Update README.md

whustan · Aug 20, 2019 · f962698 · f962698
1 parent af1a597
commit f962698
Showing 1 changed file with 17 additions and 7 deletions.
diff --git a/README.md b/README.md
@@ -1,17 +1,18 @@
 # LexiconNER
 This is the implementation of "Named Entity Recognition using Positive-Unlabeled Learning" published at ACL2019.
 
-###Set up and run
+### Set up and run
 Download glove.6B.100d.txt
 ### Environment
 pytorch 1.1.0
 python 3.6.4
 cuda 8.0
 ### Instructions for running code
-#### Phrase one <train bnPU model>
+#### Phrase one \<train bnPU model\>
 **Train**
 Print parameters
 `run feature_pu_model.py  --h`
+```html
 optional arguments:
   -h, --help            show this help message and exit
   --lr LR               learning rate
@@ -27,18 +28,24 @@ optional arguments:
                     	epochs for printing result
   --pert PERT           percentage of data use for training
   --type TYPE           pu learning type (bnpu/bpu/upu)
+```
 e.g.)
 Train on PER type of conll2003 dataset:
 `python feature_pu_model.py --dataset conll2003 --type PER`
-** Evaluating**
+** Evaluating**
+```html
 python feature_pu_model_evl.py --model saved_model/bnpu_conll2003_PER_lr_0.0001_prior_0.3_beta_0.0_gamma_1.0_percent_1.0 --flag PER --dataset conll2003 --output 1
+```
 replace the model name from the training
-python final_evl.py
+```html
+python final_evl.py 
+```
 Get the final result on all the entity type. Remember to revise the filenames to be the output file name of evaluating.
 
-#### Phrase two <train adaPU model>
+#### Phrase two \<train adaPU model\>
 **dictionary generation**
 `run python ada_dict_generation.py -h`
+```html
 optional arguments:
   -h, --help            show this help message and exit
   --beta BETA           learning rate
@@ -55,10 +62,12 @@ optional arguments:
                         use unlabeled data or not
   --pert PERT           percentage of data use for training
   --model MODEL         saved model name
+```
 e.g.)
 `python ada_dict_generation.py --model saved_model/bnpu_conll2003_PER_lr_0.0001_prior_0.3_beta_0.0_gamma_1.0_percent_1.0 --flag PER --iter 1`
 **adaptive training**
-`run python adaptive_pu_model.py -h
+`run python adaptivepumodel.py -h 
+````html
 optional arguments:
   -h, --help            show this help message and exit
   --beta BETA           beta of pu learning (default 0.0)
@@ -75,6 +84,7 @@ optional arguments:
                         file
   --model MODEL         saved model name
   --iter ITER           iteration time
+```
 e.g.)
 `python adaptive\_pu\_model.py --model saved\_model/bnpu\_conll2003\_PER\_lr\_0.0001\_prior\_0.3\_beta\_0.0\_gamma\_1.0\_percent\_1.0 --flag PER --iter 1`
-Replace saved model names and iteration times when doing adaptive learning. And in the same iteration the iter number in dictionary generation and adaptive learning should be same.
+Replace saved model names and iteration times when doing adaptive learning. And in the same iteration the iter number in dictionary generation and adaptive learning should be same.