From 62255d34f69bba8d760aec37c5524e9adc510f07 Mon Sep 17 00:00:00 2001
From: KMnO4-zx <1021385881@qq.com>
Date: Wed, 13 Dec 2023 20:18:55 +0800
Subject: [PATCH 1/9] update hello-world
---
helloworld/hello world.md | 692 +++++++++++++++++++++++++++++++++
helloworld/images/Lagent.png | Bin 0 -> 263886 bytes
helloworld/images/image-1.png | Bin 0 -> 20778 bytes
helloworld/images/image-10.png | Bin 0 -> 317777 bytes
helloworld/images/image-11.png | Bin 0 -> 78447 bytes
helloworld/images/image-12.png | Bin 0 -> 104774 bytes
helloworld/images/image-13.png | Bin 0 -> 29168 bytes
helloworld/images/image-14.png | Bin 0 -> 79071 bytes
helloworld/images/image-15.png | Bin 0 -> 445584 bytes
helloworld/images/image-16.png | Bin 0 -> 48915 bytes
helloworld/images/image-17.png | Bin 0 -> 137375 bytes
helloworld/images/image-18.png | Bin 0 -> 32320 bytes
helloworld/images/image-2.png | Bin 0 -> 46176 bytes
helloworld/images/image-3.png | Bin 0 -> 186258 bytes
helloworld/images/image-4.png | Bin 0 -> 193193 bytes
helloworld/images/image-5.png | Bin 0 -> 7589 bytes
helloworld/images/image-6.png | Bin 0 -> 134663 bytes
helloworld/images/image-7.png | Bin 0 -> 66137 bytes
helloworld/images/image-8.png | Bin 0 -> 166738 bytes
helloworld/images/image-9.png | Bin 0 -> 610173 bytes
helloworld/images/image.png | Bin 0 -> 166690 bytes
21 files changed, 692 insertions(+)
create mode 100644 helloworld/hello world.md
create mode 100644 helloworld/images/Lagent.png
create mode 100644 helloworld/images/image-1.png
create mode 100644 helloworld/images/image-10.png
create mode 100644 helloworld/images/image-11.png
create mode 100644 helloworld/images/image-12.png
create mode 100644 helloworld/images/image-13.png
create mode 100644 helloworld/images/image-14.png
create mode 100644 helloworld/images/image-15.png
create mode 100644 helloworld/images/image-16.png
create mode 100644 helloworld/images/image-17.png
create mode 100644 helloworld/images/image-18.png
create mode 100644 helloworld/images/image-2.png
create mode 100644 helloworld/images/image-3.png
create mode 100644 helloworld/images/image-4.png
create mode 100644 helloworld/images/image-5.png
create mode 100644 helloworld/images/image-6.png
create mode 100644 helloworld/images/image-7.png
create mode 100644 helloworld/images/image-8.png
create mode 100644 helloworld/images/image-9.png
create mode 100644 helloworld/images/image.png
diff --git a/helloworld/hello world.md b/helloworld/hello world.md
new file mode 100644
index 000000000..eb30441e7
--- /dev/null
+++ b/helloworld/hello world.md
@@ -0,0 +1,692 @@
+# 大模型领域的Hello World
+
+## 1. 大模型及InternLM模型简介
+
+### 1.1 什么是大模型？
+
+ 大模型通常指的是机器学习或人工智能领域中参数数量巨大、拥有庞大计算能力和参数规模的模型。这些模型利用大量数据进行训练，并且拥有数十亿甚至数千亿个参数。大模型的出现和发展得益于增长的数据量、计算能力的提升以及算法优化等因素。这些模型在各种任务中展现出惊人的性能，比如自然语言处理、计算机视觉、语音识别等。这种模型通常采用深度神经网络结构，如 `Transformer`、`BERT`、`GPT`（ Generative Pre-trained Transformer ）等。
+
+ 大模型的优势在于其能够捕捉和理解数据中更为复杂、抽象的特征和关系。通过大规模参数的学习，它们可以提高在各种任务上的泛化能力，并在未经过大量特定领域数据训练的情况下实现较好的表现。然而，大模型也面临着一些挑战，比如巨大的计算资源需求、高昂的训练成本、对大规模数据的依赖以及模型的可解释性等问题。因此，大模型的应用和发展也需要在性能、成本和道德等多个方面进行权衡和考量。
+
+### 1.2 InternLM 模型全链条开源
+
+ `InternLM` 是一个开源的轻量级训练框架，旨在支持大模型训练而无需大量的依赖。通过单一的代码库，它支持在拥有数千个 `GPU` 的大型集群上进行预训练，并在单个 `GPU` 上进行微调，同时实现了卓越的性能优化。在 `1024` 个 `GPU` 上训练时，`InternLM` 可以实现近 `90%` 的加速效率。
+
+ 基于 `InternLM` 训练框架，上海人工智能实验室已经发布了两个开源的预训练模型：`InternLM-7B` 和 `InternLM-20B`。
+
+ `Lagent` 是一个轻量级、开源的基于大语言模型的智能体（agent）框架，支持用户快速地将一个大语言模型转变为多种类型的智能体，并提供了一些典型工具为大语言模型赋能。通过 `Lagent` 框架可以更好的发挥 `InternLM` 的全部性能。`Lagent` 框架已经发布了两个开源的智能体：`Lagent-7B` 和 `Lagent-20B`。`Lagent` 框架图如下所示：
+
+
+
+ 浦语·灵笔是基于书生·浦语大语言模型研发的视觉-语言大模型，提供出色的图文理解和创作能力，结合了视觉和语言的先进技术，能够实现图像到文本、文本到图像的双向转换。使用浦语·灵笔大模型可以轻松的创作一篇图文推文，也能够轻松识别一张图片中的物体，并生成对应的文本描述。
+
+ 上述提到的所有模型，都会带领大家一起体验哦！欢迎大家来给`InternLM`: https://github.com/internLM/internLM/ 点点 star 哦！
+
+## 2. 单卡InternLM-Chat-7B对话模型web部署
+
+### 2.1 环境准备
+
+在[InternStudio](https://studio.intern-ai.org.cn/)平台中选择 A100(1/4) 的配置，如下图所示镜像选择`Cuda11.7-conda`，如下图所示：
+
+
+
+接下来打开刚刚租用服务器的`进入开发机`，并且打开其中的终端开始环境配置、模型下载和运行`demo`。
+
+
+
+进入开发机后，在页面的左上角可以切换`JupyterLab`、`终端`和`VScode`，并在终端输入`bash`命令，进入`conda`环境。如下图所示：
+
+
+
+进入`conda`环境之后，使用以下命令从本地克隆一个已有的`pytorch 2.0.1` 的环境
+
+```shell
+bash
+conda create --name internLM-demo --clone=/root/share/conda_envs/internlm-base
+```
+
+然后使用以下命令激活环境
+
+```shell
+conda activate internLM-demo
+```
+
+并在环境中安装运行demo所需要的依赖。
+
+```shell
+# 升级pip
+python -m pip install --upgrade pip
+
+pip install modelscope==1.9.5
+pip install transformers==4.35.2
+pip install streamlit==1.24.0
+pip install sentencepiece==0.1.99
+pip install accelerate==0.24.1
+```
+### 2.2 模型下载
+
+[InternStudio](https://studio.intern-ai.org.cn/)平台的`share`目录下已经为我们准备了全系列的`internLM`模型，所以我们可以直接复制即可。使用如下命令复制：
+
+```shell
+mkdir -p /root/model/Shanghai_AI_Laboratory
+cp -r /root/share/temp/model_repos/internlm-chat-7b /root/model/Shanghai_AI_Laboratory
+```
+> -r 选项表示递归地复制目录及其内容
+
+也可以使用 `modelscope` 中的`snapshot_download`函数下载模型，第一个参数为模型名称，参数`cache_dir`为模型的下载路径。
+
+在 `/root` 路径下新建目录`model`，在目录下新建 `download.py` 文件并在其中输入以下内容，粘贴代码后记得保存文件，如下图所示。并运行 `python /root/model/download.py`执行下载，模型大小为 14 GB，下载模型大概需要 10~20 分钟
+
+```python
+import torch
+from modelscope import snapshot_download, AutoModel, AutoTokenizer
+import os
+model_dir = snapshot_download('Shanghai_AI_Laboratory/internlm-chat-7b', cache_dir='/root/model', revision='v1.0.3')
+```
+
+> 注意：使用`pwd`命令可以查看当前的路径，`JupyterLab`左侧目录栏显示为`/root/`下的路径。
+
+
+
+### 2.3 代码准备
+
+首先`clone`代码，在`/root`路径下新建`code`目录，然后切换路径, clone代码.
+
+```shell
+cd /root/code
+git clone https://gitee.com/internlm/InternLM.git
+```
+
+切换commit版本，与教程commit版本保持一致，可以让大家更好的复现。
+
+```shell
+cd InternLM
+git checkout 3028f07cb79e5b1d7342f4ad8d11efad3fd13d17
+```
+
+将 `/root/code/InternLM/web_demo.py`中 29 行和 33 行的模型更换为本地的`/root/model/Shanghai_AI_Laboratory/internlm-chat-7b`。
+
+
+
+### 2.4 终端运行
+
+我们可以在`/root/code/InternLM`目录下新建一个`cli_demo.py`文件，将以下代码填入其中：
+
+```python
+import torch
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+
+model_name_or_path = "/root/model/Shanghai_AI_Laboratory/internlm-chat-7b"
+
+tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, trust_remote_code=True)
+model = AutoModelForCausalLM.from_pretrained(model_name_or_path, trust_remote_code=True, torch_dtype=torch.bfloat16, device_map='auto')
+model = model.eval()
+
+messages = []
+
+print("=============Welcome to InternLM chatbot, type 'exit' to exit.=============")
+
+while True:
+ input_text = input("User >>> ")
+ if input_text == "exit":
+ break
+ response, history = model.chat(tokenizer, input_text, history=messages)
+ messages.append((input_text, response))
+ print(f"robot >>> {response}")
+```
+
+然后在终端运行以下命令，即可体验`InternLM-Chat-7B`模型的对话能力。对话效果如下所示：
+
+```shell
+python /root/code/InternLM/cli_demo.py
+```
+
+
+### 2.5 web demo运行
+
+我们切换到`VScode`中，运行`/root/code/InternLM`目录下的`web_demo.py`文件，输入以下命令后，[**查看本教程5.2配置本地端口后**](./hello%20world.md#52-配置本地端口)，将端口映射到本地。在本地浏览器输入`http://127.0.0.1:6006`即可。
+
+```shell
+bash
+conda activate internLM-demo # 首次进入vscode会默认是base环境，所以首先切换环境
+cd /root/code/InternLM
+streamlit run web_demo.py --server.address 127.0.0.1 --server.port 6006
+```
+
+
+
+注意：要在浏览器打开`http://127.0.0.1:6006`页面后，模型才会加载，如下图所示：
+
+
+
+在加载完模型之后，就可以与InternLM-Chat-7B进行对话了，如下图所示：
+
+
+
+## 3. Lagent+InternLM-Chat-7B模型web部署
+
+Lagent 是一个轻量级、开源的基于大语言模型的智能体（agent）框架，支持用户快速地将一个大语言模型转变为多种类型的智能体，并提供了一些典型工具为大语言模型赋能。通过 Lagent 框架可以更好的发挥 InternLM 的全部性能。
+
+下面我们就开始动手实现！
+
+### 3.1 环境准备
+
+选择和第一个 `InternLM` 一样的镜像环境，运行以下命令安装依赖，如果上一个 `InternLM-Chat-7B` 已经配置好环境不需要重复安装.
+
+```shell
+# 升级pip
+python -m pip install --upgrade pip
+
+pip install modelscope==1.9.5
+pip install transformers==4.35.2
+pip install streamlit==1.24.0
+pip install sentencepiece==0.1.99
+pip install accelerate==0.24.1
+```
+
+### 3.2 模型下载
+
+[InternStudio](https://studio.intern-ai.org.cn/)平台的`share`目录下已经为我们准备了全系列的`internLM`模型，所以我们可以直接复制即可。使用如下命令复制：
+
+```shell
+mkdir -p /root/model/Shanghai_AI_Laboratory
+cp -r /root/share/temp/model_repos/internlm-chat-7b /root/model/Shanghai_AI_Laboratory
+```
+> -r 选项表示递归地复制目录及其内容
+
+也可以在 `/root/model` 路径下新建 `download.py` 文件并在其中输入以下内容，并运行 `python /root/model/download.py`执行下载，模型大小为 14 GB，下载模型大概需要 10~20 分钟
+
+```python
+import torch
+from modelscope import snapshot_download, AutoModel, AutoTokenizer
+import os
+model_dir = snapshot_download('Shanghai_AI_Laboratory/internlm-chat-7b', cache_dir='/root/model', revision='v1.0.3')
+```
+
+### 3.3 Lagent 安装
+
+首先切换路径到`/root/code` 克隆 `lagent`仓库，并通过 `pip install -e .`源码安装 `Lagent`
+
+```shell
+cd /root/code
+git clone https://gitee.com/internlm/lagent.git
+cd /root/code/lagent
+git checkout 511b03889010c4811b1701abb153e02b8e94fb5e # 尽量保证和教程commit版本一致
+pip install -e . # 源码安装
+```
+
+### 3.4 修改代码
+
+由于代码修改的地方比较多，大家直接将`/root/code/lagent/examples/react_web_demo.py` 内容替换为以下代码
+
+```python
+import copy
+import os
+
+import streamlit as st
+from streamlit.logger import get_logger
+
+from lagent.actions import ActionExecutor, GoogleSearch, PythonInterpreter
+from lagent.agents.react import ReAct
+from lagent.llms import GPTAPI
+from lagent.llms.huggingface import HFTransformerCasualLM
+
+
+class SessionState:
+
+ def init_state(self):
+ """Initialize session state variables."""
+ st.session_state['assistant'] = []
+ st.session_state['user'] = []
+
+ #action_list = [PythonInterpreter(), GoogleSearch()]
+ action_list = [PythonInterpreter()]
+ st.session_state['plugin_map'] = {
+ action.name: action
+ for action in action_list
+ }
+ st.session_state['model_map'] = {}
+ st.session_state['model_selected'] = None
+ st.session_state['plugin_actions'] = set()
+
+ def clear_state(self):
+ """Clear the existing session state."""
+ st.session_state['assistant'] = []
+ st.session_state['user'] = []
+ st.session_state['model_selected'] = None
+ if 'chatbot' in st.session_state:
+ st.session_state['chatbot']._session_history = []
+
+
+class StreamlitUI:
+
+ def __init__(self, session_state: SessionState):
+ self.init_streamlit()
+ self.session_state = session_state
+
+ def init_streamlit(self):
+ """Initialize Streamlit's UI settings."""
+ st.set_page_config(
+ layout='wide',
+ page_title='lagent-web',
+ page_icon='./docs/imgs/lagent_icon.png')
+ # st.header(':robot_face: :blue[Lagent] Web Demo ', divider='rainbow')
+ st.sidebar.title('模型控制')
+
+ def setup_sidebar(self):
+ """Setup the sidebar for model and plugin selection."""
+ model_name = st.sidebar.selectbox(
+ '模型选择：', options=['gpt-3.5-turbo','internlm'])
+ if model_name != st.session_state['model_selected']:
+ model = self.init_model(model_name)
+ self.session_state.clear_state()
+ st.session_state['model_selected'] = model_name
+ if 'chatbot' in st.session_state:
+ del st.session_state['chatbot']
+ else:
+ model = st.session_state['model_map'][model_name]
+
+ plugin_name = st.sidebar.multiselect(
+ '插件选择',
+ options=list(st.session_state['plugin_map'].keys()),
+ default=[list(st.session_state['plugin_map'].keys())[0]],
+ )
+
+ plugin_action = [
+ st.session_state['plugin_map'][name] for name in plugin_name
+ ]
+ if 'chatbot' in st.session_state:
+ st.session_state['chatbot']._action_executor = ActionExecutor(
+ actions=plugin_action)
+ if st.sidebar.button('清空对话', key='clear'):
+ self.session_state.clear_state()
+ uploaded_file = st.sidebar.file_uploader(
+ '上传文件', type=['png', 'jpg', 'jpeg', 'mp4', 'mp3', 'wav'])
+ return model_name, model, plugin_action, uploaded_file
+
+ def init_model(self, option):
+ """Initialize the model based on the selected option."""
+ if option not in st.session_state['model_map']:
+ if option.startswith('gpt'):
+ st.session_state['model_map'][option] = GPTAPI(
+ model_type=option)
+ else:
+ st.session_state['model_map'][option] = HFTransformerCasualLM(
+ '/root/model/Shanghai_AI_Laboratory/internlm-chat-7b')
+ return st.session_state['model_map'][option]
+
+ def initialize_chatbot(self, model, plugin_action):
+ """Initialize the chatbot with the given model and plugin actions."""
+ return ReAct(
+ llm=model, action_executor=ActionExecutor(actions=plugin_action))
+
+ def render_user(self, prompt: str):
+ with st.chat_message('user'):
+ st.markdown(prompt)
+
+ def render_assistant(self, agent_return):
+ with st.chat_message('assistant'):
+ for action in agent_return.actions:
+ if (action):
+ self.render_action(action)
+ st.markdown(agent_return.response)
+
+ def render_action(self, action):
+ with st.expander(action.type, expanded=True):
+ st.markdown(
+ " 插 件:" # noqa E501
+ + action.type + '
',
+ unsafe_allow_html=True)
+ st.markdown(
+ " 思考步骤:" # noqa E501
+ + action.thought + '
',
+ unsafe_allow_html=True)
+ if (isinstance(action.args, dict) and 'text' in action.args):
+ st.markdown(
+ " 执行内容:
", # noqa E501
+ unsafe_allow_html=True)
+ st.markdown(action.args['text'])
+ self.render_action_results(action)
+
+ def render_action_results(self, action):
+ """Render the results of action, including text, images, videos, and
+ audios."""
+ if (isinstance(action.result, dict)):
+ st.markdown(
+ " 执行结果:
", # noqa E501
+ unsafe_allow_html=True)
+ if 'text' in action.result:
+ st.markdown(
+ "" + action.result['text'] +
+ '
',
+ unsafe_allow_html=True)
+ if 'image' in action.result:
+ image_path = action.result['image']
+ image_data = open(image_path, 'rb').read()
+ st.image(image_data, caption='Generated Image')
+ if 'video' in action.result:
+ video_data = action.result['video']
+ video_data = open(video_data, 'rb').read()
+ st.video(video_data)
+ if 'audio' in action.result:
+ audio_data = action.result['audio']
+ audio_data = open(audio_data, 'rb').read()
+ st.audio(audio_data)
+
+
+def main():
+ logger = get_logger(__name__)
+ # Initialize Streamlit UI and setup sidebar
+ if 'ui' not in st.session_state:
+ session_state = SessionState()
+ session_state.init_state()
+ st.session_state['ui'] = StreamlitUI(session_state)
+
+ else:
+ st.set_page_config(
+ layout='wide',
+ page_title='lagent-web',
+ page_icon='./docs/imgs/lagent_icon.png')
+ # st.header(':robot_face: :blue[Lagent] Web Demo ', divider='rainbow')
+ model_name, model, plugin_action, uploaded_file = st.session_state[
+ 'ui'].setup_sidebar()
+
+ # Initialize chatbot if it is not already initialized
+ # or if the model has changed
+ if 'chatbot' not in st.session_state or model != st.session_state[
+ 'chatbot']._llm:
+ st.session_state['chatbot'] = st.session_state[
+ 'ui'].initialize_chatbot(model, plugin_action)
+
+ for prompt, agent_return in zip(st.session_state['user'],
+ st.session_state['assistant']):
+ st.session_state['ui'].render_user(prompt)
+ st.session_state['ui'].render_assistant(agent_return)
+ # User input form at the bottom (this part will be at the bottom)
+ # with st.form(key='my_form', clear_on_submit=True):
+
+ if user_input := st.chat_input(''):
+ st.session_state['ui'].render_user(user_input)
+ st.session_state['user'].append(user_input)
+ # Add file uploader to sidebar
+ if uploaded_file:
+ file_bytes = uploaded_file.read()
+ file_type = uploaded_file.type
+ if 'image' in file_type:
+ st.image(file_bytes, caption='Uploaded Image')
+ elif 'video' in file_type:
