he426100
commented
11 months ago -
跑chatglm/cli_demo效果（已实现python版等效的逐字输出效果）
-
php代码
<?php
define('BASE_PATH', DIR . '/..'); require BASE_PATH . '/vendor/autoload.php'; require DIR . '/utils.php';
use Dotenv\Dotenv; use Dotenv\Repository\Adapter; use Dotenv\Repository\RepositoryBuilder; use function Laravel\Prompts\text; use function Laravel\Prompts\info; use function Laravel\Prompts\error;
load_env();
$os = PyCore::import('os'); $platform = PyCore::import('platform'); $transformers = PyCore::import('transformers'); $AutoModel = $transformers->AutoModel; $AutoTokenizer = $transformers->AutoTokenizer; $torch = PyCore::import('torch');
$MODEL_PATH = getenv('MODEL_PATH') ?: 'THUDM/chatglm3-6b'; $TOKENIZER_PATH = getenv("TOKENIZER_PATH") ?: $MODEL_PATH; $DEVICE = $torch->cuda->is_available() ? 'cuda' : 'cpu';
$tokenizer = $AutoTokenizer->from_pretrained($TOKENIZER_PATH, trust_remote_code: true); if ($DEVICE == 'cuda') {
AMD, NVIDIA GPU can use Half Precision
// $model = $AutoModel->from_pretrained($MODEL_PATH, trust_remote_code: true)->to($DEVICE)->eval();
$model = load_model_on_gpus($MODEL_PATH, $torch->cuda->device_count());} else {
CPU, Intel GPU and other GPU can use Float16 Precision Only
$model = $AutoModel->from_pretrained($MODEL_PATH, trust_remote_code: true)->float()->to($DEVICE)->eval();}
$welcome = '欢迎使用 ChatGLM3-6B 模型，输入内容即可进行对话，clear 清空对话历史，stop 终止程序';
$past_key_values = null; $history = []; $stop_stream = false;
info($welcome);
while (true) { $query = text('用户：'); if (trim($query) == 'stop') { break; } elseif (trim($query) == 'clear') { $past_key_values = null; $history = []; info("\033c"); info($welcome); continue; } info('ChatGLM: '); try { $current_length = 0; $rs = $model->stream_chat($tokenizer, $query, history: $history, top_p: 1, temperature: 0.01, past_key_values: $past_key_values, return_past_key_values: true ); $it = PyCore::iter($rs); echo " \e[32m"; while ($next = PyCore::next($it)) { if ($stop_stream) { $stop_stream = false; break; } else { list($response, $history, $past_key_values) = PyCore::scalar($next); echo mb_substr($response, $current_length); $current_length = mb_strlen($response); } } echo "\e[39m\n"; } catch (\Throwable $e) { error($e->getMessage() ?: '执行出错了'); } }
function load_env() { $repository = RepositoryBuilder::createWithNoAdapters() ->addAdapter(Adapter\PutenvAdapter::class) ->immutable() ->make();
Dotenv::create($repository, [BASE_PATH])->safeLoad();}
hgc357341051
matyhtf
zeroxxmmbm
环境 Ubuntu 22.04.3 LTS NVIDIA-SMI 520.61.05 Driver Version: 520.61.05 CUDA Version: 11.8 3080x2
效果
Dockefile
RUN apt-get update && \ apt-get install -y --no-install-recommends build-essential git wget software-properties-common && \ add-apt-repository ppa:ondrej/php && apt-get update && \ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends php-cli php-dev && \ rm -rf /var/lib/apt/lists/*
RUN wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda.sh && \ chmod +x ~/miniconda.sh && \ mkdir -p /opt/conda && \ ~/miniconda.sh -b -u -p /opt/conda && \ rm ~/miniconda.sh && \ /opt/conda/bin/conda init bash
ENV PATH="/opt/conda/bin:$PATH"
RUN conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia && \ conda install -c huggingface transformers
RUN git clone https://github.com/swoole/phpy.git /app/phpy
WORKDIR /app/phpy
RUN phpize && \ ./configure --with-python-dir=/opt/conda && \ make install && \ echo "extension=phpy.so" > /etc/php/8.2/cli/conf.d/20_phpy.ini
RUN php -m | grep -i phpy
CMD [ "bash" ]
docker run --rm -it --gpus all --name phpy phpy bash
docker run -d -it --gpus all -v /data/app:/app -v /data/conda:/opt/conda -v /data/cache:/root/.cache --name phpy phpy
cd examples php pipeline.php
sudo apt-get update sudo apt-get install \ ca-certificates \ curl \ gnupg \ lsb-release -y curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /usr/share/keyrings/docker-archive-keyring.gpg echo \ "deb [arch=$(dpkg --print-architecture) signed-by=/usr/share/keyrings/docker-archive-keyring.gpg] https://download.docker.com/linux/ubuntu \ $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null sudo apt-get update sudo apt-get install docker-ce docker-ce-cli containerd.io -y
distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \ && curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add - \ && curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list sudo apt-get update sudo apt-get install -y nvidia-docker2
wget https://developer.download.nvidia.com/compute/cuda/11.8.0/local_installers/cuda_11.8.0_520.61.05_linux.run sudo sh ./cuda_11.8.0_520.61.05_linux.run