TypeError: forward() got an unexpected keyword argument 'return_dict'

[lulmer]

I have an error with SentenceTransformers that throws TypeError: forward() got an unexpected keyword argument 'return_dict'

Here is the code to reproduce :

model_name = 'all-MiniLM-L6-v2'  # I also tried : 'distiluse-base-multilingual-cased-v1'
m= SentenceTransformer(model_name)
m.encode('This is an English sentence') 

Version tested :

• sentence-transformers : 2.2.0 and 2.1.0
• transformers : 4.18.0 and 4.16.2

[nreimers]

Could you share a colab that shows this error? The above code works for me

[lulmer]

Hello Nils,
I couldn't reproduce the bug on colab neither.

To be more precise, it happens in a conda environnent when installing the packages from conda-forge.
Can you try by creating a small test conda environment ?

conda create --name crashtest python=3.8
conda install -c conda-forge sentence-transformers
conda activate crashtest
python3

and running the lines in the python terminal :

from sentence_transformers import SentenceTransformer
model_name = 'all-MiniLM-L6-v2'  # I also tried : 'distiluse-base-multilingual-cased-v1'
m= SentenceTransformer(model_name)
m.encode('This is an English sentence') 

[lulmer]

Here is the full detail of the libraries used in such environment :

# packages in environment at /blablabla/envs/crashtest:
#
# Name                    Version                   Build  Channel
_libgcc_mutex             0.1                        main  
_openmp_mutex             4.5                       1_gnu  
blas                      1.0                         mkl  
brotlipy                  0.7.0           py38h497a2fe_1001    conda-forge
ca-certificates           2021.10.8            ha878542_0    conda-forge
certifi                   2021.10.8        py38h578d9bd_2    conda-forge
cffi                      1.15.0           py38hd667e15_1  
charset-normalizer        2.0.12             pyhd8ed1ab_0    conda-forge
click                     8.1.3            py38h578d9bd_0    conda-forge
colorama                  0.4.4              pyh9f0ad1d_0    conda-forge
cryptography              35.0.0           py38ha5dfef3_0    conda-forge
filelock                  3.6.0              pyhd8ed1ab_0    conda-forge
freetype                  2.10.4               h0708190_1    conda-forge
future                    0.18.2           py38h578d9bd_5    conda-forge
huggingface_hub           0.6.0              pyhd8ed1ab_0    conda-forge
idna                      3.3                pyhd8ed1ab_0    conda-forge
importlib-metadata        4.11.3           py38h578d9bd_1    conda-forge
importlib_metadata        4.11.3               hd8ed1ab_1    conda-forge
intel-openmp              2021.4.0          h06a4308_3561  
joblib                    1.1.0              pyhd8ed1ab_0    conda-forge
jpeg                      9e                   h7f8727e_0  
ld_impl_linux-64          2.35.1               h7274673_9  
libffi                    3.3                  he6710b0_2  
libgcc-ng                 9.3.0               h5101ec6_17  
libgfortran-ng            7.5.0               h14aa051_20    conda-forge
libgfortran4              7.5.0               h14aa051_20    conda-forge
libgomp                   9.3.0               h5101ec6_17  
libpng                    1.6.37               h21135ba_2    conda-forge
libstdcxx-ng              9.3.0               hd4cf53a_17  
libtiff                   4.0.10            hc3755c2_1005    conda-forge
lz4-c                     1.9.3                h9c3ff4c_1    conda-forge
mkl                       2021.4.0           h06a4308_640  
mkl-service               2.4.0            py38h497a2fe_0    conda-forge
mkl_fft                   1.3.1            py38hd3c417c_0  
mkl_random                1.2.2            py38h1abd341_0    conda-forge
ncurses                   6.3                  h7f8727e_2  
ninja                     1.10.2               h4bd325d_0    conda-forge
nltk                      3.6.7              pyhd8ed1ab_0    conda-forge
numpy                     1.21.5           py38he7a7128_2  
numpy-base                1.21.5           py38hf524024_2  
olefile                   0.46               pyh9f0ad1d_1    conda-forge
openssl                   1.1.1o               h7f8727e_0  
packaging                 21.3               pyhd8ed1ab_0    conda-forge
pillow                    6.2.1            py38h6b7be26_0    conda-forge
pip                       21.2.4           py38h06a4308_0  
pycparser                 2.21               pyhd8ed1ab_0    conda-forge
pyopenssl                 22.0.0             pyhd8ed1ab_0    conda-forge
pyparsing                 3.0.9              pyhd8ed1ab_0    conda-forge
pysocks                   1.7.1            py38h578d9bd_5    conda-forge
python                    3.8.13               h12debd9_0  
python_abi                3.8                      2_cp38    conda-forge
pytorch                   1.10.2          cpu_py38hfa7516b_0  
pyyaml                    5.4.1            py38h497a2fe_0    conda-forge
readline                  8.1.2                h7f8727e_1  
regex                     2022.3.15        py38h7f8727e_0  
requests                  2.27.1             pyhd8ed1ab_0    conda-forge
sacremoses                0.0.53             pyhd8ed1ab_0    conda-forge
scikit-learn              1.0.2            py38h51133e4_1  
scipy                     1.7.3            py38hc147768_0  
sentence-transformers     2.2.0              pyhd8ed1ab_0    conda-forge
sentencepiece             0.1.95           py38hd09550d_0  
setuptools                61.2.0           py38h06a4308_0  
six                       1.16.0             pyh6c4a22f_0    conda-forge
sqlite                    3.38.3               hc218d9a_0  
threadpoolctl             3.1.0              pyh8a188c0_0    conda-forge
tk                        8.6.11               h1ccaba5_1  
tokenizers                0.11.4           py38h3dcd8bd_1  
torchvision               0.9.1           py38h3f3a366_1_cpu    conda-forge
tqdm                      4.64.0             pyhd8ed1ab_0    conda-forge
transformers              4.18.0           py38h06a4308_0  
typing-extensions         4.2.0                hd8ed1ab_1    conda-forge
typing_extensions         4.2.0              pyha770c72_1    conda-forge
urllib3                   1.26.9             pyhd8ed1ab_0    conda-forge
wheel                     0.37.1             pyhd3eb1b0_0  
xz                        5.2.5                h7f8727e_1  
yaml                      0.2.5                h516909a_0    conda-forge
zipp                      3.8.0              pyhd8ed1ab_0    conda-forge
zlib                      1.2.12               h7f8727e_2  
zstd                      1.4.9                ha95c52a_0    conda-forge

