bug fix

2024-01-16 18:38:26 -05:00
parent a5dc8b53b7
commit ed48093792
1 changed files with 1 additions and 1 deletions
@@ -76,7 +76,7 @@ class vLLMEngine:
        return total_sequences

    def _get_quantization(self):
-        quantization = os.getenv("QUANTIZATION").lower()
+        quantization = os.getenv("QUANTIZATION", "").lower()
        return quantization if quantization in ["awq", "squeezellm", "gptq"] else None
    
    def concurrency_modifier(self, current_concurrency):