search

PaccMann / paccmann_proteomics

PaccMann models for protein language modeling
MIT License
37 stars 8 forks source link

ValueError: Used --continue_from_checkpoint but no checkpoint was found in --model_name_or_path. #7

Closed 1treu1 closed 2 years ago

1treu1 commented 2 years ago

Good afternoon. I am trying to replicate the task run_seq_clf_script.sh but I get the following error:

image

The model and dataset paths are as follows: image

Do you happen to know what might be wrong with my implementation?

numpy==1.16.1 torch>=1.3 transformers==3.0.2 tokenizers==0.8.0 torchvision==0.5 tensorboardX==2.0 pandas>1.0 loguru==0.5.3 tape_proteins==0.4 seaborn==0.11.0 scikit-learn==0.23.2 scipy==1.5.3 seqeval -e . CUDA 11.0

drugilsberg commented 2 years ago

Hi @1treu1, looking at the script I believe that with the MODEL_NAME you set you do not need to pass MODEL_NAME/TASK_NAME to --model_name_or_path, but it's sufficient to pass MODEL_NAME.

1treu1 commented 2 years ago

I not understand

1treu1 commented 2 years ago

On the other hand I tried to import the dataset with language_modeling.py but I get the following error:

image

Searching a little I found that in these cases it is better to call the whole library, that is to say, change the from to import:

from ..processors.lm_utils import split_into_chunks It would be: import ..processors.lm_utils But it does not recognize the ..

drugilsberg commented 2 years ago

I not understand

Simply remove /${TASK_NAME} from the first argument you pass to the command. If you would use code snippets instead of screenshot it would be easier to help you ;)

drugilsberg commented 2 years ago

On the other hand I tried to import the dataset with language_modeling.py but I get the following error:

image

Searching a little I found that in these cases it is better to call the whole library, that is to say, change the from to import:

from ..processors.lm_utils import split_into_chunks It would be: import ..processors.lm_utils But it does not recognize the ..

This makes me think you did not install the library properly, relative imports (with .) work only inside modules i python. Did you follow the setup from the README.md? It seems not given this error. Please create the conda environment as described here: https://github.com/PaccMann/paccmann_proteomics#installation. Verify the installation by runnig the python interpreter and trying the import:

import paccmann_proteomics
1treu1 commented 2 years ago

Yes, follow the README configuration. I create the environment as described here https://github.com/PaccMann/paccmann_proteomics#installation . Volvi a crear el environment otra vez pero sigue sin funcionar

The error is the next: ImportError: attempted relative import with no known parent package

drugilsberg commented 2 years ago

ear el environment otra vez pero sigue sin funciona

can you share the output of the conda env creation? The error you shared clearly indicates that paccmann_proteomics is not installed in the environment.

