Merge pull request #899 from yfyang86/main

[Document] 更新Mac部署说明
dev
Zhengxiao Du 2 years ago committed by GitHub
commit ba8daf4c24
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

@ -20,7 +20,7 @@
对 ChatGLM-6B 进行微调的开源项目: 对 ChatGLM-6B 进行微调的开源项目:
* [InstructGLM](https://github.com/yanqiangmiffy/InstructGLM)基于ChatGLM-6B进行指令学习汇总开源中英文指令数据基于Lora进行指令数据微调开放了Alpaca、Belle微调后的Lora权重修复web_demo重复问题 * [InstructGLM](https://github.com/yanqiangmiffy/InstructGLM)基于ChatGLM-6B进行指令学习汇总开源中英文指令数据基于Lora进行指令数据微调开放了Alpaca、Belle微调后的Lora权重修复web_demo重复问题
* [ChatGLM-Efficient-Tuning](https://github.com/hiyouga/ChatGLM-Efficient-Tuning)基于ChatGLM-6B模型进行定制化微调汇总10余种指令数据集和3种微调方案实现了4/8比特量化和模型权重融合提供微调模型快速部署方法。 * [ChatGLM-Efficient-Tuning](https://github.com/hiyouga/ChatGLM-Efficient-Tuning)实现了ChatGLM-6B模型的监督微调和完整RLHF训练汇总10余种指令数据集和3种微调方案实现了4/8比特量化和模型权重融合提供微调模型快速部署方法。
* [ChatGLM-Finetuning](https://github.com/liucongg/ChatGLM-Finetuning)基于ChatGLM-6B模型进行下游具体任务微调涉及Freeze、Lora、P-tuning等并进行实验效果对比。 * [ChatGLM-Finetuning](https://github.com/liucongg/ChatGLM-Finetuning)基于ChatGLM-6B模型进行下游具体任务微调涉及Freeze、Lora、P-tuning等并进行实验效果对比。
* [ChatGLM-Tuning](https://github.com/mymusise/ChatGLM-Tuning): 基于 LoRA 对 ChatGLM-6B 进行微调。类似的项目还包括 [Humanable ChatGLM/GPT Fine-tuning | ChatGLM 微调](https://github.com/hscspring/hcgf) * [ChatGLM-Tuning](https://github.com/mymusise/ChatGLM-Tuning): 基于 LoRA 对 ChatGLM-6B 进行微调。类似的项目还包括 [Humanable ChatGLM/GPT Fine-tuning | ChatGLM 微调](https://github.com/hscspring/hcgf)

