Fix typos

Author: ZhuochengZhang98

itrg/itrg.py

@@ -14,7 +14,7 @@
 @dataclass
 class ITRGAssistantConfig(GeneratorConfig, GenerationConfig, RetrieverConfig):
     max_iteration: int = 5
-    retrieval_num: int = 5
+    retrieved_num: int = 5
 
 
 @ASSISTANTS("itrg", "iter-retgen", config_class=ITRGAssistantConfig)
@@ -26,7 +26,7 @@ class ITRGAssistant(AssistantBase):
     def __init__(self, cfg: ITRGAssistantConfig) -> None:
         # load retriever
         self.retriever = RETRIEVERS.load(cfg)
-        self.retriever.top_k = cfg.retrieval_num
+        self.retriever.top_k = cfg.retrieved_num
 
         # load generator
         self.generator = GENERATORS.load(cfg)

itrg/readme.md

@@ -8,7 +8,7 @@ This is the reproduction of the paper:
 ITRG & ITER-RETGEN iterate retrieval and generation. In each iteration, they leverages the model output from the previous iteration as a specific context to help retrieve more relevant knowledge, which may help improve model generation.
 
 <center>
-<img src="./image.png" alt="IRCoT" width="50%"/>
+<img src="./image.png" alt="ITRG" width="50%"/>
 </center>
 
 ## Running the Method
@@ -19,11 +19,11 @@ bash ./run_generator.sh
 This script will start a `Qwen2-7B-Instruct` model server on port 8000. You can change the `MODEL_NAME` in the script if you want to use a different model.
 
 
-Then, run the following command to evaluate the IRCoT on the test set of `Natural Questions`:
+Then, run the following command to evaluate the ITRG on the test set of `Natural Questions`:
 ```bash
 bash ./run.sh
 ```
-This script will run the IRCoT method on the test set of `Natural Questions` and save the results in the `results` directory. You can change the `DATASET_NAME` and the `SPLIT` variables in the script to evaluate on different datasets.
+This script will run the ITRG method on the test set of `Natural Questions` and save the results in the `results` directory. You can change the `DATASET_NAME` and the `SPLIT` variables in the script to evaluate on different datasets.
 
 ## Experiments
 

react/readme.md

@@ -7,7 +7,7 @@ This is the reproduction of the paper:
 ReAct solves knowledge-intensive tasks by taking both reasoning and retrieving actions in an interleaved manner, allowing for greater synergy between the two: reasoning traces help the model induce, track, and update action plans as well as handle exceptions, while actions allow it to interface with and gather additional information from external sources such as knowledge bases or environments.
 
 <center>
-<img src="./image.png" alt="IRCoT" width="50%"/>
+<img src="./image.png" alt="ReACT" width="50%"/>
 </center>
 
 ## Running the Method
@@ -18,11 +18,11 @@ bash ./run_generator.sh
 This script will start a `Qwen2-7B-Instruct` model server on port 8000. You can change the `MODEL_NAME` in the script if you want to use a different model.
 
 
-Then, run the following command to evaluate the IRCoT on the test set of `Natural Questions`:
+Then, run the following command to evaluate the ReACT on the test set of `Natural Questions`:
 ```bash
 bash ./run.sh
 ```
-This script will run the IRCoT method on the test set of `Natural Questions` and save the results in the `results` directory. You can change the `DATASET_NAME` and the `SPLIT` variables in the script to evaluate on different datasets.
+This script will run the ReACT method on the test set of `Natural Questions` and save the results in the `results` directory. You can change the `DATASET_NAME` and the `SPLIT` variables in the script to evaluate on different datasets.
 
 ## Citation
 If you use this code in your research, please cite the following paper: