search

naver / biobert-pretrained

BioBERT: a pre-trained biomedical language representation model for biomedical text mining
667 stars 88 forks source link

Failed to find any matching files for biobert-pretrained/biobert_v1.1_pubmed/biobert_model.ckpt #8

Closed votamvan closed 5 years ago

votamvan commented 5 years ago

Hi biobert, I got the error of not finding of file biobert_v1.1_pubmed/biobert_model.ckpt when I run the below command as written in README.md. python run_ner.py \ --do_train=true \ --do_eval=true \ --vocab_file=$BIOBERT_DIR/vocab.txt \ --bert_config_file=$BIOBERT_DIR/bert_config.json \ --init_checkpoint=$BIOBERT_DIR/biobert_model.ckpt \ --num_train_epochs=10.0 \ --data_dir=$NER_DIR/ \ --output_dir=/tmp/bioner/ Would you please provide me the file biobert_model.ckpt or tell me how to fix the issue, thanks.

jhyuklee commented 5 years ago

Hi @votamvan, The path might be wrong if you didn't set env variable $BIOBERT_DIR. Please check your path is properly set ($BIOBERT_DIR), and make sure that your files are there. Thank you.

votamvan commented 5 years ago

Hi @jhyuklee,

Thank you very much for your quick reply. I understand the variable $BIOBERT_DIR and already set it in my bash shell. The problem is about the missing file . After unpacking file biobert_v1.1_pubmed.tar.gz. I cannot find it anywhere. I enclose here content of unpacked folder for your review. I guess is forgot to attach in the package or do I have to run some script to create the file . Please help me, thanks.

biobert_v1.1_pubmed xxxxx$ pwd /Users/xxxxx/biobert-pretrained/biobert_v1.1_pubmed biobert_v1.1_pubmed xxxxx$ ls -l total 871944 -rw-r--r--@ 1 xxxxx staff 947465 Feb 26 15:25 model.ckpt-1000000.meta -rw-r--r--@ 1 xxxxx staff 8224 Feb 26 15:25 model.ckpt-1000000.index -rw-r--r--@ 1 xxxxx staff 433241088 Feb 26 15:25 model.ckpt-1000000.data-00000-of-00001 -rw-r--r--@ 1 xxxxx staff 313 May 16 15:47 bert_config.json -rw-r--r--@ 1 xxxxx staff 213450 May 16 15:47 vocab.txt

jhyuklee commented 5 years ago

For that version, please try init_checkpoint=$BIOBERT_DIR/model.ckpt. Sorry for the inconsistency. We will update the readme.

vnnp commented 4 years ago

Hi @votamvan, In this case, it worked for me only with init_checkpoint=$BIOBERT_DIR/model.ckpt-1000000.index

v-loves-avocados commented 4 years ago

I had the same problem. There was not a model.ckpt fille or biobert_model.ckpt. Instead, there are these three ckpt files:

model.ckpt-1000000.data-00000-of-00001 model.ckpt-1000000.index model.ckpt-1000000.meta

Thank you for your input.

wonjininfo commented 4 years ago

Hi @v-loves-avocados , You need to change parameter --init_checkpoint= according to the file name. Please ignore extensions(?) such as .meta or .data~~ and use --init_checkpoint=$BIOBERT_DIR/model.ckpt-1000000 instead. Thanks!

v-loves-avocados commented 4 years ago

@wonjininfo Thank you! It works! Successfully replicated the Relation Extraction results.