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]

Enh(models): Handle local LM context length errors

Kilian Liereteee7181942e1
diff --git a/sweagent/agent/models.py b/sweagent/agent/models.py
index 98061860..c6390d51 100644
--- a/sweagent/agent/models.py
+++ b/sweagent/agent/models.py
@@ -639,6 +639,10 @@ class LiteLLMModel(AbstractModel):
raise ContextWindowExceededError from e
except litellm.exceptions.ContentPolicyViolationError as e:
raise ContentPolicyViolationError from e
+ except litellm.exceptions.BadRequestError as e:
+ if "is longer than the model's context length" in str(e):
+ raise ContextWindowExceededError from e
+ raise
self.logger.info(f"Response: {response}")
cost = litellm.cost_calculator.completion_cost(response)
choices: litellm.types.utils.Choices = response.choices # type: ignore