Initial

README.md

@@ -45,36 +45,11 @@ VEGA_AE
 - pip install -r requirements.txt
 
 
-## 4. Fine-Tuning
-
-**Run fine-tuning with:**
-```
-$ bash run_fine_tuning.sh
-```
-
-Customize parameters for fine-tuning by modifying following options in the ```run_fine_tuning.sh```.
-```
-  --model_name_or_path ../../saved_models/UnixCoder \
-  --train_filename ../../dataset/train.jsonl \
-  --dev_filename ../../dataset/valid.jsonl \
-  --output_dir ../../saved_models/New_Fine_Tuned_Model \
-  --beam_size 4 \
-  --train_batch_size 96 \
-  --eval_batch_size 80 \
-  --learning_rate 6e-5 \
-  --num_train_epochs 50 \
-  --mse_loss_weight 0.9 \
-  --ce_loss_weight 0.1
-```
-The fine-tuned model will be saved in ```--output_dir```.
-
-## 5. Code Generation
+## 4. Code Generation
 
 We have provided a fine-tuned model in ```./saved_models/Fine_Tuned_Model```.
 
-We have also provided a script fot functionality test, which only generates a single function for RI5CY, taking less than 3 minutes with 8 V100 GPU.
-
-
+We have also provided a script fot functionality test, which only generates a single function for RI5CY (Recorded as PULP in our dataset), taking less than 3 minutes with 8 V100 GPU.
 
 **Run functionality test with:**
 
@@ -130,6 +105,32 @@ The inference result will be saved in ```--output_dir/result.jsonl```.
 
 The ```--output_dir``` parameter specifies the directory where the fine-tuned model is saved, such as ```./saved_models/Fine_Tuned_Model```.
 
+
+
+## 5. Fine-Tuning (**Optional**)
+
+**Run fine-tuning with:**
+```
+$ bash run_fine_tuning.sh
+```
+
+Customize parameters for fine-tuning by modifying following options in the ```run_fine_tuning.sh```.
+```
+  --model_name_or_path ../../saved_models/UnixCoder \
+  --train_filename ../../dataset/train.jsonl \
+  --dev_filename ../../dataset/valid.jsonl \
+  --output_dir ../../saved_models/New_Fine_Tuned_Model \
+  --beam_size 4 \
+  --train_batch_size 96 \
+  --eval_batch_size 80 \
+  --learning_rate 6e-5 \
+  --num_train_epochs 50 \
+  --mse_loss_weight 0.9 \
+  --ce_loss_weight 0.1
+```
+The fine-tuned model will be saved in ```--output_dir```.
+
+
 ## 6. Reproducing Results in the Experiment
 
 We provide the scripts to reproduce each Figure/Table from the paper, along with the corresponding output result files, in the following table:

run_fine_tuning.sh

@@ -7,10 +7,9 @@ python ./Scripts/UnixCoder/run_one_model.py \
   --dev_filename ../../dataset/valid.jsonl \
   --output_dir ../../saved_models/New_Fine_Tuned_Model \
   --beam_size 4 \
-  --train_batch_size 96 \
-  --eval_batch_size 80 \
+  --train_batch_size 64 \
+  --eval_batch_size 48 \
   --learning_rate 6e-5 \
   --num_train_epochs 50 \
   --mse_loss_weight 0.9 \
-   --ce_loss_weight 0.1
-~                             
+  --ce_loss_weight 0.1