diff --git a/DGM_outer.py b/DGM_outer.py
index 2c87f4b..2848af6 100644
--- a/DGM_outer.py
+++ b/DGM_outer.py
@@ -220,7 +220,7 @@ def get_full_eval_threshold(output_dir, archive):
def main():
parser = argparse.ArgumentParser(description="Darwin Godel Machine!")
- parser.add_argument("--max_generation", type=int, default=80, help="Maximum number of evolution iterations.")
+ parser.add_argument("--max_generation", type=int, default=4, help="Maximum number of evolution iterations.")
parser.add_argument("--selfimprove_size", type=int, default=2, help="Number of self-improvements attempts per DGM generation.")
parser.add_argument("--selfimprove_workers", type=int, default=2, help="Number of parallel workers for self-improvement attempts.")
parser.add_argument(
diff --git a/README.md b/README.md
index 9fd9250..a3b2d34 100644
--- a/README.md
+++ b/README.md
@@ -1,3 +1,6 @@
+***In-progress adaptation of DGM research. Original repo/paper found below.***
+
+
Darwin Gödel Machine:
Open-Ended Evolution of Self-Improving Agents
@@ -59,7 +62,7 @@ cd ../../
# Prepare Polyglot
# Make sure git is properly configured in your environment with username and email
-python polyglot/prepare_polyglot_dataset.py
+python -m polyglot.prepare_polyglot_dataset
```
## Running the DGM
diff --git a/llm.py b/llm.py
index d6b4da1..23f467f 100644
--- a/llm.py
+++ b/llm.py
@@ -16,10 +16,11 @@
"gpt-4o-mini-2024-07-18",
"gpt-4o-2024-05-13",
"gpt-4o-2024-08-06",
- "o1-preview-2024-09-12",
+ "o4-mini-2025-04-16",
+ #"o1-preview-2024-09-12",
"o1-mini-2024-09-12",
- "o1-2024-12-17",
- "o3-mini-2025-01-31",
+ #"o1-2024-12-17",
+ #"o3-mini-2025-01-31",
# OpenRouter models
"llama3.1-405b",
# Anthropic Claude models via Amazon Bedrock
@@ -49,6 +50,8 @@ def create_client(model: str):
Returns:
Tuple[Any, str]: A tuple containing the client instance and the client model name.
"""
+ return openai.OpenAI(), "o1-mini-2024-09-12"
+
if model.startswith("claude-"):
print(f"Using Anthropic API with model {model}.")
return anthropic.Anthropic(), model
diff --git a/prompts/self_improvement_prompt.py b/prompts/self_improvement_prompt.py
index eb96087..0a95827 100644
--- a/prompts/self_improvement_prompt.py
+++ b/prompts/self_improvement_prompt.py
@@ -314,8 +314,8 @@ def get_diagnose_prompt_swe(entry_id, commit, root_dir, out_dir, dataset, patch_
diagnose_prompt_out = diagnose_prompt_contextlength
else:
# Get user prompt for the entry
- md_logs, eval_logs, predicted_patches = find_selfimprove_eval_logs(entry_id, out_dir, commit_id=commit)
- md_log, eval_log, predicted_patch = process_selfimprove_eval_logs(md_logs, eval_logs, predicted_patches)
+ md_logs, eval_logs, predicted_patches, eval_results = find_selfimprove_eval_logs(entry_id, out_dir, commit_id=commit)
+ md_log, eval_log, predicted_patch, eval_result = process_selfimprove_eval_logs(md_logs, eval_logs, predicted_patches, eval_results)
entry = next((e for e in dataset if e['instance_id'] == entry_id), None)
answer_patch = entry['patch']
test_patch = entry['test_patch']