@ -191,14 +191,48 @@ model = AutoModel.from_pretrained("THUDM/chatglm-6b-int4",trust_remote_code=True
如果遇到了报错 `Could not find module 'nvcuda.dll'` 或者 `RuntimeError: Unknown platform: darwin` (MacOS) ,请[从本地加载模型](README.md#从本地加载模型) 如果遇到了报错 `Could not find module 'nvcuda.dll'` 或者 `RuntimeError: Unknown platform: darwin` (MacOS) ,请[从本地加载模型](README.md#从本地加载模型)
### Mac 上的 CPU 部署和加速
Mac直接加载量化后的模型会出现问题例如`clang: error: unsupported option '-fopenmp'这是由于Mac由于本身缺乏omp导致的此时可运行但是单核。
以[chatglm-6b-int4](https://huggingface.co/THUDM/chatglm-6b-int4)量化模型为例需要做如下配置即可在Mac下使用OMP
#### 第一步:安装`libomp`
```bash
# 第一步: 参考`https://mac.r-project.org/openmp/`
## 假设: gcc(clang)是14.x版本其他版本见R-Project提供的表格
curl -O https://mac.r-project.org/openmp/openmp-14.0.6-darwin20-Release.tar.gz
sudo tar fvxz openmp-14.0.6-darwin20-Release.tar.gz -C /
```
此时会安装下面几个文件:`/usr/local/lib/libomp.dylib`, `/usr/local/include/ompt.h`, `/usr/local/include/omp.h`, `/usr/local/include/omp-tools.h`
#### 第二步:配置`gcc`编译项
然后针对`chatglm-6b-int4`, 修改[quantization.py](https://huggingface.co/THUDM/chatglm-6b-int4/blob/main/quantization.py),主要是把硬编码的`gcc -O3 -fPIC -pthread -fopenmp -std=c99`命令修改成`gcc -O3 -fPIC -Xclang -fopenmp -pthread -lomp -std=c99`[对应代码](https://huggingface.co/THUDM/chatglm-6b-int4/blob/63d66b0572d11cedd5574b38da720299599539b3/quantization.py#L168)见下:
```python
# 第二步: 找到包含`gcc -O3 -fPIC -pthread -fopenmp -std=c99`的这一行,并修改成
compile_command = "gcc -O3 -fPIC -Xclang -fopenmp -pthread -lomp -std=c99 {} -shared -o {}".format(source_code, kernel_file)
```
> 补充说明:可以用`platform.uname()[0] == 'Darwin'`做OS的判断从而使得[quantization.py](https://huggingface.co/THUDM/chatglm-6b-int4/blob/main/quantization.py)有兼容性。
> 注意:如果你之前运行`ChatGLM`项目失败过最好清一下Huggingface的缓存i.e. 默认下是 `rm -rf ${HOME}/.cache/huggingface/modules/transformers_modules/chatglm-6b-int4`。由于使用了`rm`命令,请明确知道自己在删除什么。
### Mac 上的 GPU 加速 ### Mac 上的 GPU 加速
对于搭载了Apple Silicon的Mac以及MacBook可以使用 MPS 后端来在 GPU 上运行 ChatGLM-6B。需要参考 Apple 的 [官方说明](https://developer.apple.com/metal/pytorch) 安装 PyTorch-Nightly。 对于搭载了Apple Silicon的Mac以及MacBook可以使用 MPS 后端来在 GPU 上运行 ChatGLM-6B。需要参考 Apple 的 [官方说明](https://developer.apple.com/metal/pytorch) 安装 PyTorch-Nightly。
目前在 MacOS 上只支持[从本地加载模型](README.md#从本地加载模型)。将代码中的模型加载改为从本地加载,并使用 mps 后端 目前在 MacOS 上只支持[从本地加载模型](README.md#从本地加载模型)。将代码中的模型加载改为从本地加载,并使用 mps 后端
```python ```python
model = AutoModel.from_pretrained("your local path", trust_remote_code=True).half().to('mps') model = AutoModel.from_pretrained("your local path", trust_remote_code=True).half().to('mps')
``` ```
即可使用在 Mac 上使用 GPU 加速模型推理。 即可使用在 Mac 上使用 GPU 加速模型推理。如果出现关于`half`的报错比如在MacOS 13.3.x上可以改成
```python
model = AutoModel.from_pretrained("your local path", trust_remote_code=True).float().to('mps')
```
> 注意上述方法在非量化版中运行没有问题。量化版模型在MPS设备运行可以关注[这个](https://github.com/THUDM/ChatGLM-6B/issues/462)ISSUE这主要是[kernel](https://huggingface.co/THUDM/chatglm-6b/blob/658202d88ac4bb782b99e99ac3adff58b4d0b813/quantization.py#L27)的原因,可以解包这个`ELF`文件看到是CUDA的实现。
### 多卡部署 ### 多卡部署
如果你有多张 GPU但是每张 GPU 的显存大小都不足以容纳完整的模型那么可以将模型切分在多张GPU上。首先安装 accelerate: `pip install accelerate`,然后通过如下方法加载模型: 如果你有多张 GPU但是每张 GPU 的显存大小都不足以容纳完整的模型那么可以将模型切分在多张GPU上。首先安装 accelerate: `pip install accelerate`,然后通过如下方法加载模型:

