_gdrive()
"""## Uploading txt file
Upload **any smaller text file** (<10 MB) and update the file name in the cell below, then run the cell.
"""
file_name = "shakespeare.txt"
if not os.path.isfile(file_name):
url = "https://raw.githubusercontent.com/karpathy/char-rnn/master/data/tinyshakespeare/input.txt"
data = requests.get(url)
with open(file_name, 'w') as f:
f.write(data.text)
gpt2.copy_file_from_gdrive(file_name)
"""## training ... may take a while. TensorFlow magic, saving checkpoints, be sure to save if using coLab so its not a bunch of wasted time..
* **`restore_from`**: Set to `fresh` to start training from the base GPT-2, or set to `latest` to restart training from an existing checkpoint.
* **`sample_every`**: Number of steps to print example output
* **`print_every`**: Number of steps to print training progress.
* **`learning_rate`**: Learning rate for the training. (default `1e-4`, can lower to `1e-5` if you have <1MB input data)
* **`run_name`**: subfolder within `checkpoint` to save the model. This is useful if you want to work with multiple models (will also need to specify `run_name` when loading the model)
* **`overwrite`**: Set to `True` if you want to continue finetuning an existing model (w/ `restore_from='latest'`) without creating duplicate copies.
"""
sess = gpt2.start_tf_sess()
gpt2.finetune(sess,
dataset=file_name,
pip install requests
pip install tqdm
print(tf.keras.__version__)
print(tf.__version__)
!ls
"""# Loading Dataset"""
gpt2.mount_gdrive()
drive.mount("/content/drive")
gpt2.copy_file_from_gdrive('emily_dickinson_titleless.txt')
"""The data set loaded above was taken from [JenLooper's Kraggle data set](https://www.kaggle.com/jenlooper/emily-dickinson-poetry). In my earlier test iterations, I had tried to incorporate an API between my Colab Notebook and Kraggle. However, I realised it failed quite often while I was running it. So I extracted out the .txt file and placed it within my Google Drive instead."""
with io.open('emily_dickinson_titleless.txt', encoding='utf-8') as f:
text = f.read().lower()
print('corpus length:', len(text))
text[:100]
len(text[:50])
!ls
chars = sorted(list(set(text)))
print('total chars:', len(chars))