+ st.video(file_bytes, caption='Uploaded Video')
+ elif 'audio' in file_type:
+ st.audio(file_bytes, caption='Uploaded Audio')
+ # Save the file to a temporary location and get the path
+ file_path = os.path.join(root_dir, uploaded_file.name)
+ with open(file_path, 'wb') as tmpfile:
+ tmpfile.write(file_bytes)
+ st.write(f'File saved at: {file_path}')
+ user_input = '我上传了一个图像，路径为: {file_path}. {user_input}'.format(
+ file_path=file_path, user_input=user_input)
+ agent_return = st.session_state['chatbot'].chat(user_input)
+ st.session_state['assistant'].append(copy.deepcopy(agent_return))
+ logger.info(agent_return.inner_steps)
+ st.session_state['ui'].render_assistant(agent_return)
+
+
+if __name__ == '__main__':
+ root_dir = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+ root_dir = os.path.join(root_dir, 'tmp_dir')
+ os.makedirs(root_dir, exist_ok=True)
+ main()
+```
+
+### 3.5 Demo 运行
+
+```shell
+streamlit run /root/code/lagent/examples/react_web_demo.py --server.address 127.0.0.1 --server.port 6006
+```
+
+用同样的方法我们依然切换到`VScode`页面，运行成功后，[**查看本教程5.2配置本地端口后**](./hello%20world.md#52-配置本地端口)，将端口映射到本地。在本地浏览器输入`http://127.0.0.1:6006`即可。
+
+我们在 `Web` 页面选择 `InternLM` 模型，等待模型加载完毕后，输入数学问题 已知 `2x+3=10`，求`x` ,此时 `InternLM-Chat-7B` 模型理解题意生成解此题的 `Python`代码，`Lagent` 调度送入 `Python` 代码解释器求出该问题的解。
+
+
+
+## 4. 双卡InternLM-XComposer-7B图文模型web部署
+
+### 4.1 环境准备
+
+首先在 [InternStudio](https://studio.intern-ai.org.cn/) 上选择A100(1/4)*2的配置。如下图所示：
+
+
+
+接下来打开刚刚租用服务器的`进入开发机`，并在终端输入`bash`命令，进入`conda`环境，接下来就是安装依赖。
+
+进入`conda`环境之后，使用以下命令从本地克隆一个已有的`pytorch 2.0.1` 的环境
+
+```shell
+conda create --name xcomposer-demo --clone=/root/share/conda_envs/internlm-base
+```
+
+然后使用以下命令激活环境
+
+```shell
+conda activate xcomposer-demo
+```
+
+接下来运行以下命令，安装 `transformers`、`gradio` 等依赖包。请严格安装以下版本安装！
+
+```shell
+pip install transformers==4.33.1 timm==0.4.12 sentencepiece==0.1.99 gradio==3.44.4 markdown2==2.4.10 xlsxwriter==3.1.2 einops accelerate
+```
+### 4.2 模型下载
+
+[InternStudio](https://studio.intern-ai.org.cn/)平台的`share`目录下已经为我们准备了全系列的`internLM`模型，所以我们可以直接复制即可。使用如下命令复制：
+
+```shell
+mkdir -p /root/model/Shanghai_AI_Laboratory
+cp -r /root/share/temp/model_repos/internlm-xcomposer-7b /root/model/Shanghai_AI_Laboratory
+```
+> -r 选项表示递归地复制目录及其内容
+
+也可以安装`modelscope`，下载模型的老朋友了
+
+```shell
+pip install modelscope==1.9.5
+```
+
+在 `/root/model` 路径下新建 `download.py` 文件并在其中输入以下内容，并运行 `python /root/model/download.py`执行下载
+
+```python
+import torch
+from modelscope import snapshot_download, AutoModel, AutoTokenizer
+import os
+model_dir = snapshot_download('Shanghai_AI_Laboratory/internlm-xcomposer-7b', cache_dir='/root/model', revision='master')
+```
+
+### 4.3 代码准备
+
+在 `/root/code` `git clone InternLM-XComposer` 仓库的代码
+
+```shell
+cd /root/code
+git clone https://gitee.com/internlm/InternLM-XComposer.git
+cd /root/code/InternLM-XComposer
+git checkout 3e8c79051a1356b9c388a6447867355c0634932d # 最好保证和教程的commit版本一致
+```
+
+### 4.4 Demo 运行
+
+在终端运行以下代码：
+
+```shell
+cd /root/code/InternLM-XComposer
+python examples/web_demo.py \
+ --folder /root/model/Shanghai_AI_Laboratory/internlm-xcomposer-7b \
+ --num_gpus 1 \
+ --port 6006
+```
+
+> 这里`num_gpus 1`是因为InternStudio平台对于`A100(1/4)*2`识别仍为一张显卡。但如果有小伙伴课后使用两张3090来运行此demo，仍需将`num_gpus`设置为 `2` 。
+
+[**查看本教程5.2配置本地端口后**](./hello%20world.md#52-配置本地端口)，将端口映射到本地。在本地浏览器输入`http://127.0.0.1:6006`即可。我们以`又见敦煌`为提示词，体验图文创作的功能，如下图所示：
+
+
+
+接下来，我们可以体验以下图片理解的能力，如下所示~
+
+
+
+## 5. 通用环境配置
+
+### 5.1 pip、conda 换源
+
+更多详细内容可移步至[MirrorZ Help](https://help.mirrors.cernet.edu.cn/)查看。
+
+#### 5.1.1 pip 换源
+
+临时使用镜像源安装，如下所示：`some-package` 为你需要安装的包名
+
+```shell
+pip install -i https://mirrors.cernet.edu.cn/pypi/web/simple some-package
+```
+
+设置pip默认镜像源，升级 pip 到最新的版本 (>=10.0.0) 后进行配置，如下所示：
+
+```shell
+python -m pip install --upgrade pip
+pip config set global.index-url https://mirrors.cernet.edu.cn/pypi/web/simple
+```
+
+如果您的 pip 默认源的网络连接较差，临时使用镜像源升级 pip：
+
+```shell
+python -m pip install -i https://mirrors.cernet.edu.cn/pypi/web/simple --upgrade pip
+```
+
+#### 5.1.2 conda 换源
+
+镜像站提供了 Anaconda 仓库与第三方源（conda-forge、msys2、pytorch 等，各系统都可以通过修改用户目录下的 .condarc 文件来使用镜像站。
+
+不同系统下的.condarc目录如下：
+
+- `Linux`: `${HOME}/.condarc`
+- `macOS`: `${HOME}/.condarc`
+- `Windows`: `C:\Users\\.condarc`
+
+注意：
+
+- `Windows` 用户无法直接创建名为 `.condarc` 的文件，可先执行 `conda config --set show_channel_urls yes` 生成该文件之后再修改。
+
+快速配置
+
+```shell
+cat <<'EOF' > ~/.condarc
+channels:
+ - defaults
+show_channel_urls: true
+default_channels:
+ - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main
+ - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/r
+ - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/msys2
+custom_channels:
+ conda-forge: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
+ pytorch: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
+EOF
+```
+
+### 5.2 配置本地端口
+
+首先我们需要配置一下本地的 `SSH Key` ，我们这里以`Windows`为例。
+
+步骤①：在本地机器上打开`Power Shell`终端。在终端中，运行以下命令来生成SSH密钥对：（如下图所示）
+```shell
+ssh-keygen -t rsa
+```
+
+
+步骤②： 您将被提示选择密钥文件的保存位置，默认情况下是在`~/.ssh/`目录中。按`Enter`键接受默认值或输入自定义路径。
+
+步骤③：公钥默认存储在 `~/.ssh/id_rsa.pub`，可以通过系统自带的 `cat` 工具查看文件内容：（如下图所示）
+
+```shell
+cat ~\.ssh\id_rsa.pub
+```
+
+> `~` 是用户主目录的简写，`.ssh` 是SSH配置文件的默认存储目录，`id_rsa.pub` 是SSH公钥文件的默认名称。所以，`cat ~\.ssh\id_rsa.pub` 的意思是查看用户主目录下的 `.ssh` 目录中的 id_rsa.pub 文件的内容。
+
+
+
+步骤④：将公钥复制到剪贴板中，然后回到 `InternStudio` 控制台，点击配置SSH Key。如下图所示：
+
+
+
+步骤⑤：将刚刚复制的公钥添加进入即可。
+
+
+
+步骤⑥：在本地终端输入以下指令.6006是在服务器中打开的端口，而33090是根据开发机的端口进行更改。如下图所示：
+
+```shell
+ssh -CNg -L 6006:127.0.0.1:6006 root@ssh.intern-ai.org.cn -p 33090
+```
+
+
+
+### 5.3 模型下载
+
+#### 5.3.1 hugging face
+
+使用`huggingface`官方提供的`huggingface-cli`命令行工具。安装依赖:
+
+```shell
+pip install -U huggingface_hub
+```
+
+然后新建python文件，填入以下代码，运行即可。
+
+- resume-download：断点续下
+- local-dir：本地存储路径。（linux环境下需要填写绝对路径）
+
+```python
+import os
+
+# 下载模型
+os.system('huggingface-cli download --resume-download internlm/internlm-chat-7b --local-dir your_path')
+```
+
+#### 5.3.2 modelscope
+
+使用`modelscope`中的`snapshot_download`函数下载模型，第一个参数为模型名称，参数`cache_dir`为模型的下载路径。
+
+注意：`cache_dir`最好为绝对路径。
+
+安装依赖：
+
+```shell
+pip install modelscope==1.9.5
+pip install transformers==4.35.2
+```
+
+在当前目录下新建python文件，填入以下代码，运行即可。
+
+```python
+import torch
+from modelscope import snapshot_download, AutoModel, AutoTokenizer
+import os
+model_dir = snapshot_download('Shanghai_AI_Laboratory/internlm-chat-7b', cache_dir='your path', revision='master')
+```
+
+#### 5.3.3 Openxlab
+
+Openxlab 可以通过指定模型仓库的地址，以及需要下载的文件的名称，文件所需下载的位置等，直接下载模型权重文件。
+
+使用python脚本下载模型首先要安装依赖，安装代码如下：`pip install -U openxlab` 安装完成后使用 download 函数导入模型中心的模型。
+
+```python
+from openxlab.model import download
+download(model_repo='OpenLMLab/InternLM-7b', model_name='InternLM-7b', output='your local path')
+```
+
+## 6. 课后作业
+
+助教老师会在社群中进行公布哦！
\ No newline at end of file
diff --git a/helloworld/images/Lagent.png b/helloworld/images/Lagent.png
new file mode 100644
index 0000000000000000000000000000000000000000..6e66a4aa0aca9efc389a1b12362aa13f7e5c1f30
GIT binary patch
literal 263886
zcmcG$2|SeD`#)Y7`hKNflh)YYFNy!L9)HfQ41VmcOa-+Y!v!na3
zfn+5l!2qKA_6~S=vIogU7y<*|;qGKdH}DB=gVxX#G-&WAC2c2X2RR9@)IB}j94#EN
zNIY3sT3Z$(F9DGQw?&b1P|sbIB$C&3A_jBE+r!&C7}Rr
zX}EZM5J4wx1$hZc2^nbxNeKn!{r_;MEhPo|N`ZT(_D=TR_)U{B8FwVxJ8v-UE~W2{
zMfoSbv8Hi25cF57!Y8_a7h>j
zsfRI=GDeY1jY%#Lsw@R7EDe)G8OrFG`htEGq^qnG&C$ZZ8K#A{J-mRPJY0wJes
zX6fN+EH7&W@isC*dl0csjvl7cW-e$29b>!!;73~yg4OkK*Hf_2(ZIT)<+WY?6fDfl
zE%Y6vXf$JceLp7z!r2usPjkmpU?g=7pfNDGnHgxw+Ac;U7}e0-5)ZTorYDCbki89j
zbp1$@Xeo*VQXQy_ww}4Pyb<`+F#wGd_=C91%bQWaaU@h98lnQQ7oJ;)|r+9-V)516y2fvXq6*WJw60H-aj<6&v4YeClZ(=su3@x~f!