@ -188,6 +188,36 @@ model = AutoModel.from_pretrained("THUDM/chatglm-6b-int4", trust_remote_code=Tru
If your encounter the error `Could not find module 'nvcuda.dll'` or `RuntimeError: Unknown platform: darwin`(MacOS), please [load the model locally](README_en.md#load-the-model-locally). If your encounter the error `Could not find module 'nvcuda.dll'` or `RuntimeError: Unknown platform: darwin`(MacOS), please [load the model locally](README_en.md#load-the-model-locally).
### CPU Deployment on Mac
The default Mac enviroment does not support Openmp. One may encounter such warning/errors when execute the `AutoModel.from_pretrained(...)` command `clang: error: unsupported option '-fopenmp'`.
Take the quantified int4 version [chatglm-6b-int4](https://huggingface.co/THUDM/chatglm-6b-int4) for example, two extra steps are needed.
#### STEP 1: Install `libomp`
```bash
# STEP 1: install libopenmp, check `https://mac.r-project.org/openmp/` for details.
# Assumption: `gcc(clang) >= 14.x`, read the R-Poject before run the code:
curl -O https://mac.r-project.org/openmp/openmp-14.0.6-darwin20-Release.tar.gz
sudo tar fvxz openmp-14.0.6-darwin20-Release.tar.gz -C /
```
Four files (`/usr/local/lib/libomp.dylib`, `/usr/local/include/ompt.h`, `/usr/local/include/omp.h`, `/usr/local/include/omp-tools.h`) are installed accordingly.
#### STEP 2: Configure `gcc` with `-fopenmp`
Next, modify the [quantization.py](https://huggingface.co/THUDM/chatglm-6b-int4/blob/main/quantization.py) file of the `chatglm-6b-int4` project. In the file, change the `gcc -O3 -fPIC -pthread -fopenmp -std=c99` configuration to `gcc -O3 -fPIC -Xclang -fopenmp -pthread -lomp -std=c99` (check the corresponding python code [HERE](https://huggingface.co/THUDM/chatglm-6b-int4/blob/63d66b0572d11cedd5574b38da720299599539b3/quantization.py#L168)), i.e.:
```python
# STEP 2: Change the line contains `gcc -O3 -fPIC -pthread -fopenmp -std=c99` to:
compile_command = "gcc -O3 -fPIC -Xclang -fopenmp -pthread -lomp -std=c99 {} -shared -o {}".format(source_code, kernel_file)
```
> Notice: `platform.uname()[0] == 'Darwin'` could be used to determine the OS type and further polish the python script.
> Notice: If you have executed the `ChatGLM` project and failed, you may want to clean the cache of Huggingface before your next try, i.e. `rm -rf ${HOME}/.cache/huggingface/modules/transformers_modules/chatglm-6b-int4`. Since `rm` is used, please MAKE SURE that the command deletes the right files.
### GPU Inference on Mac ### GPU Inference on Mac
For Macs (and MacBooks) with Apple Silicon, it is possible to use the MPS backend to run ChatGLM-6B on the GPU. First, you need to refer to Apple's [official instructions](https://developer.apple.com/metal/pytorch) to install PyTorch-Nightly. For Macs (and MacBooks) with Apple Silicon, it is possible to use the MPS backend to run ChatGLM-6B on the GPU. First, you need to refer to Apple's [official instructions](https://developer.apple.com/metal/pytorch) to install PyTorch-Nightly.
@ -195,8 +225,17 @@ Currently you must [load the model locally](README_en.md#load-the-model-locally)
```python ```python
model = AutoModel.from_pretrained("your local path", trust_remote_code=True).half().to('mps') model = AutoModel.from_pretrained("your local path", trust_remote_code=True).half().to('mps')
``` ```
For Mac users with Mac OS >= 13.3, one may encounter errors related to the `half()` method. Use the `float()` method instead:
```python
model = AutoModel.from_pretrained("your local path", trust_remote_code=True).float().to('mps')
```
Then you can use GPU-accelerated model inference on Mac. Then you can use GPU-accelerated model inference on Mac.
> Notice: there is no promblem to run the non-quantified version of ChatGLM with MPS. One could check [this issue](https://github.com/THUDM/ChatGLM-6B/issues/462) to run the quantified version with MPS as the backend (and get `ERRORS`). Unpacking [kernel](https://huggingface.co/THUDM/chatglm-6b/blob/658202d88ac4bb782b99e99ac3adff58b4d0b813/quantization.py#L27) as an `ELF` file shows its backend is `cuda`, which fails on MPS currently (`torch 2.1.0.dev20230502`).
### Multi-GPU Deployment ### Multi-GPU Deployment
If you have multiple GPUs, but the memory size of each GPU is not sufficient to accommodate the entire model, you can split the model across multiple GPUs. If you have multiple GPUs, but the memory size of each GPU is not sufficient to accommodate the entire model, you can split the model across multiple GPUs.

Loading…
Cancel
Save