Skip to content

SWE-agent takes a GitHub issue and tries to automatically fix it, using your LM of choice. It can also be employed for offensive cybersecurity or competitive coding challenges. [NeurIPS 2024]

Config: Rename/clarify sbl submission config

Kilian Lieret22669f2a6ab8
diff --git a/config/sweagent_heavy.yaml b/config/sweagent_heavy_sbl_250212.yaml
similarity index 94%
rename from config/sweagent_heavy.yaml
rename to config/sweagent_heavy_sbl_250212.yaml
index bf977032..801f97c2 100644
--- a/config/sweagent_heavy.yaml
+++ b/config/sweagent_heavy_sbl_250212.yaml
@@ -1,3 +1,7 @@
+# Used for our SWE-Bench lite benchmark submission from 12 Feb 2025
+# Used together with swe-agent as
+# sweagent run-batch --num_workers=12 --instances.type=swe_bench --instances.subset=lite --instances.split=test
+# --instances.shuffle=True --instances.evaluate=True --instances.deployment.docker_args=--memory=10g --config config/retry_heavy_v3.yaml --instances.slice :100
# This template is heavily inspired by anthropic's computer use demo
agent:
type: retry
@@ -181,4 +185,4 @@ agent:
temperature: 1.
per_instance_cost_limit: 30
completion_kwargs:
- reasoning_effort: "high"
\ No newline at end of file
+ reasoning_effort: "high"