z={xwqoQODM7k3X$qOS>{SwvS61da)&FOP+JVl{O%o(6J9$`Q8P3rVOkk{Q
zLc;>jq74Hk1f!D8fKr&^ygWU@H@rH<$kmCaE2p8MU}_*~j>g+VOw64SPB?o5EqPf-
zoHhz+uLZN9dbmMM;~`6v#v)DeG)=0U7SI|PLQCEgj`i@eM;q#To0yox
zXW;oQ6wAWW^)1A>M@Ts8HHZt_H`ua_qct%D-ra1KD-
zoKP5TZ>X-8mjVvYG&(7H7hj;=+Ga30KMhOSpI2VqE($cFpRuf@E5VQ`<%x&+;r%dB
zCqsD>%@9fP!Qk8t$&!}luHKqxS$QK&cIh#EvcQz?ub&D4yf!_;87oG}sQ
zi0nBBu!q7u9W|v)^t2>_j{DH?riKm}ZC?$F
zfulCs&ynOHEk&cs*~6V3^!@N=2yJ^BhD>vnlR_y#onaV1q`9^S(!(33XCSBG;)0_=F}fxMPpkt8sih8;m9xgAWq@d9q%6?B7HGJmw-y9SH8jGwQ+)|A4HqO8=cJ4Cwsa(s?LlzY!r=_H
zC0*2A+$~&X93?FzF=p-x3hJ(=77lVW8E0uJq@)R2Pr;062r26iN@{t*6`0maMp|eaAqlclcsHCn1m^-rLOpRt3T`-0vbnB-mXE2N6U@}b6Km)s
zO>mKO_11HR>FY{E@ep66B^0Ws;7xULwe)e3lJg--qnZAUf`q#I7`YojyqKZHnIh|>
z;b?@jmvsV>Rolc9Zbp$dbY@0bCp5%S)qgO@el_R;M+t4?FQ9Et9v^-dwD>K(#F0-W=zCU&~i=~s4>(G?<@zGve)xPK#|5&
zD8)7KPLEbEKirzWRQ?8Z>tWISpqYZC`tujFAS_z}Z=r
z3bD|TV`3tEin_Xqh9gneL(kC6Oxi;hVFGbSn|kQTXjz!Z%X^u7LmlxHR|*Dj1o3sJ
zIzsil&9!8(rUV_F9@8WpQJ0GHB6-r7?%5fRG9@XX4Rk$q
z5t8+05Szlc&+}BXS-CfcLaOIA2M3|!7m^cXLXiNmRw0*Hmc`E3_;O=mDbE1c=
z3)B?m=H%(DW3Quwl6KY6wJ?{)=rg6M=Pikl_VHA3@RjrM1Be4A2OJhNs>&OKFOu3$
zI6ngzPTClN2yG1?n57gBK?M)+`Z{t@T*RlH7V11GX21ASCnHqy(I{zO~s=JHtsD>c2TPI@YMNU^n8~J!fuVv6|)lMQP@>
z65$e$&>Jn2BHxUwVNaGsoUkj!)wsF>^Wt_@5+rc
zLao=B7Z}@+VBSTa(9MmQ?Y645*W;}}oBQ+2q_N;pu|~tLETQq&xSN?l(9gU&L0D!xE6H(mkYs#u=m{Fh7BtJpD(l*(jMi}n}nJZWX`PloWQ)-Jz{ke
z*2y6(8;|RFIBCcZhwxFO@|3EdauDsu_r;-aP3$+T;u`#Z^LeD#-g`bt?r@YnU^)DN^uotiE(q=;mussNpE-=bOQG&92iiqcxtGE=H#=nxwM(5CH
zyY!OU6+X6KZC=toi6agN3e6cjDZ^EwiejvvRZDiAan}0x!wQVE*a*G5tvIAzdL{oL
zCdT@*WkXMP^4x1*;c;DewaEOXuzSqNkHl!frOF
z1qr9v-4AA777I=5lq2o?cZRUR#^c>rl5iL-AyK$AeQtZ#1yV=;8$kjxZSU1~bH*11
zL|*5s$A7}GkupoPu>Qgg13P%&ZfG+sBIhvTkZ|Lg1VKDXkY?M!@$vm13~R=+^;GAN
z7+}AIRaAU<{s<2v-AIt2v1A>;YEy%`G|m6%ehvHRv-17_&UAk&o4Wft=e;3@5*>>I;5F;?H-OXgt40^Ft2C0YMyyebwj
zOiDnX@24tJlaYi4lXF9VutoysSYGZcbQ1Xk^#+S(5!xRgQOB-Qd4xObsNlGfHCg-L
zdxb5bEW629Wi%PqaDDHC#Vx3YiY$7?w?^JKai+)rZNLB`P`k9NI6)@sy`?u#af?<%
z4@+m;ATW~1-p}|y5m1O)1h_1pza>a*l!II&O1a>1jj4lbFdVw
z;*Tz0M?x0H*|F)b?f*W&gd$EE@f4@UWrvhBIze#6!FO~Wq^ps8ucdCG)|)?4nozwR
zS%_ufHgyyJ+korTrV}d1j}*J}?zI{5=C~65?->Vg;PeG}1;+~;Wi!rGI_EEi{t00M
zhAicKFe7-jTRZ=~oWL6mXwc%M}`15ogdyy1_h8_bR#b$mgcJ>eO74QSFUFeo~&pLPt
zB?eC8i1vgL9yuIQWX+m!g^YAoYP}NkN9Gt(8RD4Ai|A4QxbKu?wZ;uto4rQoJ>k^D
zs+p-dvGhArGY>@mWC;Y~AX~GsDrfCy?FKt?bKex3vU3-@pT<|zKMWkSCuDFl*#1l}
z?mPim!lyyQ>+Qm}AA6Rr+rv6}9+jq3mYg!DD(n$|E(tZ?(5(0z*OKt;)NG3+)o21(
zFq}w%YD&`K4$iY)#`l`=
z{DE8E!g6!hzQPvGScgACIZ-Sq;di&gZLf*o_+N@Uw)~0k;$qZn;MKyiNivW%ABI9$
zPTQTKzy3su7vn?1vm<@OvV6W~IZg5wy7Ns>VAd*hW!EGaG%4|4LpJDz>wRbLc5La`+GS2r_Wb&VtILMp}>ut8a6bS&@STD7<75jI{to%v+f~ryq=Re4Vs*i
zS0zc}+L-6Ngwbnx8(RH`uz#56dK{RCt4A;K{Kh=cP;AQz`bLG%Gk46_o-B-oi|u6>
z348#o)=r3M{JFPP-{PvrzP63DNkBgoTQiCv!TBQoDKf=3ta$#tVQ@ytOkOg@Mn|^K
zUy85LEqpYm+DzDQUcPJ{dqF(F=5=&R9=4xVZam||Cu6f
z;if=lo=}rs@6l7*sqIxf_M>h{e3FN_(7TYA?;nqf4fol*Z%BuJ#%My`W;FlG@qbXm
zhwoWF2Z}AM75Op69LhK-da26r)3MVDfhqL5c)8f=zWgJ1#{`bXDyAOKVWR|A9JjM*D*YB!1a$t0bXV^eff8)#s
zy=*LDIpsqC02LmZfnKLNs72_X;xIWNfa8z05e7{lHgSWZ#ohY~hd=#83X^Wawy!YH
zthX8aQ~n6!kT{AAH32m9moPtx6!>sTts(jHw*M}-P=v5rLxl-Z{_@ZKTT~0UwYS|U
zW+UQs>8Dge^0x>QBse#$1>R*8b>yTJbfR#XMXj6{r?@D#eAYIL{cK_p|YSMF0jDTZM^{Px%IcX`0@?Y+-G|HJSN
zrGPiL%hE?SmVw428X89-!3Vj!Pb2#tpT?~g(X)mWUnjI{GS1LSTRxM+av*O9e=6wK
z?NMq#v-Z2OnEr!83Iv1IopElyxj{_BB{jA%koEhCY_0#^W`%C(VFEV;H^;3Lx6+5t
z8NoWCYL2=a41YhV)}VO1=N9{a75$db^X&TyB^G%ke|3}Q27Hukz20~2^=00NX@Z%t
zJ2DR=lhf1=oFb+EZPg!p0+;IJZE*3Y601>xLOOR{d;1Tv67>cKpjD%;xA*5Pp#q^?
zY#1xq@$yRqP+hZv-{T4b#D2;8F!IaG#8JQt7!gOc4PG{e*EM`nv+8lGUV|55Ioh4{
z%V%x^Pe#CAWi7UEEXaOd(W`MKBNFI_igLQdx7%`2>wR6XtaaQ59^&{5I2Zzv;;hB=#?5R4yN!XjUyU9o2B-1s5n5#!3{$8dHr-oYX9JjL|H;H
z@$>A^So8S0aOH@pUkme}X@MIgY?Ki~8-NPM)bVhE=ByNQWAXqd?_R(09=c=xx{PWx
zv`S#~u}sY1zU@)QA9UgJ4{mfL;Nn7oR#&b$8*H9tA9Sl2M#sJBhUCVI9%OpI!6>G0
z|1WMml!YDrXf|19#0bfd#~*33-*w=cniBl@wi9M{4Az%2_~8wN@hdi6JqX5om+SCp
zW9A%S{ru6VfGPi1m%m#S?0Mu1b?A~flaLkS%?{8#p;%@#)2n>CxdC>&6GyrCf@ql?HXj{FVI!VEs{br$-D36|sVQ>Zq!m$`AB^D{&I-Kj7_08&2({}>
z!e|ROlr|*&Qx}eGp)Z0|F2z_MTOaW>?|+E0CYh6
zM9?<*9e`BBa*!lD%+KUbO_&QT#oUz_#GUJ*
zdiq~X$_Dc!MK)7r0ixjOl=!KiDE^@k%l#vT24WuKnO@r$61dl
zQ;2Boal5XNggp`n3XEGnqa|eSvFXcC8#hi&+FQBZ793xm+G0w)1
zA+in*oa!y&5Ej2=VY)82Ry$I^Afc3eP)87#%8shj++3zr1+-fq-q27@YbG$2i~9#J
z^WXHyT!uD^0=fTONp?x*S;#NkyE2%OjRgy->^7RTR<;?#hcrD)yB?
