update the req and chatglm_tokenizer.py

basic_demo/glm_server.py

@@ -673,7 +673,6 @@ if __name__ == "__main__":
         gpu_memory_utilization=0.9,
         enforce_eager=True,
         worker_use_ray=False,
-        engine_use_ray=False,
         disable_log_requests=True,
         max_model_len=MAX_MODEL_LENGTH,
     )

basic_demo/requirements.txt

@@ -1,23 +1,23 @@
 torch>=2.4.0
 torchvision>=0.19.0
-transformers==4.44.0
-huggingface-hub>=0.24.5
+transformers>=4.45.0
+huggingface-hub>=0.25.1
 sentencepiece>=0.2.0
 jinja2>=3.1.4
-pydantic>=2.8.2
-timm>=1.0.8
+pydantic>=2.9.2
+timm>=1.0.9
 tiktoken>=0.7.0
 numpy==1.26.4 # Need less than 2.0.0
-accelerate>=0.33.0
-sentence_transformers>=3.0.1
-gradio>=4.42.0 # web demo
-openai>=1.43.0 # openai demo
+accelerate>=0.34.0
+sentence_transformers>=3.1.1
+gradio>=4.44.1 # web demo
+openai>=1.51.0 # openai demo
 einops>=0.8.0
 pillow>=10.4.0
 sse-starlette>=2.1.3
 bitsandbytes>=0.43.3 # INT4 Loading
 
-# vllm==0.5.4 # using with VLLM Framework
-# flash-attn>=2.6.1 # using with flash-attention 2
+# vllm>=0.6.2 # using with VLLM Framework
+# flash-attn>=2.6.3 # using with flash-attention 2
 # PEFT model, not need if you don't use PEFT finetune model.
-# peft>=0.12.2 # Using with finetune model
+# peft>=0.13.0 # Using with finetune model

basic_demo/trans_cli_demo.py

@@ -37,11 +37,7 @@ MODEL_PATH = os.environ.get('MODEL_PATH', 'THUDM/glm-4-9b-chat')
 #     return model, tokenizer
 
 
-tokenizer = AutoTokenizer.from_pretrained(
-    MODEL_PATH,
-    trust_remote_code=True,
-    encode_special_tokens=True
-)
+tokenizer = AutoTokenizer.from_pretrained(MODEL_PATH,trust_remote_code=True)
 
 model = AutoModel.from_pretrained(
     MODEL_PATH,

basic_demo/vllm_cli_demo.py

@@ -30,7 +30,6 @@ def load_model_and_tokenizer(model_dir: str, enable_lora: bool):
         gpu_memory_utilization=0.9,
         enforce_eager=True,
         worker_use_ray=True,
-        engine_use_ray=False,
         disable_log_requests=True
         # 如果遇见 OOM 现象，建议开启下述参数
         # enable_chunked_prefill=True,