1treu1 commented 2 years ago
Yes, of course packages in environment at /home/ubuntu/anaconda3/envs/paccmann_proteomics: Name Version Build Channel
_libgcc_mutex 0.1 conda_forge conda-forge
_openmp_mutex 4.5 1_gnu conda-forge
biopython 1.79 pypi_0 pypi
boto3 1.20.44 pypi_0 pypi
botocore 1.23.44 pypi_0 pypi
ca-certificates 2021.10.8 ha878542_0 conda-forge
certifi 2021.10.8 pypi_0 pypi
charset-normalizer 2.0.10 pypi_0 pypi
click 8.0.3 pypi_0 pypi
cycler 0.11.0 pypi_0 pypi
filelock 3.4.2 pypi_0 pypi
fonttools 4.29.0 pypi_0 pypi
idna 3.3 pypi_0 pypi
importlib-metadata 4.10.1 pypi_0 pypi
jmespath 0.10.0 pypi_0 pypi
joblib 1.1.0 pypi_0 pypi
kiwisolver 1.3.2 pypi_0 pypi
ld_impl_linux-64 2.36.1 hea4e1c9_2 conda-forge
libffi 3.4.2 h7f98852_5 conda-forge
libgcc-ng 11.2.0 h1d223b6_12 conda-forge
libgomp 11.2.0 h1d223b6_12 conda-forge
libnsl 2.0.0 h7f98852_0 conda-forge
libstdcxx-ng 11.2.0 he4da1e4_12 conda-forge
libzlib 1.2.11 h36c2ea0_1013 conda-forge
lmdb 1.3.0 pypi_0 pypi
loguru 0.5.3 pypi_0 pypi
matplotlib 3.5.1 pypi_0 pypi
ncurses 6.3 h9c3ff4c_0 conda-forge
numpy 1.16.1 pypi_0 pypi
openssl 3.0.0 h7f98852_2 conda-forge
paccmann-proteomics 0.0.1 dev_0
packaging 21.3 pypi_0 pypi
pandas 1.3.5 pypi_0 pypi
pillow 9.0.0 pypi_0 pypi
pip 20.2.4 py_0 conda-forge
protobuf 3.19.3 pypi_0 pypi
pyparsing 3.0.7 pypi_0 pypi
python 3.7.12 hf930737_100_cpython conda-forge
python-dateutil 2.8.2 pypi_0 pypi
python_abi 3.7 2_cp37m conda-forge
pytz 2021.3 pypi_0 pypi
readline 8.1 h46c0cb4_0 conda-forge
regex 2022.1.18 pypi_0 pypi
requests 2.27.1 pypi_0 pypi
s3transfer 0.5.0 pypi_0 pypi
sacremoses 0.0.47 pypi_0 pypi
scikit-learn 0.23.2 pypi_0 pypi
scipy 1.5.3 pypi_0 pypi
seaborn 0.11.0 pypi_0 pypi
sentencepiece 0.1.96 pypi_0 pypi
seqeval 1.2.2 pypi_0 pypi
setuptools 60.5.0 py37h89c1867_0 conda-forge
six 1.16.0 pypi_0 pypi
sqlite 3.37.0 h9cd32fc_0 conda-forge
tape-proteins 0.4 pypi_0 pypi
tensorboardx 2.0 pypi_0 pypi
threadpoolctl 3.0.0 pypi_0 pypi
tk 8.6.11 h27826a3_1 conda-forge
tokenizers 0.8.0 pypi_0 pypi
torch 1.10.1 pypi_0 pypi
torchvision 0.5.0 pypi_0 pypi
tqdm 4.62.3 pypi_0 pypi
transformers 3.0.2 pypi_0 pypi
typing-extensions 4.0.1 pypi_0 pypi
urllib3 1.26.8 pypi_0 pypi
wheel 0.37.1 pyhd8ed1ab_0 conda-forge
xz 5.2.5 h516909a_1 conda-forge
zipp 3.7.0 pypi_0 pypi
zlib 1.2.11 h36c2ea0_1013 conda-forge
drugilsberg commented 2 years ago

Thanks a lot, could you paste the output you get when doing the following:

which python

Open python running:

python

Run the following in the python shell:

import paccmann_proteomics
1treu1 commented 2 years ago

Of course.

image

drugilsberg commented 2 years ago

could you try to repeat the import starting the interpreter from another folder? I'm afraid you are currently importing the module from the path and not from the actual site-packages location

1treu1 commented 2 years ago

Done, I did it from desktop Captura de pantalla de 2022-02-01 17-11-59

drugilsberg commented 2 years ago

ok then I really do not understand what is going wrong with the relative imports error you reported above. can you try to run from the repo root:

python paccmann_proteomics/run_language_modeling.py --help
1treu1 commented 2 years ago

Of course.

image

image

image

1treu1 commented 2 years ago

I solved the problem by changing "from ..processors.lm_utils import split_into_chunks" por "from paccmann_proteomics.data.processors.lm_utils import split_into_chunks" without quotation marks

drugilsberg commented 2 years ago

I solved the problem by changing "from ..processors.lm_utils import split_into_chunks" por "from paccmann_proteomics.data.processors.lm_utils import split_into_chunks" without quotation marks

Hi @1treu1 I'm happy you solved it but the fact that you had to adapt the code means that the installation is somehow off. As we tested you should be able to run scripts leveraging the relative imports. Anyhow happy you can use our code in your research.