zRyR2WwR8s1zP#C_En9xBmoY%1{L%bE{}8y8cfYJsrG9mzjj5lLU*
zTXxL8jPfr%H%H_Y^z0(FB$Bh}X9ih|6KB<;E)@M;*bisqesg`FH$$0pMpcTdtrw4P
z-MacLEJr@B-NvN#bZVvBUzO(K+ovt;6~6%-T2!@lPvo!o_me<8a9IYIqc8lj?ju0e
zUKa6YI>8zg53Rg7yFaQ3JuRMCCo)zH@ZA1yd@j2n@kDh&WrO3j;2XMu@k!>tP^#BBC%G7V(_fV4q
zI_s=)RqB@v(*eCKpCT1x1;_WH$-mt5|IrB=JUO2?fGm8~n%OfI-*5Q#;|aKh(ifSl
zGSBbWRoG-9lg04)WY%3?x5BSZv+a6lrRHaJQL4$u2Fc)v&x6k_>Twp8OdcN{lwoDt
zjX5eb>ILt;4-M*mo_8I0^n`B5E=(lqzUrp9zUwm#cn=A)zqvQSxbkk`;VX=frBu7v
zhymn~IUzLuGu%EbX2$5r&LYoEEAjp08#Bg>5bhfOz3(54mFCmmadOw%ndn3Yjgm1I
zjO_S%4D4rw?AG|b1>=t=Ds9KUku
z@6Aqbd!YbWSoVv1Hq=Z_a0fyZmrAySHB`h|@QnQ0w+h`22rC!ZY;gj6Sq$N>`#sYx
z3-|r~rry?SRg)3JN$N*4c1H%;SI#E3i+omJKqQNdvcoS6tx7az*>_#VZJmW^#3XgA
z^u^5?$J_4lnran_zCLRmppyNHpYU;pD9nEbR90Cd{_0|xkGx-CcQYa3r
zIB?`mD@zLxnbVaA9jN@!wBhzqfl-cnhtp5|FVpwO*X&hrIu1`l6lO6-i^;G~E4bf4
z0ZU2l%NF~C&(_Je&IQ8nX_)<(62o5{>~`#9aA0~VopuaSc+<8P{L{4=p}t84`Rvy=
zy#*di?vtOk&YMLhkFw&gH2+>5R=W#I8F_3QN^(pB1XV&lZ}2_zM_`;Uj=JdmIx+#m
z*S3xDwZToCLnum(q
zJNXt#%|a^2M<4C0kL1AI{!3No&B+IFS3x}9s6OO&rTNiXZN`&*BfAe59y!(c?Daak
zSIl=Cn=Y?Jx(+PuLL5a`vf?*4qYYFa@@eqghO~E|F@gz>|9vUhydW9T8D3j=W9u(9
z(>;c3cKlDa2Ckq77R!XP$8Kd|hCU_72kn)Lor>S7Fk6deq;jrWS3C}TIgms?*Wmiv
zcBC}xT@nAYQOEiS4vdwSQ}*jyS%Jd%GjP?)58G!kI`OEuv(G+SK05bfcReYB4HJ8{
z-IgI%Ur=3S~cDBXOi|6eE(e9{t4sXk?JO4QIlvw~+|E
z9+Aux{^e%>7^Du|9@Z)G)JJKuKfd2fxfUtUZg__J#kZKA$+?uU{ExIPb4CCGwSE_&~zvzlHk9%{K>^
zD}tpYBB2`!VzdMTfklmq^UF=j{uzF58+-uXBl96367DA)4Xr-7mxYzz5@e+6sVqTf
z4H!4K2HIXH>qU&fm47e;z?HvG;8YMMy&+N_q+Rt&dZ#j*JR{Xs)!t;z9tNgO8O8>O4=coroDMCtDdZg10GJ
z4o#$5?>yBN`Om5i?Cz`o2kC8KEfKRnb6K_Yi_A3Q+pwhUrjT@3ToTS`+=_hkwJ1Ej
z4Nnmcobo((Lf!rF+7`I^O$8J?_a5pE?{JA$xZ>Fl0$p(wQ#kjwx@vvJSS_7&alUFix>#2$C;fusw;It8D8iP?;o{swrdS}FRLbt9}EW7ZL0q&#om6s9ibJt>W^q25??k3Z}EWu$B9AUslmFZeM%
ziGP_!Dc5B@3USE2OD&VB->W#W2Ohhf?9G805E>l|DKJQy^?gH^UZm}XuWpM`RcFj_
zj(Dd4aG=*0vrf?)z+8{5!~c1
ztzU^^6be9sf~h?-u>VA30R}h+#=(XTMMa(!eU7pf9W{uQnu>3KcjvZ6QlxOnZiP-U
zs_M1jMfJ#ztHqM(*T5OWMEWXAUDk=w`gFMC*O}q?I##~maQaiukzKnn8K%ZPAB(*+
zC2A6z)P|Og8voJQfB+X&Ja29lB=8C^Zv+ja7l8cBzN8zQPx!q(=xb=Q8Zzn=PxgsG
zl5=km6T%6IWtHfez?1mOkipXUS<^!fm3&zHK09Gz^4%h2gC941I4)~Q?7H>ZtxCkQ
zP=lEjAkc{9-p>qF>Y($a$C
zU{}RT^~YC@&Rv~$!GXk@rQs~gk}PT;#bUCw*WJ>idah6|=)lfoSwiAe^+G>AGtH-;
z))TzGdhbzKe9B;%&@!3isLvVayxHiB3i0^3vAtBTg`w!PDj0@?D@?%(I~^_i;l
z`*BAkzJhOHLV@46JM_blI?&J~aYdd{4PiNZReyNqqW8}R={gHIF$}_a151}5-FkPOm
zAsE=too{1*^Ou#^uxJR|9Qpv!Hz_*(H}$&A3;Uc)@ySZFrv_eT*##f=)KeUcxD_!K
zNA`-kU=8aOK_-9p{FKkPjO!78D81Me&*`H8H)`;^K}Yh{vGs<>QP5J$zlHGvw$C!S
z)z94-?b(FdR5jc73&q7;mwP5ScqFH_RyumE*OG_^d3e
zsJMIf@r0tjFZsO3wEA)PPBw7^Prhh(b(YI#?dXGhzt8inz44S)d^IF|E4}FJjOcFV
zZ*REA`f?2!?PqGfAEVUVB^^jQP!3J{iTQ)N^Vv=hHy_7M7AQp@k~L{fQ*l7$AaN5D>n_g
zg5vX#Ug^k&ifFpj=&h?{BGBQTQ%Ih~W&Yggw}{N~2)C!De{S9mh~KGQVg
zi8{zOxrVB@3657?Bf)e+;nz;!dt^?Tm={n+8CsO_?+M75D^Mmla4z1n>Ex@Lf0Y;kva!z5Ry}X5mR5w|!GE+b5c%))6UH{qx6*4*MU{Nv7fx16(Wz?tV9Kg@GqE+Ww7
zV-K=|Fk}~uLig0pXXPXF5OeqavMad@C4L+<1_dzs_s)1Ho-~!&Jr8lI=iTl_%P}5%
z4Vq<}=e~Kgr0@Z_wKg^pq2Sc#Y5r50#`m;$7=Y@yfN{SPh=EE_5%Ph@-|7y}eZFRR
zV(y5X+YL@-$+-+Yena)LJxuTsQh#R?4ibnH{d$GYeffrCDScJr^|UDb%+A@e-EG^L
zN@I~-7hw8RVmt~hlRmN;Mbgs_HP_7A>>sDcQTD>+6|QiXvI~ws`u(U=D3)8a9Bf(n
zJbqYRF+Y*a0BJ{oP1MKXwg*e)e&f+)EPE@Km$N&De{_bfMwLG&ai?`9q~_@*V>6o0
zyb-pri!aM`1}A9}DW5q`!$d5{+#eZkiw^9xT|@LS3lc+XBcMJx@ApkU
z>R=ulhEZq(>D~JX{=?xYjpD31r+p@6327C*B=db8D-pG
z{#Jif*13hR`G%Cn%y2W37LadS-Cm409U*S49mG!L3(@B=Yin($yU9$6dS)2;)~ypB#q}#owZqyJ%fmE
z)Z8B_?GVF^EPYb`@%4EL_GJ{k2qa1V`n~5ndFbEJPaZ}d9&skq)lo)3iNb%@F~lJ&ZK{F;%VDYRUtDHDuy-{
zcEj&$yA^@Ss*?7{T6-Dbk7H_LTaYWSv`jEO+A^`dN2e;A!GBeBtql&Uvz8ep?;XIUGF?N
z;?u@%So@(|P$$)s9udPgwgdq?oK+D60YmyF9`Bx#zRxXWn^}5wq-P;rr)7f)SQt
zf;9zamsUBL$AwgcaH)?B0ceck5x2Zv><=&zBe$&I(yTJMHEym|bv@@@ku#w&Uf8nA
z*K1;043+iJvc$DOl`_`;R-Kif9xYe6@VT%n@r+#_rpRd&yvkS@2+ltmT)fatcWg?;
zXXloh%#C%R@@-Z4J)dmt0XV;;(~{sg{rT+i)#>2%wqB1~adyq#Bi*-{4kY;WSZ8iF
z$QG=|bFTfk^eVk^e<3H=_qC<`+BLE4XC6ndKU!TJ&##>wxgG4^MlLRlp8P!V+Ncf(3gVb
z2`3e;B9|K0IPy@!yR@Ta}oXNhro1E7kJ^JAt8T9fE0MAkrETrEyfJJv-*KF{6
z1u-Sl>hsdpVhw3}z`AeG7Fmbo1pipm37r4tyu93$U$U6}a@qB5*YZ;R?c$f_R<7ej
zLY80ip{xQ5rQ`jB2WutuwZ+tg5>im}=~YUmZSW{1aApV_SoT?I^~YdqkINtjs#=?Z
z$~M;bmwB(T-ZNtdy{#i>Ah~w+TSXqy+a5V{OL64ojgcj1n{Uzq`e)eMX
z$mLdEJa}$&%^~ICTjTzt+rK8a-EHM95Gr6;SB)K{E@x*|weNj%?W7i3dpm%{BoRj5)@I-7LAU__#~zK=S9VwG1!SNSIOzz4qh$?!9pcN4Ar^je3s(VqrSploIg?~La&C#09Q{96Y~U-pY8CQJR2o(
z^dkYh)gc*|^p)va)%gMOUB2I5_79{f1}u6`)~>JNJqTTSHggQ*8DahbI-DCLJ1JpP
zL6**J0uJA(T~y
z_8pgtzFM&u7v4-fB6pPCxuO&c&O^9#cYf${uU)0`2Yh{+*K~Q#e|a>cqilLCpC*;T
zA=fRyqgz&6_Ri^WA242n!FT$Jli|_MN4M;9r3Q98d1q9|68GDZ(xUl|CkN2jkpycdJ0=%4r|>xKQ^^<_JUi@aMsA4!klTN+uKG-Q<`
zN1R<-nLk2Wqz(iwWZFNE2xVQHZ8I#a0nkG4gM*ltd#+VkCy$lee&3ky{B<$>*-tO9
zS}4@zJ>|9aRX^v>!q4XQ;)5kqma*
zPrOxQ1pHSUT?Rz~qxv=_zvY32?m)#aU$TFp8}vu(+@3O%*uxd;GXqZ5ipahFD{&9>(|8mzCdb?|%qLqI
zQ2tiq1d6vQS;CWxlfMPuV>$iWPbAZM}KvLrA(xm9~y>%&WIrF?FUGMlU@omjjwR_w>*99wryT{xx
