Got error in the first cell:
!!pip install --disable-pip-version-check -q sagemaker==2.35.0
!conda install -q -y pytorch==1.6.0 -c pytorch
!pip install --disable-pip-version-check -q transformers==3.5.1
Anyone else experience the same issue ? I was using ‘Data Science 3’ image with ‘Python 3’ kernel.
Hi @taohuang67 ,
Try replacing second line by !conda install -q -y conda pytorch==1.6.0 -c pytorch I think pytorch isn’t getting loaded on conda. Also please check with the tornado and pyTorch preinstalled version. There might be a problem there.
Thank you,
Harsh Der
Hi Harsh,
Thanks for the help !
I restarted the kernel with ‘Data Science 2.0’, which includes python 3.8, and got pass the 1st cell. But at the second cell, got the following error with ‘import sagemaker’
TypeError Traceback (most recent call last)
Cell In[5], line 1
----> 1 import sagemaker
File /opt/conda/lib/python3.8/site-packages/sagemaker/init.py:18
14 from future import absolute_import
16 import importlib_metadata
—> 18 from sagemaker import estimator, parameter, tuner # noqa: F401
19 from sagemaker.amazon.kmeans import KMeans, KMeansModel, KMeansPredictor # noqa: F401
20 from sagemaker.amazon.pca import PCA, PCAModel, PCAPredictor # noqa: F401
File /opt/conda/lib/python3.8/site-packages/sagemaker/estimator.py:28
26 from six.moves.urllib.parse import urlparse
27 import sagemaker
—> 28 from sagemaker import git_utils, image_uris
29 from sagemaker.analytics import TrainingJobAnalytics
30 from sagemaker.debugger import TensorBoardOutputConfig # noqa: F401 # pylint: disable=unused-import
File /opt/conda/lib/python3.8/site-packages/sagemaker/image_uris.py:22
19 import re
21 from sagemaker import utils
—> 22 from sagemaker.spark import defaults
24 logger = logging.getLogger(name)
26 ECR_URI_TEMPLATE = “{registry}.dkr.{hostname}/{repository}”
File /opt/conda/lib/python3.8/site-packages/sagemaker/spark/init.py:16
13 “”“Placeholder docstring”“”
14 from future import absolute_import
—> 16 from sagemaker.spark.processing import PySparkProcessor, SparkJarProcessor # noqa: F401
File /opt/conda/lib/python3.8/site-packages/sagemaker/spark/processing.py:35
32 from urllib.parse import urlparse
34 from sagemaker import image_uris
—> 35 from sagemaker.local.image import _ecr_login_if_needed, _pull_image
36 from sagemaker.processing import ProcessingInput, ProcessingOutput, ScriptProcessor
37 from sagemaker.s3 import S3Uploader
File /opt/conda/lib/python3.8/site-packages/sagemaker/local/init.py:16
13 “”“Placeholder docstring”“”
14 from future import absolute_import
—> 16 from .local_session import ( # noqa: F401
17 file_input,
18 LocalSagemakerClient,
19 LocalSagemakerRuntimeClient,
20 LocalSession,
21 )
File /opt/conda/lib/python3.8/site-packages/sagemaker/local/local_session.py:23
20 import boto3
21 from botocore.exceptions import ClientError
—> 23 from sagemaker.local.image import _SageMakerContainer
24 from sagemaker.local.entities import (
25 _LocalEndpointConfig,
26 _LocalEndpoint,
(…)
30 _LocalTransformJob,
31 )
32 from sagemaker.session import Session
File /opt/conda/lib/python3.8/site-packages/sagemaker/local/image.py:38
35 from six.moves.urllib.parse import urlparse
37 import sagemaker
—> 38 import sagemaker.local.data
39 import sagemaker.local.utils
40 import sagemaker.utils
File /opt/conda/lib/python3.8/site-packages/sagemaker/local/data.py:26
22 from six import with_metaclass
24 from six.moves.urllib.parse import urlparse
—> 26 import sagemaker.amazon.common
27 import sagemaker.local.utils
28 import sagemaker.utils
File /opt/conda/lib/python3.8/site-packages/sagemaker/amazon/common.py:23
19 import sys
21 import numpy as np
—> 23 from sagemaker.amazon.record_pb2 import Record
24 from sagemaker.deprecations import deprecated_class
25 from sagemaker.deserializers import SimpleBaseDeserializer
File /opt/conda/lib/python3.8/site-packages/sagemaker/amazon/record_pb2.py:36
16 _sym_db = _symbol_database.Default()
19 DESCRIPTOR = _descriptor.FileDescriptor(
20 name=“record.proto”,
21 package=“aialgs.data”,
(…)
25 ),
26 )
29 _FLOAT32TENSOR = _descriptor.Descriptor(
