import_template erro Expecting value: line 1 column 1 (char 0)

[wuzxc1230123]

import pandas as pd
from autots import AutoTS 

symbol='BTCUSDT'


forecast_length=1

df = pd.read_csv(symbol+"-Reverse-Eval.csv")
df=df[-200:]
print(df.head())

# at = AutoTS(
#     forecast_length=forecast_length,
#     frequency='infer',
#     prediction_interval=0.9,
#     ensemble='all',
#     model_list="fast",  # "superfast", "default", "fast_parallel"
#     transformer_list="fast",  # "superfast",
#     drop_most_recent=1,
#     max_generations=15,
#     subset=100,
#     num_validations=2,
#     validation_method="backwards"
# )

at = AutoTS(
    forecast_length=forecast_length,
    frequency='infer',
    prediction_interval=0.9,
    ensemble='all',
    model_list="fast",  # "superfast", "default", "fast_parallel"
    transformer_list="fast",  # "superfast",
    drop_most_recent=1,
    max_generations=1,
    subset=1,
    num_validations=0,
    validation_method="backwards"
)

model = at.fit(
    df,
    date_col='StartTime',
    value_col='Label',
    id_col='Symbol'
)
# prediction = model.predict()
# print(prediction.forecast)

example_filename = "example_export.csv"  # .csv/.json
model.export_template(example_filename, models='best')


# print(prediction.forecast)


at2 = AutoTS(
    forecast_length=forecast_length,
    frequency='infer',
    prediction_interval=0.9,
    ensemble='all',
    model_list="fast",  # "superfast", "default", "fast_parallel"
    transformer_list="fast",  # "superfast",
    drop_most_recent=1,
    max_generations=1,
    subset=1,
    num_validations=0,
    validation_method="backwards"
)
model2 = at2.import_template(example_filename, method='only') # method='add on'

model2.fit_data(df,    
    date_col='StartTime',
    value_col='Label',
    id_col='Symbol')

prediction2 = model2.predict()
print(prediction2.forecast)

prediction2 = model2.predict() Expecting value: line 1 column 1 (char 0)

[winedarksea]

You should try reformatting your code here on Github with it inside a block of: ``` right now the markdown you have it turning code comments into headers

https://docs.github.com/en/get-started/writing-on-github/getting-started-with-writing-and-formatting-on-github/basic-writing-and-formatting-syntax

Edit: I've done it for you

[wuzxc1230123]

你应该尝试在 Github 上重新格式化你的代码，并将其放在以下块中： ``` 现在，你已经将代码注释转换为标题了

https://docs.github.com/en/get-started/writing-on-github/getting-started-with-writing-and-formatting-on-github/basic-writing-and-formatting-syntax

编辑：我已经为你做到了

I’m a little curious. The examples on the official website cannot use already trained models.

[winedarksea]

Okay, here's the issue, and yes, it's a bit confusing, so sorry about that:

• import_template loads the initial template for a model search (.fit) which is a "seed" for a generation search. It allows a gradually evolving template
• import_best_model imports just the best model and then loads it ready for prediction

so you need to replace

model2 = at2.import_template(example_filename, method='only') # method='add on'

with

model2 = at2.import_best_model(example_filename)

and it should work

[wuzxc1230123]

thanks