amazon-sagemaker - 语法错误（amazon-sagemaker-stock-prediction/dbg-custom-rnn.ipython）

Question

我正在SageMaker Notebook 实例上从https://github.com/aws-samples/amazon-sagemaker-stock-prediction/blob/master/notebooks/dbg-custom-rnn.ipynb运行下面的代码单元。有关更多信息，这是我在 GitHub 上发布的错误的链接：https ://github.com/aws-samples/amazon-sagemaker-stock-prediction/issues/7 。

%%time
Instantiate estimator with container image of artifact and backend EC2 instance(s)
rnn = Estimator(image,
role, 1, 'ml.c5.18xlarge',
output_path=output_location,
base_job_name = base_job_name,
sagemaker_session=session)

rnn.set_hyperparameters(**hyperparameters)
Train the model
rnn.fit(data_location)
estimator_job = rnn.latest_training_job.job_name
model_archive = "{}/{}/output/{}/output/model.tar.gz".format(artifactname,interval,estimator_job)
print("Estimator created at completion of training job {}".format(estimator_job))

我遇到了以下错误：

Parameter image_name will be renamed to image_uri in SageMaker Python SDK v2.
's3_input' class will be renamed to 'TrainingInput' in SageMaker Python SDK v2.
2020-08-12 15:58:15 Starting - Starting the training job...
2020-08-12 15:58:17 Starting - Launching requested ML instances......
2020-08-12 15:59:38 Starting - Preparing the instances for training...
2020-08-12 16:00:16 Downloading - Downloading input data...
2020-08-12 16:00:22 Training - Downloading the training image...
2020-08-12 16:01:13 Uploading - Uploading generated training model.2020-08-12 16:01:08.786584: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'libcudart.so.10.1'; dlerror: libcudart.so.10.1: cannot open shared object file: No such file or directory
2020-08-12 16:01:08.786650: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.
Starting the training.
Hyperparameters file : {"target_stock": "BMW", "lag": "10", "interval": "D", "batch_size": "4096", "covariate_columns": "StartPrice, MinPrice, MaxPrice", "percent_train": "85.0", "covariate_stocks": "CON, DAI, PAH3, VOW3", "dropout_ratio": "0.1", "num_epochs": "1000", "target_column": "EndPrice", "horizon": "5", "num_units": "256"}
Hyperparameters initialized
Loading data from : /opt/ml/input/data/training/resampled_stockdata.csv
Loading data from : /opt/ml/input/data/training/resampled_stockdata.csv
Training data loaded
100 Stock symbols found.
Records for 65 trading days found.
0-CON#0111-DAI#0112-PAH3#0113-VOW3
Exception during training: invalid syntax (core.py, line 314)
Traceback (most recent call last):
File "/opt/program/train", line 255, in train
traindata.to_csv(os.path.join(model_path, trainfile))
File "/usr/local/lib/python3.5/dist-packages/pandas/core/generic.py", line 3020, in to_csv
formatter.save()
File "/usr/local/lib/python3.5/dist-packages/pandas/io/formats/csvs.py", line 157, in save
compression=self.compression)
File "/usr/local/lib/python3.5/dist-packages/pandas/io/common.py", line 344, in _get_handle
from s3fs import S3File
File "/usr/local/lib/python3.5/dist-packages/s3fs/__init__.py", line 1, in <module>
from .core import S3FileSystem, S3File
File "/usr/local/lib/python3.5/dist-packages/s3fs/core.py", line 8, in <module>
from fsspec import AbstractFileSystem
File "/usr/local/lib/python3.5/dist-packages/fsspec/__init__.py", line 10, in <module>
from .mapping import FSMap, get_mapper
File "/usr/local/lib/python3.5/dist-packages/fsspec/mapping.py", line 2, in <module>
from .core import url_to_fs
File "/usr/local/lib/python3.5/dist-packages/fsspec/core.py", line 314
out[0] = (f"{out[0][1]}://", out[0][1], out[0][2])
^
SyntaxError: invalid syntax
2020-08-12 16:01:19 Failed - Training job failed
UnexpectedStatusException Traceback (most recent call last)
<timed exec> in <module>

~/anaconda3/envs/python3/lib/python3.6/site-packages/sagemaker/estimator.py in fit(self, inputs, wait, logs, job_name, experiment_config)
497 self.jobs.append(self.latest_training_job)
498 if wait:
--> 499 self.latest_training_job.wait(logs=logs)
500
501 def _compilation_job_name(self):

~/anaconda3/envs/python3/lib/python3.6/site-packages/sagemaker/estimator.py in wait(self, logs)
1114 # If logs are requested, call logs_for_jobs.
1115 if logs != "None":
-> 1116 self.sagemaker_session.logs_for_job(self.job_name, wait=True, log_type=logs)
1117 else:
1118 self.sagemaker_session.wait_for_job(self.job_name)

~/anaconda3/envs/python3/lib/python3.6/site-packages/sagemaker/session.py in logs_for_job(self, job_name, wait, poll, log_type)
3075
3076 if wait:
-> 3077 self._check_job_status(job_name, description, "TrainingJobStatus")
3078 if dot:
3079 print()

~/anaconda3/envs/python3/lib/python3.6/site-packages/sagemaker/session.py in _check_job_status(self, job, desc, status_key_name)
2669 ),
2670 allowed_statuses=,
-> 2671 actual_status=status,
2672 )
2673

UnexpectedStatusException: Error for Training job dbg-custom-rnn-D-BMW-2020-08-12-15-58-15-812: Failed. Reason: AlgorithmError: Exception during training: invalid syntax (core.py, line 314)
Traceback (most recent call last):
File "/opt/program/train", line 255, in train
traindata.to_csv(os.path.join(model_path, trainfile))
File "/usr/local/lib/python3.5/dist-packages/pandas/core/generic.py", line 3020, in to_csv
formatter.save()
File "/usr/local/lib/python3.5/dist-packages/pandas/io/formats/csvs.py", line 157, in save
compression=self.compression)
File "/usr/local/lib/python3.5/dist-packages/pandas/io/common.py", line 344, in _get_handle
from s3fs import S3File
File "/usr/local/lib/python3.5/dist-packages/s3fs/__init__.py", line 1, in <module>
from .core import S3FileSystem, S3File
File "/usr/local/lib/python3.5/dist-packages/s3fs/core.py", line 8, in <module>
from fsspec import AbstractFileSystem
File "/usr/local/lib/python3.5/dist-packages/fsspec/__init__.py", line 10, in <module>
from .mapping import FSMap, get_mapper
File "/usr/local/lib/python3.5/dist-packages/fsspec/map

谁能为我解释如何解决这个错误。太感谢了！

标签： amazon-sagemakerpredictive

Answer 1

将 s3_input 重命名为 TrainingInput 并删除警告消息。可能这会帮助你https://github.com/aws/sagemaker-python-sdk/pull/1680