chore: reformat

lib/questions_eval/bash/experiments/tiny_mimoracle.sh

@@ -0,0 +1,3 @@
+python run_mimoracle.py -m model=gpt-4o samples=18 num_questions=5
+# python run_mimoracle.py -m model=gpt-4o-mini samples=18 num_questions=5
+# python run_mimoracle.py -m model=llama3.1-405b-local samples=18 num_questions=5

lib/questions_eval/notebooks/test_mimorable.ipynb

@@ -365,14 +365,6 @@
    "source": [
     "sample_df.to_csv('mimoracle_sample.csv', index=False)"
    ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "a5675a76-ed7b-444e-a1e0-5b0c790969ba",
-   "metadata": {},
-   "outputs": [],
-   "source": []
   }
  ],
  "metadata": {

lib/questions_eval/run_mimoracle.py

@@ -1,17 +1,15 @@
-import itertools
-import pdb
-
 import hydra
-import pandas as pd
 import wandb
+import itertools
+import pandas as pd
+from pandas import json_normalize
+from tqdm import tqdm
 from dotenv import load_dotenv
 from langchain.output_parsers import OutputFixingParser
 from langchain.schema import StrOutputParser
 from langchain_core.output_parsers import JsonOutputParser
 from langchain_core.prompts import ChatPromptTemplate
 from omegaconf import DictConfig, OmegaConf
-from pandas import json_normalize
-from tqdm import tqdm
 
 load_dotenv()
 
@@ -275,7 +273,7 @@ def main(cfg: DictConfig):
     }
     for key, value in log_dict.items():
         wandb.run.summary[key] = value
-    wandb.log({"dataset/evaluation_mimoracle": wandb.Table(dataframe=df_joined)})
+    wandb.log({"dataset/evaluation_mimoracle_gpt4o": wandb.Table(dataframe=df_joined)})
     wandb.finish()