nli-MiniLM2-L6-H768:可用于零样本分类任务，该项目基于 MiniLMv2 架构，在 multi_nli 和 snli 数据集上训练，支持 PyTorch 框架，具备高效推理能力，通过 openMind 库可便捷部署使用。【此简介由AI生成】

与 openMind 配合使用

环境变量

# source environment variable
source /usr/local/Ascend/ascend-toolkit/set_env.sh
export OPENMIND_FRAMEWORK=pt

pip install openMind Library

OpenMind Library 可通过 pip 进行安装，请根据实际环境选择相应命令进行安装。

需要注意的是，由于 torch npu 依赖 torch，在 aarch64 环境下可通过 pip 直接安装，而在 x86 环境下则需要通过特定 URL 下载 CPU 版本，因此两种环境下的安装命令有所不同。具体安装代码已在下文进行区分呈现。

# aarch64
pip install openmind[all]
# x86
pip install openmind[all] --extra-index-url https://download.pytorch.org/whl/cpu

推理

from openmind import AutoTokenizer, AutoModelForCausalLM
import torch
import torch_npu

model_dir = "HangZhou_Ascend/nli-MiniLM2-L6-H768"
tokenizer = AutoTokenizer.from_pretrained(model_dir, device_map="auto", trust_remote_code=True)
# Set `torch_dtype=torch.float16` to load model in float16, otherwise it will be loaded as float32 and might cause OOM Error.
model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto",  trust_remote_code=True, torch_dtype=torch.float16)
model = model.eval()
response, history = model.chat(tokenizer, "1+1=", history=[], meta_instruction="")
print(response)

与 openMind 配合使用

环境变量

# source environment variable
source /usr/local/Ascend/ascend-toolkit/set_env.sh
export OPENMIND_FRAMEWORK=pt

pip install openMind Library

OpenMind Library 可通过 pip 进行安装，请根据实际环境选择相应命令进行安装。

# aarch64
pip install openmind[all]
# x86
pip install openmind[all] --extra-index-url https://download.pytorch.org/whl/cpu

推理

from openmind import AutoTokenizer, AutoModelForCausalLM
import torch
import torch_npu

model_dir = "HangZhou_Ascend/nli-MiniLM2-L6-H768"
tokenizer = AutoTokenizer.from_pretrained(model_dir, device_map="auto", trust_remote_code=True)
# Set `torch_dtype=torch.float16` to load model in float16, otherwise it will be loaded as float32 and might cause OOM Error.
model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto",  trust_remote_code=True, torch_dtype=torch.float16)
model = model.eval()
response, history = model.chat(tokenizer, "1+1=", history=[], meta_instruction="")
print(response)