run_classifier.py invalid arguments

NVIDIA / sentiment-discovery

Unsupervised Language Modeling at scale for robust sentiment classification

Other

1.06k stars 202 forks source link

run_classifier.py invalid arguments #72

Open crazylazylife opened 4 years ago

crazylazylife commented 4 years ago

I downloaded the sst_clf_16.pt file to perform classification from a csv file. Thw code I wrote:

python run_classifier.py --load_model sst_clf_16.pt --fp16 --data upload_processed_tweet.csv --text-key text --write-results output.csv

The error I got was that the --load_model is an "unrecognized argument":

run_classifier.py: error: unrecognized arguments: --load_model sst_clf_16.pt

krishh-konar commented 4 years ago

try using --load instead of --load_model.

usage: run_classifier.py [-h] [--model MODEL] [--lr LR]
                         [--constant-decay CONSTANT_DECAY] [--clip CLIP]
                         [--epochs EPOCHS] [--tied] [--seed SEED]
                         [--log-interval N] [--save SAVE] [--load LOAD]

Saumajit commented 4 years ago

@krishh-konar I have tried to run_classifier on semeval data for prediction. I have used --data, --load, --text-key and --write-results as the arguments in the command line for running the run_classifier.py It is showing me /bin/bash: line 113 segmentation fault (core dumped) Can you please provide the command line in detail?