30 name=“Float32Tensor”,
31 full_name=“aialgs.data.Float32Tensor”,
32 filename=None,
33 file=DESCRIPTOR,
34 containing_type=None,
35 fields=[
—> 36 _descriptor.FieldDescriptor(
37 name=“values”,
38 full_name=“aialgs.data.Float32Tensor.values”,
39 index=0,
40 number=1,
41 type=2,
42 cpp_type=6,
43 label=3,
44 has_default_value=False,
45 default_value=,
46 message_type=None,
47 enum_type=None,
48 containing_type=None,
49 is_extension=False,
50 extension_scope=None,
51 options=_descriptor._ParseOptions(descriptor_pb2.FieldOptions(), _b(“\020\001”)),
52 file=DESCRIPTOR,
53 ),
54 _descriptor.FieldDescriptor(
55 name=“keys”,
56 full_name=“aialgs.data.Float32Tensor.keys”,
57 index=1,
58 number=2,
59 type=4,
60 cpp_type=4,
61 label=3,
62 has_default_value=False,
63 default_value=,
64 message_type=None,
65 enum_type=None,
66 containing_type=None,
67 is_extension=False,
68 extension_scope=None,
69 options=_descriptor._ParseOptions(descriptor_pb2.FieldOptions(), _b(“\020\001”)),
70 file=DESCRIPTOR,
71 ),
72 _descriptor.FieldDescriptor(
73 name=“shape”,
74 full_name=“aialgs.data.Float32Tensor.shape”,
75 index=2,
76 number=3,
77 type=4,
78 cpp_type=4,
79 label=3,
80 has_default_value=False,
81 default_value=,
82 message_type=None,
83 enum_type=None,
84 containing_type=None,
85 is_extension=False,
86 extension_scope=None,
87 options=_descriptor._ParseOptions(descriptor_pb2.FieldOptions(), _b(“\020\001”)),
88 file=DESCRIPTOR,
89 ),
90 ],
91 extensions=,
92 nested_types=,
93 enum_types=,
94 options=None,
95 is_extendable=False,
96 syntax=“proto2”,
97 extension_ranges=,
98 oneofs=,
99 serialized_start=29,
100 serialized_end=101,
101 )
104 _FLOAT64TENSOR = _descriptor.Descriptor(
105 name=“Float64Tensor”,
106 full_name=“aialgs.data.Float64Tensor”,
(…)
175 serialized_end=175,
176 )
179 _INT32TENSOR = _descriptor.Descriptor(
180 name=“Int32Tensor”,
181 full_name=“aialgs.data.Int32Tensor”,
(…)
250 serialized_end=247,
251 )
File /opt/conda/lib/python3.8/site-packages/google/protobuf/descriptor.py:561, in FieldDescriptor.new(cls, name, full_name, index, number, type, cpp_type, label, default_value, message_type, enum_type, containing_type, is_extension, extension_scope, options, serialized_options, has_default_value, containing_oneof, json_name, file, create_key)
555 def new(cls, name, full_name, index, number, type, cpp_type, label,
556 default_value, message_type, enum_type, containing_type,
557 is_extension, extension_scope, options=None,
558 serialized_options=None,
559 has_default_value=True, containing_oneof=None, json_name=None,
560 file=None, create_key=None): # pylint: disable=redefined-builtin
→ 561 _message.Message._CheckCalledFromGeneratedFile()
562 if is_extension:
563 return _message.default_pool.FindExtensionByName(full_name)
TypeError: Descriptors cannot not be created directly.
If this call came from a _pb2.py file, your generated code is out of date and must be regenerated with protoc >= 3.19.0.
If you cannot immediately regenerate your protos, some other possible workarounds are:
More information: Changes announced May 6, 2022 | Protocol Buffers Documentation
Would appreciate your insights.
Thank you !
After testing different combinations of the packages, I completed the C2W1 lab with the following updates:
#!pip install --disable-pip-version-check -q sagemaker==2.173.0
#!conda install -q -y pytorch==1.10.0 -c pytorch
#!pip install --disable-pip-version-check -q transformers==4.27.4
sklearn ‘1.2-1’
And made the corresponding version updates in prepare_data.py.
Using the ‘Data Science 3’ image with Python 3.
Thanks for the help !
Some correction, it should be:
!conda install -q -y pytorch==2.0.1 -c pytorch
Thanks for the correction. Delighted to find your problem got solved.
Congratulations for completing the assignment!
Regards,
Harsh Der.
I got this to work after a number of fixes. I really suggest you update this course, as many things are not working out of the box (at least for me).
Thank you for your feedback @cwestegren.
If i can support in any way, i would be happy to do so! 
Just let me know!
