openreasoner · wangst0181 · Dec 10, 2024
diff --git a/benchmark/plots.md b/benchmark/plots.md
diff --git a/benchmark/tables.md b/benchmark/tables.md
diff --git a/data/README.md b/data/README.md
@@ -4,30 +4,6 @@ A simple implementation of data generation of [Improve Mathematical Reasoning in
 
 ## Generating data
 
-Edit the `config.yaml` file to set the model and other parameters. Input json file should be a list of problems and answers with key `problem` and `final_answer`.
-
-```yaml
-input:
-  json_file_path: 'extracted_problems_and_answers.json'
-
-output:
-  file_prefix: 'math'
-  log_file_path: 'processing_log.log'
-
-processing:
-  initial_rollouts: 20
-  num_rollouts: 20
-  max_iterations: 100
-
-model:
-  # supported model_types: "hf", "openai", "anthropic"
-  model_type: "hf"
-  model_name: "Qwen/Qwen2.5-Math-7B-Instruct"
-  model_args:
-    max_tokens: 200
-    temperature_range: [0.7, 1.0]
-```
-
 ```bash
 python gen_data.py
 ```

diff --git a/data/config.yaml b/data/config.yaml