[nreimers]

I ran this on my Windows machine using Anaconda and it works, giving me back the embeddings.

Maybe re-installing your conda? Or maybe installing sentence transformers via pip?

[lulmer]

Ok I finally found a workaround, I share it here so it can help someone that could have the same problem.
The problem was caused by conda using some packages from /home/my_username/.local/lib/python3.8/site-packages/ instead of /home/my_username/miniconda3/envs/crashtest/lib/python3.8/site-packages/.
as you can see in the full error traceback:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/my_username/miniconda3/envs/crashtest/lib/python3.8/site-packages/sentence_transformers/SentenceTransformer.py", line 164, in encode
    out_features = self.forward(features)
  File "/home/my_username/.local/lib/python3.8/site-packages/torch/nn/modules/container.py", line 100, in forward
    input = module(input)
  File "/home/my_username/.local/lib/python3.8/site-packages/torch/nn/modules/module.py", line 550, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/my_username/miniconda3/envs/crashtest/lib/python3.8/site-packages/sentence_transformers/models/Transformer.py", line 66, in forward
    output_states = self.auto_model(**trans_features, return_dict=False)
  File "/home/my_username/.local/lib/python3.8/site-packages/torch/nn/modules/module.py", line 550, in __call__
    result = self.forward(*input, **kwargs)
TypeError: forward() got an unexpected keyword argument 'return_dict'

Maybe I had an old pytorch version installed in my /home/my_username/.local/lib/python3.8/site-packages/ or I don't know?

The following command solved my problem:
conda env config vars set PYTHONNOUSERSITE=1 -n crashtest
The line can be found here

Thank you anyway @nreimers, sorry for disturbing :)