z-Benv#F+1;1qZuJOao^X88M*vjI>t?6hLhJ
z{elJPZbp@P4-9YKdFvBBU)8@l?0WRZqmk(0V@@tz+mU4`k@v<`SHQ}XH>Sh$X
z?pGY|nobV#!_D41*n;xOsPLQba9&&Tu#@;NOX^jpxs1;uqAF7&=>i_h5xZ+W7
zcu&>Xa0{WK7+`dm8wJ@BAo9E!E!JpKSVvPVV`&C7evkHxAC
zn&mNL>ebEy9A+%AxH)y7Os$v`)+V{P=(_S6PK67?$aS%8Zqwb*>VkdjOco;tT
zmpp~G<=*!dsS{t8o^G>>&22sZCB?1qQqy;AY--}mo-ob-ISU|ewY<~Se_<;3njz^=X^St&THdA^Cz`oQpZp1ypandt*P2+zewtT%VY&Y<&7-;L%O_@4hSC?Uf>T#dDqU}3722_Cpgg46ryP);
zC~osWa9hNew^AQxq*wSJ$L{2)#mJWBj!H*Y&0`iDVX3}f;AtZP%R}%UuOxm4)uLB@
ztdnClpIe`m_(YeSIoniKt-2mKH`S>#@1xj!{>F$MHLZu@o^-&i7Qno=9@Uh{9XaAI
zsh7bolv2kj=#sBjpX)RpX&RMT@l;>$UG_h#x){AMe99v$u!*y=RpeV-`HrwNs`&&)
zq`)gqP=QtjrLcmKQi98XM;eug`Qn>OnH+lidDw!#@8#QP?_&*u=rD_vGL<7-DS?Ym
z&u;(7$ngfh+o_M?8)x(t#RmG`Uyln?d|cpvr!`xmN^x|X^!(zjxyBUB67Ph%=pz}2
z{oeoOXD}ZSw9;^Ad13A8ONhqaDKC)bxrq)3?SXX;T@ha#u?SxVj=T$H@y?$=yH|_X
zhAux40KrF|{^1f@Erxm)44Zi8d^;gruR3|hUDd@87plT?#D!{H-X6m=C|+~gI{BQl
zXd&Z$GODC{mVs>JY&oIw5oGTzDht*bTOQaoEU&LkuaiNHi_CgBSz^f^J^gra?C^85
z+-_K>oJA$iE3p*+<~~?kmo-WM4FjvXklpUCTI4z@p-#T;W>L^L%o1(k8Xz
z{PmFTvOkFS-GT)F~&8`HNV=C@q%m4p$O
z;5h3P&1joH$L#y2QBp=`U1wv$T?p90Oz61W^SJ8s2kIq^z;q~^L-F$#rR_&P@D|p5
zs>pL~`(%xA93ZElJOjPQxO>h$=V0(36RGDoOwQ1LpUf9GzdOK|F>f-kyB{*FZ!wqL
z0ZDeqdNpd#S}In%&3)WX&?R-;)4Lm*ICthj+}z%Q)%2MbWmWSzg*K{rFND%DX8}d2
z85}I^uW=`)dcD7S@)Z9{cuuY#rIxTV{*Co$D;c`ZS3B@@1^+_hdC5oZ?Aoi<0jEmi
zN>Bo|>8Oli)#A9vXIIkIBQGl{IIjMHb>`QsM^%ukl+Cn=`OPV_49!V1kBpqj>;W~9
zb{pJE;0C8)>KB{k2Oi?v9;7|uDOQ;*Zd+S=M`~5!WuK>&>_4vU0t%t*)#~O6a|ssl
z?8DblH%r8_mmcDX9hbzv-oh8Xxrc)0nK$=eK=#$Fuld_}KHX8E<{_7YI$e6s!)OMe
zB-sgk+xgn_!Y)5-sPD9R9@v2W4HK$(U0jL2f@rtf#}1IRC=KCb#oCZ!
z&n`bK?Z#fd?tq?i$tI^YCyIT(<$Hu(#V%-RC|Sjy+WFvIEcePj0Mv;IGESL5NGy+4
zIv*IMCd;ZGZV;Or@Sm=>3|?)wV;$rzEPWGX^rWXnqd_sdx9;5L8lw>(3jto3f9d;r
z^8Jn(zB8lOl?|6X;S#q!OSjCZZzUCRiRSN|#qi|It?DFpE!`4Inl7S%y?Mb*Y~$b4
zY`tAs&8^=pUqz5IA5`bjYx!CX8+x{0lGr>&NJyj_JS@kbGI?L5{zUNKrWje1{M1$<
zoXU6y#?{AlCCAQ)=Y@a3D+kNq3+>Ywf5407*aEVqRrH)qX}bBORc1X$Ur|jBnYDO}
z>tKXFPW|-@YF~8MXb$yYt?$IrrE49N=iSpXiYV`v8N_#-UA@YRlu0@=&yu@a1eaQI
zYykjv{_GQ}jgf3z3+rLS^@I;!LoNM@E7H*QVcx5Cl|&_
z4@W%k$1ET5sVmQD8cytvI*D8DD;s#3X2=^_3IxGWn56xe0t@_J1
zz$nshZv3TKFstM0`8u{`u*ocN=M!_U*?Z-jpkXbd!l#dtG~egb+!c4)uEsk0u>Ab{
zmoKuqB&@2*qbB@bFZQR|h@*)KVFSK{(Yrf}EppHwmB?6B*EUWC0X&gsf=}{XNYm`v
zXc0(E2^pzW#+lk5cfr<}$X9f>P++65`RRsQ%#Lp@sY>G@pGK6YaT_QInsfNgG|P8+
z)UG}dMkc-x(+mFbt^Q
zsW~q6i
zUuXwP$k#l6VPT8NZBP~dD{y(^$4>uqeuS|7*wI4UQu57YS=T(_%Sl3$<7Zi?xU)GY
zJEYk#x7&(K`fm>1rWB05s_e|q;HUN*ceuLU2b(rcV$%{gR`P69)0Ss?c`g>8!(|O0
zEaa?sDE@@lQ8;jWJ?Isp(s6JRwPmlOp+rzxmiKJabj<-Pi>3MA?*lk(+uSwVIS2K5
za!=Rgnog;eqbc8wQkw1n)Jvtfj<>F_X11y>>D>;P#h(kAE$>Q0rSgxzs4FFqW^D+0
zR#vfgMSHofd+^%PuJ%~JIgvWF;|O&&!ns#E#!W?S)0o>IM7pGVS(Jc>25iCEm<$(!NX(Kgh5#KDnqA(W$A_A)4Ab4=J71IzXCJi-po_>>)J53u`<-q)6|GT-e05SJYZMB$h3WB8TE;xsrjAJy_~PlY(g
zc-(hVn)a&a&sbYs$tV^DR;k~USC$AcdPI0{G5#L%Om#ac(88AHI3e}1BL`N6iZbjt
z-OztGd|fq-=lfJ&Wf@2W;$|Ng?Un3yXkY;czwEP>Jq~^`En!yQp0S^qoy4yA_tQo)
z4N^PgTZei<*(}S-qgr;z;pvgsOX}0$|4`xX@*IqoA8Qo3t!LT$$o@I*;J}3*u<4au
zO6*;x7G=%MclUrAm-0&YoX`5|f`7}MiIDWYW?{_3awWj4WV`oO6jndl`=*=qR8&F2
z`aYxbr3A(*LaGq*+gW$@Cp8kU);qp;oGRYkDl5Ed8?vL9U7OG^`}KB60%ep}vQ|g<
zeozK!Zc8u1@>Y0>`=Px-BEk6u3m?EjDE0IkThAo;uU#wIJ(QWv9cJ|~7oR7z&il?h
z)+f=SggB0DP<(&n&Efy-B7iGV?u+Dxp6@4;N~Z7PuT`tor(_)?RS1s1%WpOA<>{Zs
zSUs6ftlZhiay6(q>uXE}>3UuC|3lT8$3xxqZ~Si2rqXhwyFw}{Ya6nMO7_Bx-Be`X
z8DcC+xFbtS%AO_6FvBnggGsXQ>tO6;n;Da3#xjiGaX-)VdwTxr_3{U==KGyr+FbvsV$V?mTiJz{6^>D0(j-)++|GZyma
zAhVqQ9q%Yn4b@X#^6I`9eP{8|$c;_!j>DFN^O`0&ON_Mp%xmK(QrgAW?Z^h?`2v|r
zPUWFdA4;E2gN$-}@v{PpOdeGRvfg>c4)_ws&Mc&GETiOx=a)IXn)>oMp3~P%5k5AXGJ%I-cuC+VNE+HhaX@e^@ZM6rRRjUG>+2j?evYiuS
z=^oFry)f<{{fqCZQv#Tn#(tGKX*>JjHU|
z6*DgUR@?!;PGX$y!cbQ|suQFDQ#N^Ita;XFECA0OY+!{s{%nrNuB=1E@Eh4jo+Pvh
z^&LD%ttd+O+klFI$2C#%lAUosQeo{E_`RJs8qS}7v27BblNM+iz+y6}j(wQ40H(0YQr!}lL-$<2IAy2Wu7s(I4@
z$ue8wG;XssO#|yQpwHoLf$}EH2SLkOn4>(;CA}1`UTc+41PCEZ%dh8Y;F{LLTxSQa
zB_K|2zgH;+QS1}L_BvmJ4b4CIFp+Mj_IuXm#N7+=cdga>pGD#3KkUXBCdPxR
zuPnt!z?0f5e&*AF*h`>Bgb&?&X0OLSGr2dma{XWyWS?!*n)DKKQa|_?7qxs4f*k|YY8VSOmrPya36%H
z&SH^GnW^VD^n@Igi(x?U5=68Gv`v66lFEz@|Lz-FCDM&H;T9v_R3STyl
z2>tBJA6T5t;|4qtzWc^e9f?{EUgVSkqnTa22y36^k
z0XF1Yy;VVLM}jShI1`}-rnvl%SJQ83u~)+9Qr$;mRQWy}n%e)7khG)@=pU~f_^Us^`t+n3j
zFvPozB$p10&`jg)Fw$=0-v(qWO@Vy}e3EKvjFuMe0OvreM?{rO{BXWsAtZ<4P#exg
zVzbDI2QF=|l6hVS&hpVQ-b=_o&80maH4dp(+9tb)_1vTS6$Pv~Ow~vANDvISM8cQ1
z^+WShbW?4FIpy_|?{v-2q0kkYJzsr;5_^%M%B!7fJ6luF1kCM5E6ZIu>@knLCok(m
zVoSO97=IhXiCb3QDUGq@E--);r^
z1j`@sO#znh=N4SM+gM}v>t7|M9`XRPKi@5;-8d8$sG+~$V{Qa8xfQO4jpZ-kc#Pmm
zATR2fD@q~I?C+I${l+0BMf$cyzvJESh8xFg9vc1b-^Z3@XH{Juq(5giu!A;;Rctej
znyNKtkB5>Y97?jxYm5a4*!lS3QH#pEjs@IkrJv12XtRilp1*$_f90%-Nns~ON|9zg
z@}eQCT3JuVI$%PU*I~wx1+RFf#XZMN`~eGMXCJ_w^J}O7^nt~NKG-$|j}a~b&0maI
zA(3ZcGPad!W4ZL@>5
zvt(Xhj8Vv}kSD=SCoAzYd3|qe5llYzc#4CUm&b?LScU`IGqIFeV7xV%d~L{Ep7l%?
