Update app.py

app.py

@@ -13,7 +13,7 @@ os.environ["RWKV_CUDA_ON"] = '1' # if '1' then use CUDA kernel for seq mode (muc
 
 from rwkv.model import RWKV
 model_path = hf_hub_download(repo_id="BlinkDL/rwkv-5-world", filename=f"{title}.pth")
-model = RWKV(model=model_path, strategy='cuda fp16')
+model = RWKV(model=model_path, strategy='cuda fp16i8 *8 -> cuda fp16')
 from rwkv.utils import PIPELINE, PIPELINE_ARGS
 pipeline = PIPELINE(model, "rwkv_vocab_v20230424")