move token limit conf to bridge_all.py

config.py

@@ -91,7 +91,7 @@ AVAIL_LLM_MODELS = ["gpt-3.5-turbo-1106","gpt-4-1106-preview","gpt-4-vision-prev
                     "gpt-3.5-turbo-16k", "gpt-3.5-turbo", "azure-gpt-3.5",
                     "api2d-gpt-3.5-turbo", 'api2d-gpt-3.5-turbo-16k',
                     "gpt-4", "gpt-4-32k", "azure-gpt-4", "api2d-gpt-4",
-                    "chatglm3", "moss", "claude-2", "deepseekcoder"]
+                    "chatglm3", "moss", "claude-2"]
 # P.S. 其他可用的模型还包括 ["zhipuai", "qianfan", "deepseekcoder", "llama2", "qwen", "gpt-3.5-turbo-0613", "gpt-3.5-turbo-16k-0613",  "gpt-3.5-random"
 # "spark", "sparkv2", "sparkv3", "chatglm_onnx", "claude-1-100k", "claude-2", "internlm", "jittorllms_pangualpha", "jittorllms_llama"]
 
@@ -114,9 +114,6 @@ CHATGLM_PTUNING_CHECKPOINT = "" # 例如"/home/hmp/ChatGLM2-6B/ptuning/output/6b
 LOCAL_MODEL_DEVICE = "cpu" # 可选 "cuda"
 LOCAL_MODEL_QUANT = "FP16" # 默认 "FP16" "INT4" 启用量化INT4版本 "INT8" 启用量化INT8版本
 
-# 设置deepseekcoder运行时输入的最大token数（超过4096没有意义），对话过程爆显存可以适当调小
-MAX_INPUT_TOKEN_LENGTH = 2048
-
 # 设置gradio的并行线程数（不需要修改）
 CONCURRENT_COUNT = 100
 

request_llms/bridge_all.py

@@ -552,7 +552,7 @@ if "deepseekcoder" in AVAIL_LLM_MODELS:   # deepseekcoder
                 "fn_with_ui": deepseekcoder_ui,
                 "fn_without_ui": deepseekcoder_noui,
                 "endpoint": None,
-                "max_token": 4096,
+                "max_token": 2048,
                 "tokenizer": tokenizer_gpt35,
                 "token_cnt": get_token_num_gpt35,
             }

request_llms/bridge_deepseekcoder.py

@@ -8,7 +8,6 @@ from .local_llm_class import LocalLLMHandle, get_local_llm_predict_fns
 from threading import Thread
 import torch
 
-MAX_INPUT_TOKEN_LENGTH = get_conf("MAX_INPUT_TOKEN_LENGTH")
 def download_huggingface_model(model_name, max_retry, local_dir):
     from huggingface_hub import snapshot_download
     for i in range(1, max_retry):
@@ -94,8 +93,8 @@ class GetCoderLMHandle(LocalLLMHandle):
         history.append({ 'role': 'user', 'content': query})
         messages = history
         inputs = self._tokenizer.apply_chat_template(messages, return_tensors="pt")
-        if inputs.shape[1] > MAX_INPUT_TOKEN_LENGTH:
-            inputs = inputs[:, -MAX_INPUT_TOKEN_LENGTH:]
+        if inputs.shape[1] > max_length:
+            inputs = inputs[:, -max_length:]
         inputs = inputs.to(self._model.device)
         generation_kwargs = dict(
                                     inputs=inputs,