z?@eLg9j4vr2^IXw>o1u~-v}3w0Pa|*HkwafY1i0Ld4n|>^8BQjdOFEstY>sX8NhW{
zTsX0MQ$$6WZAzy|9+4wnxaxyk3*_%uIFieE80hH1!5hn*$B?&owkRF;EWXmUlA}jV
zd<5JI+84*SKk1xz8_rel-7%#{tWM=9X8SR!B3tBKJtCt2jjs&z-IM-sOW+bPP*LU;
zv#AIfcGscLv1%`7m&)(vel8b|K
zcmAmMS=J0+;d=k)sv-)In2Xmx!hNA}@dG#Zk!MKFEia*|*su)_65hfxaMFs;T
z7R|Aua%{x%7CPSEF@aqo`#)_YmkWgFVN1$PsTSzs0Q*Hwp~y)S=@1;)&(+V*THNc5
z^c(DcjuolU;*oS8eKSwj)-t&-4bm?~2hZnVs&m&J14mL876y&d+*8d=8l>Ag`pjG}
z^Li!UWO4K%En0jk4FxHwGK>$9Z*b>g%am;IoFQA(d-Cvoy4?t)AI>Ktye$^%R5Sg<
z8)ytD(p~x=K_)Pif-`q+)zDnd1AAh~nrGq)*|oZ$zfB}C^TlBBtc71{g%`2aJ~`dj
zqL-#>F+lsCRdmQii*wW@{ZJ)9G46Qu^mDg>IFqvS`y5hooCQb?54|
zE?TEVo3dlx_&3TWSVQ|(Kij5u*(5p7AQCw^d2%h}>7=>>RnL$EqcE6$;9cS$85|}E
zCUqXMXdad?KJGLzd7TEVhW=PVwYcL;oJcb%Yg7ca?>n?eLOBeQQd;N)7O!e;ZfqcX
zV-`~7$GHCIGUtIY|CP3CNDFQM<3lE71##^(5GGL9Qy2bOlKtEeklVlh?
zO?DyZg!B8^8LC3L_Vi{&ZGmZmIgB(X)v4#Ay`)zkI^xng%E9rH&%rcHel94_7=>f0k;gzsj
zcM$#KytAXzMQT1jDDO223!V;0-^Q-zW_1MA<4#}$CdaBBT9EcRaU{OcEhT*%x7ticEi=eUJegm>Lab41!LCu`p5J@Adc-er}aO^Mr`RjYgmf2
zfD?}8f`X=O*<8?XHUdwtm&XtH%Mtkz^ruDKTrAS@*wgTFw#0DEZ1c5cw~(`BUrAHb
z*dM7b0I}-VO_mij4E<%ujk9e}`vR1LP)noIqa3FM5?9mNRXaq3)P;C^
z%-ky|b8v0i+W}?ZwfzLqBvQs8d`;v3%UP0g884$8c*zj=^xVWxdr6c5YI-vI-@H2JNbqRC7)t>;a8O}eSX#Iz4D(oezhlgb>|6lz>*IuR=z5&JA(~e8szQE
z4heW?;X&A0VtI`}m4IFRKDBaj5Im_;H3nA*~!Fr+tsx3
zQdaof+-chLj^UD1M@%{%{)s)pV~A~UgHOpLAqSYq7s)7l4fvzCY6Ehp^yTv%c9Rf8
zywjnE`U491lqdmZB|~o7vX%DdoA*ZuX77@wXLzeuIZb37dLQCP&$11qLk=A4&WyP-
zmOEIfD_a-1v*PmjunC#UVo-A4oUS5%{4VmR+1&y94M&z4!JGCik}{u@1|6#oc)#fk
z6mY+VYxHxO-&6{=y2op=)o(@Bc6
z=K&M2vo))K$Ry{;v|i$E4YCV`jg$OYRlUtLOYw?Zi^>}dzYRV>nr|(doGY!`-!8wV
zrC{ewtP9n5IU7Y4q#(V?cam+7us?rGY{_1iY`EZTf?{8mAhxM(8
zU-$7k19*PC&>()s|1>7iW=x1+u+_F5tLE#iCv5Hf4w7qOWoM|h`;;3*>&S3>$r2%=+L>o}D}Iys9zEO3w0RQw
z;@i;TXZt;SA9IFabq0|=Kj=k*yLOaq_FixaHow8)aD)wpON@p(q05$K6N+tKZC?oAtB@peHxBRHT+m;mMfOS#;Ib
za!bY5^0Xex*sk|`=9{*?1$1JX`EXtR9TxSMOw98s$$!i0$y<%@F5yU)%$oiPpy@kv
zgP4WezzlCj74WA2Au%U1{=`bp_UI{Fc=LtXEBdWIisWGu+}u_;JAK_S
zM5c0d(4{Pu9}M-An)>%L-r}t_!wk=6-^?n%oma+f@#kl)vaeSjC%(OwrK%d5t93FX
zPVn~M+B+oraL#IS)tW3K%{^_!zn{`asdniwb;N92vJ`Y@{nuO%5|g}GC##UTgwO@`
zsI%EOtf+|Z-`hGUWXe%-e>M$2CKqV_MKx3vv*CZp1hHYl?SXobs$HCOXrO0*hS}QU
zNFQK^8peQKqNlWB9d%yz=I3nd05L(z>(XzjI?sI^2hzAsCkCthw#iI?eaT^Oq#40`
zv-D9BZktL{tpXK$}o}`=(yKW
zCt_xHeW2NlLzKR}c9LeACu;AxH?lDY)Sgx=?#Mk>Cz4$EwjTWrU%g%-NR@(Hf)Q1Q|wFjuA}
zZ@iAqo6j%dbJcN{bjhjXqy<-pl21!&8%9l~41QAHSvTN0u9Ia_H}R}a>?SW^dx<50
z^cnsOT{GmeU4irzuobey|78XF!aUtMQt4j2VCYT^8U`rved@iT#6CxB!Cxt6^HrJz
z=y6F9#x0i1DMiE%?JW;K@BQ$i*@?fV;32`GBjT(4BhNzf69O55s@%BKmRq{}g0>k(
zj0$YROqIge4U02s{o-d!dc*0mQ{1G{PCc5OtfyRH)Uzr}fgtZ!(ehm;mr0`5Bd!l&
zgFyzbN?G(WSmuM_*)#n_Yjm^8YeqdI7HSC{b>Asx&lfZ+V&Haf
zM(9e>xxnYp0#ok8f2-qxIb{A`n75kHiME&TnCk~9{ITmP~3`KHH5Rc5dKU4?W^RI0l!634~4&sf9_k`pkJZk5s0K(
zJvm!pJV2(0e}o1$oA5W3Uu8XUu&E)+bQeym8VqgHWJ_lo?mWs6bWBybwO*H)2}lxiDYtp82eO
zr9COKx4H;%eK}r8lJYX-_%B?vwH!BS{mEbWA&_s8=+3)8Fpi8CN|KZGK|gyG5ubSxwU^-=7}yv){L0k<*8R^}Q{
zKHL$|fZc!DCQob@gLHw8V3#!Vy&t)HWt@+x?UX!udAtItS&>}q*vepSEA
z!GNFfM%G`(tcT)_Ec4dJ@E5Z~$>7xCH4(}y#~u~6c7`;^$hZ5-XjkHNz!Y=o!JV|l
z&p_5eX3+6z+}uZuipMv9yg5B3(QJvjWz+h1<>JW;Y#2XHgrpQ
zbM8{ZO{b`pS>9^-R?)ct8je(P0Joi-G^Vrx5?+3_A+U3lH6YkEBkR>=5K&}D#R}ih
z{DWNl`H>P8R~x^mk!1AL`c+a|w6L%NW=jp5^opdC7R*S_O1fOdy*mDTtMP^K{Zx~J
z#W62#r8oOHk&@vPz0@oE2)0e-0=cyRcpN{o+>Yh*!%JgS;O=_z)d%=Up~0vW$%k^1
z4~R-}i`tWJOxG`i8sLJo=}yrj?e>t1A)NKoW{pGDUn~;R4+5lZ;>>>wL}PpQ!o>*J
z=3l+EU9wmCw5h1UbF^^B;~6EJY6tZOYPqZ(RiZ2%?YMhM-#FOUy-b`?DZva{5F_FSe@KE)%>n88*BQ
z1C3ALNM3fxdY(e{iGY0Cu^Uw1bYLVl(*IS5MXdqW!TexJlFAwQ}PYy%V+4QZGf6lSf=Q94i&Y>*4g@NAbA9;~f9
zpT;HVC`nwJYM|^gxy>EO=rNFDik78*@~*lMemA_+@mAjb9Q*5nj(WzNPm(PkZG3Z$
z1Fva1F0|nHuKMYumx_G#BEEHw9m2mjd@5z%Z*N4!9bkR>f%Hn7$sA)!2pvIcuW??l
z*CS=enQCsjs>~WqDID%fjc2_~dtdzM(LFS=jq8m+V(mwNioi$#P?={q^p|
ztZ{BJz~^aLY9IZt{og3Nhtr6hrs$8}`us}xef5=VpOX{=_$5|Husm3{+sMJgyw|lu
z9Yj?7(}I@Ezv?Ck3y*6j?PO{ou2aco;Fzj{#u?
zX1hfi(Ow|vUSKgfem)n-iw8!3TmWnEP3OL^nMq?WeJkrm0CD06z)QZ73a|MLQMB8r
z$t=)TmYakj<=CBuZYSj@;2IXPEE(=ZDmIdz`DC>^U=>qGE>%_D1v%k094fBH$Gr}b
z+5u<6j8>1^+L%p4S`bPdiCvlM2)5V;WXm|6C$^7;SI``zu|PyLKu5>OX1Aqp3lYOd`gBm7X>GdY02Q_<2L%v`3n
z1k4_U21Xl~#VC%p?wue~S4mJ4ap-z^N|t*K*UvMLHTCJQ#$O!y>Z!dHrT7lgJr&1(
zShHvOhP8!$_Zf}0y_?_>9D3mM|4MyuGWQ+A_!w+w)%^faZ4y2}v)ebWItrYgcIQ>?
