AngledLuffa
commented
1 year ago when the content in "filename" is empty， the CREC new will computer twice。
Please clarify what you mean by this. sizeof(CREC) will be the size of the C object, not the data file itself.
In terms of git usage, it will be easier to follow newly posted issues if you:
- use ``` to denote code blocks. it will look like this:
asdf {
stuff;
}- link to the lines of code you are interested in. if you click on the line number on github, three dots should show up. one of the options will be
copy permalink. for example
https://github.com/stanfordnlp/GloVe/blob/a6f2b947358bd5a0bdb825af3ebe429ab43a3cb5/src/common.h#L7
in file cooccur.c line 177-184 for (i = 0; i < num; i++) { sprintf(filename,"%s_%04d.bin",file_head,i); fid[i] = fopen(filename,"rb"); if (fid[i] == NULL) {log_file_loading_error("file", filename); free_fid(fid, num); free(pq); return 1;} fread(&new, sizeof(CREC), 1, fid[i]); new.id = i; insert(pq,new,i+1); } when the content in "filename" is empty， the CREC new will computer twice。 let's see the corpus has only the following words: "training the vectors can be speeded up by increasing the training parallelism to match the number of physical CPU cores available" there will generate two bin files, and one of files is empty.