From d94908f9d73647df259aa5021495556671097072 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E9=9B=BE=E8=81=AA?= Date: Sat, 1 Mar 2025 19:12:13 +0800 Subject: [PATCH] update func inference_vllm --- cosyvoice/llm/llm.py | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/cosyvoice/llm/llm.py b/cosyvoice/llm/llm.py index e5a961c..19d7097 100644 --- a/cosyvoice/llm/llm.py +++ b/cosyvoice/llm/llm.py @@ -536,8 +536,7 @@ class Qwen2LM(TransformerLM): for request_output in request_outputs: if str(request_output.request_id) != str(request_id): continue - - print(f"request output: {request_output}") + # print(f"request output: {request_output}") out_token = list(request_output.outputs[0].token_ids)[-1] if next_fill_index != -1 and len(out_tokens) == next_fill_index: top_ids = self.speech_token_size + 2 @@ -575,7 +574,7 @@ class Qwen2LM(TransformerLM): for request_output in request_outputs: if str(request_output.request_id) != str(request_id): continue - print(f"request output: {request_output}") + # print(f"request output: {request_output}") top_ids = list(request_output.outputs[0].token_ids)[-1] out_tokens.append(top_ids) if top_ids >= self.speech_token_size: