Fix eos_token_id error

[ypwhs]

tokenizer.eos_token_id outputs 20002, but the eos_token_id defined in the config.json of ChatGLM-6B is 150005, so it needs to be changed to 150005.

{
  "_name_or_path": "THUDM/chatglm-6b",
  "architectures": [
    "ChatGLMModel"
  ],
  "auto_map": {
    "AutoConfig": "configuration_chatglm.ChatGLMConfig",
    "AutoModel": "modeling_chatglm.ChatGLMForConditionalGeneration",
    "AutoModelForSeq2SeqLM": "modeling_chatglm.ChatGLMForConditionalGeneration"
  },
  "bos_token_id": 150004,
  "eos_token_id": 150005,
  "hidden_size": 4096,
  "inner_hidden_size": 16384,
  "layernorm_epsilon": 1e-05,
  "max_sequence_length": 2048,
  "model_type": "chatglm",
  "num_attention_heads": 32,
  "num_layers": 28,
  "position_encoding_2d": true,
  "torch_dtype": "float16",
  "transformers_version": "4.23.1",
  "use_cache": true,
  "vocab_size": 150528
}

[mymusise]

Thanks for this PR!

It looks like a bug in ChatGLM Tokenizer config, this may be the reason for not stopping when generating sentences.
#55 #60

But I'd instead read eos_token_id from the config file than a magic number

config = AutoConfig.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True, device_map='auto')
config.eos_token_id

[ypwhs]

Thanks for this PR!

It looks like a bug in ChatGLM Tokenizer config, this may be the reason for not stopping when generating sentences. #55 #60

But I'd instead read eos_token_id from the config file than a magic number

config = AutoConfig.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True, device_map='auto')
config.eos_token_id

Your solution is better.

[dumpmemory]

It should be 150005. please check stream_chat function