Copy and Run Alphageometry

[greentree327]

https://github.com/google-deepmind/alphageometry

I would like to copy and run this github repository on Ubuntu / google colab / kaggle, but I faced error every time when I Run AlphaGeometry, which is defined in the github and below.

python -m alphageometry
--alsologtostderr
--problems_file＝$(pwd)/examples.txt
--problem_name＝orthocenter
--mode＝alphageometry
″${DDAR_ARGS[@]}″
″${SEARCH_ARGS[@]}″
″${LM_ARGS[@]}″

Can anyone share their full, successful running code?

[HeikeKimoyori]

Same with me

[HeikeKimoyori]

lm.LanguageModelInference(vocab_path, ckpt_init, mode='beam_search')
File "/home/ubuntu/NSL-KDD/Data5et-master/N5L-KDD/alphageometry-main/alphageometry-main/1m_inference,py", line 81, in init for idx in self.encode_list(['.'，';']):
File "/home/ubuntu/NSL-KDD/Data5et-master/N5L-KDD/alphageometry-main/alphageometry-main/1m_inference,py", line 97, in encode_list result= [self.vocab.encode(x) for x in inputs_strs]
File "/home/ubuntu/NSL-KDD/Data5et-master/N5L-KDD/alphageometry-main/alphageometry-main/1m_inference,py", line 97, in result = [self.vocab.encode(x) for x in inputs_strs]
File "/home/ubuntu/NSL-KDD/Data5et-master/N5L-KDD/alphageometry-main/alphageometry-main/lib/python3,10/site-packages/seqio/vocabularies,py", line 96, in encode return self._encode(s)
File "/home/ubuntu/NSL-KDD/Data5et-master/N5L-KDD/alphageometry-main/alphageometry-main/1ib/python3,10/site-packages/seqio/vocabularies,py", line 455, in _encode return self.tokenizer.EncodeAsIds(s)
File "/home/ubuntu/NSL-KDD/Data5et-master/N5L-KDD/alphageometry-main/alphageometry-main/1ib/python3,10/site-packages/seqio/vocabularies,py", line 426, in tokenizer return self. model context().tokenizer
File "/home/ubuntu/NSL-KDD/Data5et-master/N5L-KDD/alphageometry-main/alphageometry-main/1ib/python3,10/site-packages/seqio/vocabularies,py", line 336,in _model_context self._model = self._load_model(
File "/home/ubuntu/NSL-KDD/Data5et-master/N5L-KDD/alphageometry-main/alphageometry-main/1ib/python3,10/site-packages/seqio/vocabularies,py", line 361, in _load_model sp_model = f.read()
File "/home/ubuntu/NSL-KDD/Data5et-master/N5L-KDD/alphageometry-main/alphageometry-main/1ib/python3,10/site-packages/tensorflow/python/lib/io/file_io,py", line 116, in read self._preread_check()
File"/home/ubuntu/NSL-KDD/Data5et-master/NSL-KDD/alphageometry-main/alphageometry-main/1ib/python3,10/site-packages/tensorflow/python/1ib/io/file_io,py", line 77, in preread check self._read_buf = _pywrap_file_io.BufferedInputstream(
tensorflow,python, framework,errors_impl,NotFoundError: /geometry,757.model; No such file or directory 0.0
(alphageometry-main) (tf_env) ubuntu@MM-0-12-ubuntu:~/NSL-KDD/DataSet-master/NSL-KDD/alphageometry-main/alphageometry-main$ It is said there is no file named geometry.757.model in the current folder. But I am sure that it exists here...

[tpgh24]

I created a fork of AlphaGeometry and made various improvements, including a Kaggle notebook. It should run on Colab too with some minor modifications, but the free version of Kaggle reliably provides 2 T4 GPUs, while in the free version of Colab, one can almost never get GPUs. Check it out at https://github.com/tpgh24/ag4masses.