ImportError: cannot import name 'TransformerEncoderLayer' from 'fairseq.modules

Thank you for your work and your attentive answers to every question.

The problem occurs when I run sh train_caption_stage1.sh . Something went wrong with fairseq

Traceback (most recent call last): File “…/…/train.py”, line 29, in <module> from fairseq import ( File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/quantization_utils.py”, line 8, in <module> from fairseq.modules.quantization import pq, quantization_options, scalar File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/modules/init.py”, line 39, in <module> from .transformer_layer import TransformerDecoderLayer, TransformerEncoderLayer File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/modules/transformer_layer.py”, line 15, in <module> from fairseq.models.transformer import ( File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/init.py”, line 236, in <module> import_models(models_dir, “fairseq.models”) File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/init.py”, line 218, in import_models importlib.import_module(namespace + “.” + model_name) File “/home/carlos/anaconda3/envs/GenerativeImage2Text/lib/python3.7/importlib/init.py”, line 127, in import_module Traceback (most recent call last): return _bootstrap._gcd_import(name[level:], package, level) File “…/…/train.py”, line 29, in <module>

File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/speech_to_text/init.py”, line 7, in <module> from .convtransformer import * # noqa File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/speech_to_text/convtransformer.py”, line 19, in <module> from fairseq import (
from fairseq.modules import LayerNorm, PositionalEmbedding, TransformerEncoderLayer File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/quantization_utils.py”, line 8, in <module> ImportError: cannot import name ‘TransformerEncoderLayer’ from ‘fairseq.modules’ (/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/modules/init.py) from fairseq.modules.quantization import pq, quantization_options, scalar File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/modules/init.py”, line 39, in <module> from .transformer_layer import TransformerDecoderLayer, TransformerEncoderLayer File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/modules/transformer_layer.py”, line 15, in <module> from fairseq.models.transformer import ( File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/init.py”, line 236, in <module> import_models(models_dir, “fairseq.models”) File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/init.py”, line 218, in import_models importlib.import_module(namespace + “.” + model_name) File “/home/carlos/anaconda3/envs/GenerativeImage2Text/lib/python3.7/importlib/init.py”, line 127, in import_module return _bootstrap._gcd_import(name[level:], package, level) File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/speech_to_text/init.py”, line 7, in <module> from .convtransformer import * # noqa File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/speech_to_text/convtransformer.py”, line 19, in <module> from fairseq.modules import LayerNorm, PositionalEmbedding, TransformerEncoderLayer ImportError: cannot import name ‘TransformerEncoderLayer’ from ‘fairseq.modules’ (/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/modules/init.py) Traceback (most recent call last): File “…/…/train.py”, line 29, in <module> from fairseq import ( File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/quantization_utils.py”, line 8, in <module> from fairseq.modules.quantization import pq, quantization_options, scalar File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/modules/init.py”, line 39, in <module> from .transformer_layer import TransformerDecoderLayer, TransformerEncoderLayer File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/modules/transformer_layer.py”, line 15, in <module> from fairseq.models.transformer import ( File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/init.py”, line 236, in <module> import_models(models_dir, “fairseq.models”) File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/init.py”, line 218, in import_models importlib.import_module(namespace + “.” + model_name) File “/home/carlos/anaconda3/envs/GenerativeImage2Text/lib/python3.7/importlib/init.py”, line 127, in import_module return _bootstrap._gcd_import(name[level:], package, level) File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/speech_to_text/init.py”, line 7, in <module> from .convtransformer import * # noqa File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/speech_to_text/convtransformer.py”, line 19, in <module> from fairseq.modules import LayerNorm, PositionalEmbedding, TransformerEncoderLayer ImportError: cannot import name ‘TransformerEncoderLayer’ from ‘fairseq.modules’ (/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/modules/init.py) Traceback (most recent call last): File “…/…/train.py”, line 29, in <module> from fairseq import ( File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/quantization_utils.py”, line 8, in <module> from fairseq.modules.quantization import pq, quantization_options, scalar File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/modules/init.py”, line 39, in <module> from .transformer_layer import TransformerDecoderLayer, TransformerEncoderLayer File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/modules/transformer_layer.py”, line 15, in <module> from fairseq.models.transformer import ( File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/init.py”, line 236, in <module> import_models(models_dir, “fairseq.models”) File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/init.py”, line 218, in import_models importlib.import_module(namespace + “.” + model_name) File “/home/carlos/anaconda3/envs/GenerativeImage2Text/lib/python3.7/importlib/init.py”, line 127, in import_module return _bootstrap._gcd_import(name[level:], package, level) File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/speech_to_text/init.py”, line 7, in <module> from .convtransformer import * # noqa File “/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/models/speech_to_text/convtransformer.py”, line 19, in <module> from fairseq.modules import LayerNorm, PositionalEmbedding, TransformerEncoderLayer ImportError: cannot import name ‘TransformerEncoderLayer’ from ‘fairseq.modules’ (/home/carlos/cv_project/OFA/OFA/fairseq/fairseq/modules/init.py) ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 0 (pid: 16406) of binary: /home/carlos/anaconda3/envs/GenerativeImage2Text/bin/python3 Traceback (most recent call last): File “/home/carlos/anaconda3/envs/GenerativeImage2Text/lib/python3.7/runpy.py”, line 193, in _run_module_as_main “main”, mod_spec) File “/home/carlos/anaconda3/envs/GenerativeImage2Text/lib/python3.7/runpy.py”, line 85, in _run_code exec(code, run_globals) File “/home/carlos/anaconda3/envs/GenerativeImage2Text/lib/python3.7/site-packages/torch/distributed/launch.py”, line 193, in <module> main() File “/home/carlos/anaconda3/envs/GenerativeImage2Text/lib/python3.7/site-packages/torch/distributed/launch.py”, line 189, in main launch(args) File “/home/carlos/anaconda3/envs/GenerativeImage2Text/lib/python3.7/site-packages/torch/distributed/launch.py”, line 174, in launch run(args) File “/home/carlos/anaconda3/envs/GenerativeImage2Text/lib/python3.7/site-packages/torch/distributed/run.py”, line 755, in run )(*cmd_args) File “/home/carlos/anaconda3/envs/GenerativeImage2Text/lib/python3.7/site-packages/torch/distributed/launcher/api.py”, line 131, in call return launch_agent(self._config, self._entrypoint, list(args)) File “/home/carlos/anaconda3/envs/GenerativeImage2Text/lib/python3.7/site-packages/torch/distributed/launcher/api.py”, line 247, in launch_agent failures=result.failures, torch.distributed.elastic.multiprocessing.errors.ChildFailedError:

[Carlos25130]

Will this issue help #217 ? Thank you very much for your answer, I refer to #217, reduced the pip version and reinstalled facirseq, which successfully solved the problem. My previous pip version was 22.2.2 and I downgraded to 21.2.4.

[yangapku]

Could you please share the pip (pip list) and conda (conda list) info for us to have a check?