diff --git a/composte-hermes.yml b/composte-hermes.yml index 857c0b3..2454755 100644 --- a/composte-hermes.yml +++ b/composte-hermes.yml @@ -25,7 +25,7 @@ services: - --port=8005 - --quantization=awq - --tensor-parallel-size=2 - - --gpu-memory-utilization=0.84 + - --gpu-memory-utilization=0.86 - --max-model-len=65536 - --hf-overrides={"rope_scaling":{"rope_type":"yarn","factor":2.0,"original_max_position_embeddings":32768}} - --max-num-seqs=4