大模型-GLM-4-9B

     

目录

     

什么是 GLM-4-9B

快速入门

---

   什么是 GLM-4-9B

      GLM-4-9B是由Zhipu AI推出的GLM-4系列中最新训练模型的开源版。在对语义，数学，推理，代码和知识的数据集的评估中，GLM-4-9B及其人类偏好一致的版本GLM-4-9B-CHAT在Llama-3-8B上表现出了出色的性能。除多轮对话外，GLM-4-9B-CHAT还具有高级功能，例如Web浏览，代码执行，自定义工具调用（功能调用）和长文本推理（最多支持128K上下文）。这一代模型增加了多语言支持，支持了包括日语，韩语和德语在内的26种语言。我们还启动了基于GLM-4-9B的1M上下文长度（约200万个汉字）和多模型GLM-4V-9B的GLM-4-9B-CHAT-1M模型。

     GLM-4V-9B具有中文和英语的对话能力，高分辨率为1120*1120。在各种多模式评估中，包括中文和英语的全面能力，感知和推理，文本识别和图表理解，GLM-4V-9B与GPT-4-Turbo-2024-04-09，Gemini 1.0 Pro相比，表现出卓越的性能QWEN-VL-MAX和Claude 3 Opus

快速入门

使用以下方法快速调用GLM-4-9B-Chat语言模型

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

device = "cuda"

tokenizer = AutoTokenizer.from_pretrained("THUDM/glm-4-9b-chat", trust_remote_code=True)

query = "你好"

inputs = tokenizer.apply_chat_template([{"role": "user", "content": query}],
                                       add_generation_prompt=True,
                                       tokenize=True,
                                       return_tensors="pt",
                                       return_dict=True
                                       )

inputs = inputs.to(device)
model = AutoModelForCausalLM.from_pretrained(
    "THUDM/glm-4-9b-chat",
    torch_dtype=torch.bfloat16,
    low_cpu_mem_usage=True,
    trust_remote_code=True
).to(device).eval()

gen_kwargs = {"max_length": 2500, "do_sample": True, "top_k": 1}
with torch.no_grad():
    outputs = model.generate(**inputs, **gen_kwargs)
    outputs = outputs[:, inputs['input_ids'].shape[1]:]
    print(tokenizer.decode(outputs[0], skip_special_tokens=True))

使用VLLM后端进行推理：

from transformers import AutoTokenizer
from vllm import LLM, SamplingParams

# GLM-4-9B-Chat
# If you encounter OOM, you can try to reduce max_model_len or increase tp_size
max_model_len, tp_size = 131072, 1
model_name = "THUDM/glm-4-9b-chat"
prompt = [{"role": "user", "content": "你好"}]

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
llm = LLM(
    model=model_name,
    tensor_parallel_size=tp_size,
    max_model_len=max_model_len,
    trust_remote_code=True,
    enforce_eager=True,
    # if you encounter OOM in GLM-4-9B-Chat-1M, you can try to enable the following parameters
    # enable_chunked_prefill=True,
    # max_num_batched_tokens=8192
)
stop_token_ids = [151329, 151336, 151338]
sampling_params = SamplingParams(temperature=0.95, max_tokens=1024, stop_token_ids=stop_token_ids)

inputs = tokenizer.apply_chat_template(prompt, tokenize=False, add_generation_prompt=True)
outputs = llm.generate(prompts=inputs, sampling_params=sampling_params)

print(outputs[0].outputs[0].text)

使用以下方法快速调用GLM-4V-9B多模态模型

import torch
from PIL import Image
from transformers import AutoModelForCausalLM, AutoTokenizer

device = "cuda"

tokenizer = AutoTokenizer.from_pretrained("THUDM/glm-4v-9b", trust_remote_code=True)

query = 'display this image'
image = Image.open("your image").convert('RGB')
inputs = tokenizer.apply_chat_template([{"role": "user", "image": image, "content": query}],
                                       add_generation_prompt=True, tokenize=True, return_tensors="pt",
                                       return_dict=True)  # chat mode

inputs = inputs.to(device)
model = AutoModelForCausalLM.from_pretrained(
    "THUDM/glm-4v-9b",
    torch_dtype=torch.bfloat16,
    low_cpu_mem_usage=True,
    trust_remote_code=True
).to(device).eval()

gen_kwargs = {"max_length": 2500, "do_sample": True, "top_k": 1}
with torch.no_grad():
    outputs = model.generate(**inputs, **gen_kwargs)
    outputs = outputs[:, inputs['input_ids'].shape[1]:]
    print(tokenizer.decode(outputs[0]))

---

企业架构、技术方案、案例资源-CSDN博客