zf1D)mKz*3vj%M{t<9)Un2{3raFN72`xX8ZzHmLYa5V6p74T>8G@vh~;20CmRRD4Mr
z5l&-63m!-*N+Gbb!Tf(Q{$8w{muToHc(wf6wmv5)m}y-6*R4Lx`&(K)c}D7Bx=_62
zSQwJecEYvGmw%xwJJ(vqzp!
zjPK(r5ysZLfGO1sPP?Kb7S59o)+|Zm!n-_IVq5@O(H+7yWZ?MT(Gu~L8=WNMy7*}r
z?7{IKT*|{8@9jACOg?}~HkT6>=z=(l;=q}Mb1TJC$mjVF0q5Bw02!#u)g!+uKK;HQ
zQeY~@eyy?k5Z>qq*42t%-#Ph2!7Sj{@=+XVhDW2m0lYy7aoO;u^W={=;63ZOGj%lT
z%+oPjUeJH@zN?*3GZ-R5gp0sV
zltk~GhXvW%ad*4Ig@ofZ9CvwbDL&p1Bxg$*l%}D+?@BlWw_Be&H*`W
zlEaSIY?Nbsb+>H|jb(srNdBIE2kx{#2S7;z#9SOorEA$Tse{<-o?a-E;;;OEjboXo
zJY=o6+vVxGp0=}+mYlcE&d3b*BO#&%=H=Zt+JM;5SNLURQB1Y}O{89uv<^t(Y7HM2
zi&d~`J;N<&so(;m6xy@_kluRC-Ub~nW}@#o(>hr6025nWFxRCn0Es#~AF%0yRO@FZ
z7Lp!-TNfMs$feG5KR+0l%X9|9b~`FGoEJf@f;cBZ9S^3pham^$(=LE~;I&16KL1Zy
zj+bTWo7ep7q2hFXjg}SeSHHK1>|Q(L5f^>!khOe)m$r`)xwQOlgvKN(?@tYKbSmwA&J@S}&rtRS!R9{B7|%UW5ut3N`gjG=wfY
zOO$;kxy)Gz&Rb{*3@Gjd$9((a|KD41%a&TYY0YW(=*VJoEu~TM`QqU_1^MW>Sooho
z`LQ^~&~h`fOX3=~*e!!3Vfw8|KfV^SR2YHuYkiC~-G0dTm~n{q_!*aP=$=}B>|$p|
z2u~~G))&q-SS>pnnt)$DPxxmWOYN&@Hm3~d6jv}P3x40;;iJ0u`6!LA4_<@o^{c(#
z`iYHGd=Fki{$9t}uZr&m5zo;C-*kzp?&=}QVSPUAXr+Cz#HQ_~0+_A9$xTP-(jeOq
zsjmTvL7&}7@tB(kcSx}CdCb@s?X7(ZK)?xKQaWy}q%p%!WExFv>XbYy5735|9Xlth
z%`4@}^8w@)Z~G1jgBmOEXR~`e1g)#!6lh9d1tK`jVEcj=srJW22e#t$sqF?Ob|mu#
zl9Hu5R9Qb9BBx#Fvm$XJ0VU|TlVJ2K)n~*X&(FW%_A6N9ZQ+T9mPlB)1YtmNDU@7x
zuwlMm5HHKFR;@dd$moyMIN!I^3EqIxg;wTJqzj`xFNakhwaoSTRchXG6WG`aEb&Iu
z!l^R62P@uSr6qBGZYmlY?*Q?7va!zs(?iSj1dkDvqe#g&Ts+-gN~KJT4v^614LHz8
zME6$hzxLv3mhTbJWO~3mNeS%2r7oxsTY!1EqdUj~$?nol1LJ301T8sNz7}W8SvzCI2{o50vTs?_e)<4L$anO^B&jzL
zQ(+MV;ssF@o%65rx0`te19@&1ro$V~?A>1wN&Q*#=C>-8*bm?G;w)osseOL6P_M!F
z_rI&E{*Cb5ZQQ@LKVLxXkTFa0;92bZ)@~2b3|fQWSQ10H>Bqmfs6|NMF9sjs(7DY#
z>%6#XJXR}70u3puwQL$%JA8n;&T=JE-X=0|6Y{pD88!Eh-5wv?-q&ti#u~v7PyJ=b
zCu5cGkYd#tkMJ|!U1tO2~h^=MfGK5VJxqK&{v_g`4%GG&ZN-0d)mO*<9=ncu28
zC}3D%+8%W8Bv$tmA@^7bK{!lvlT~MI{4yQ>m~in9?6(yi4Crw}+BRy?Jn3Qo=~cX@O3AhMv}f(8kPGXa%x7pmNd)8tnh8
zfQia-GrR5+ELT!F1aEb0nL!O#GrwV;&;S@3u^%4Xb7uoC8GX)Abs-rD}gonGxDN4+x&+$3;xApa~%HGFctJjZ+ES*9?UZA|=&jIANf*
z{~j5Dd^sR2E*mD`GwhBu%+vou#7DHmC@=a?J$8v+8J
zFNhB%LKx|sU^XThUujE{At515llv1N+hZAjpS*ODq~?V($cbJAf{opiK7Rtk1wtaG;@o|JV4D0|bIyisc_Qr-W@n^L{&GNdlOMXqs
z;9Agx{nc>}M(8QizX;%#QI`3Lk)+Z}25RZmPW6{b0ymda<;qLtz#TPVBgMMDVdfH!
zYt}L56yi_0`WWS{T!R=7E}yZDPCmJ{@1FZGO{iDyDLo5wUDsmn%z%TmBjIF?%R@KE
zj@BU4c;uh$v-<~Q%Zq0nwkuqZ92?C1u~<=7SBLJ&x@oOQh1K4kcxoFxmmg-kkgA>%
zPnb$cNH74?s5k#?nqRQe&Z=j_H49G*J@{Og;Aoy7d)URYaABpZz%kPea*Y;p
z!zQdCK__yl4s8hco}*x7=%0b~san7~
zm|(_6S*+TF{4ef;;eeeRPFpd(LsG49q1ExidIN#~8#E!$DA8=-v5R*3*L)GWlj$nx
z%84LIDbJozlz?*Ba0VubKpYxQAeN+_{gg41^6#&mT1bO;u)K$4Q~c
zDr`)@N1B&=s3PnH+4xbvmK*&d!Fs7w_JPP)xN7hHfLIG{u--IkN%u1FOi;`lXw-$AddmZ~CXo=+zP_1~LG#K#=z!b-n=Sm!M2!17$UR~g!
zNEQ~6UnXG)_fHGJmCDJRqubzfW^EPo)mHBPg40E2yWgrmU`YadQ)c)g+(t*t+(EEs
z4hmWui@q;4nV++aPiq@PEw}?=@n6mnuRN2m6GcKgHl{6W0d
z%EB*im4VdSI84@M@^$oAS8MxEb%G}!i5vj1`uHtA>tH){-23u3j+!y`j%7G2bgJ$zg0W+TwkE8V4DYBX!*tM-M=Pk
z_Z3V44rRBTW8usc_wiCI^4Sct&X@QC&*l%;(2hE`|9?KVh;i(rR_z@qdQ@=A7XSn1
z1mk!&jAK|VJY6fkJ&5zi5mTaYlS?J}(3$_2T7W(O+Hzb+W3i)n3js>)O%o<>6`hkA
zp&wwg?n0sSQGW^~%>Pf*PO#G<Ohpq
z3yC~C7rir3>l`A5d6L$q|0lROB>sDw)pEWDGB6*ZEBvWqwf&V(Z*|j8bnSvnF6N!p
zB7Cd$6^iEx+_7g?+W-uS{*qRd4l}qe&2x9839Lo(nsHD4^M=fHRRwUh`F3sWN3cFa
zF5ul`h9~;G?C4Qc7RDT+Hw;QTy{on2f(KMR&eFe7Xe4
z#G97&yO}RAwp2ikqah>x$3bO=z{)`Z|MV6~pP(LxTM8?!x8&w7u#}dA+C(DFINb51
zSKGti@oHg5-}t(5wwewM6;uJNUc{_4pIFDciq#~Sw$JgPUTu!ogfIX^SjOLAsSbW&
zAx0|Sf&k^U@OtM6!;1{{cTu*06N3!icW_e<4dHX`L&n>Vc>zKx9XRcsuQ)i(V_eZp
z#Epl=3p3E4!v%FnH{C|Lzi$R)?tjKnP%vE0B3^v*`<@hh{qV44oleTFRYMT(q#-5w?E&2APNQ}1>jpjeFsENfKO~8H9!1{
zEb3Il&{yL@<@3s7UM&mgl6>p2br2`>jA(gM!Grf2jN3YvYE!!;a
zlcQU&{i4hr&V;Px2h}Iv_`LFIZP_1g-%|%?>sHAj5j#)nxRqJcw=CQe5nqTk_~Xz>
z@5Qwxu%V57xqndnlP%-?5xgY{mWir>aohP+il~t4@8a(~CFx}8{uTZ7*N?_Hxi6wY
z2h*3k-Z(gK#TbqU&IzssU1m@R4H7K8KT{OI@$EgN-P_Sy?PZ%IhG5+0vbQivl}FNB
z>E@66-9|QU5koiar}1=@Zc1}3*`-4wZUXzi#{URo^9NXUuP&um;C
zTKp0Rc4$-3D{~2GuBprbkF8R-iSK`Bn#qG=5-F!oYd!IYO&Zp+)wP3-RUAcqc1{KU
z*9$&$;eeegWq+_I>FV=elVpso%uj`Od3y&fYKC#dBGVaxf`X
zY@{Zc&#qP2oGQrR?`WXRHihNCmyH2Lu>o&t;5#rPtzF*^yZO91YrD?}Sgs2%_Z@dTKMQJR
zd#0Nf_3*fbd7h7JNQFQ!AQ3*t)5Sm3OG5%Nh~jM|@n1ltxg{Z8F)
z$baK!0+#d|GU@FIoga32d-ZX~>)G=>R|4mS){bUdSV6f*XWShHg0_lW$|eqwL@51Z
zsxv&&K909++k4YhLKH}FP(Ic8y{;}|Fes7j0=++6pBQC^TTfwaf&%Chn>7bhT%D-H
zG`dyQ1DZ}T%-1aIy&`!_MqzfRgq5guY)EnFGym1pCIl?#8}`1HR??Yn>PpV9CT
z4cGi8r3UYtb5w*
zdQ-75$}h?RtsKOP!oz(H)-qljSo&$I^LZK$d+Oc~2a%$t)P{Up+dQ9erPQ46QY$kC
zUvES4UhH$-{D`tpYL#i_H-9h{TE-9e*$Q`0wcoO4k=9Ujj%$;vO4?>%xIcg#OJ1Em
zvz9Z!%8LNP@8Zw*jN>1zJwM<-8cI?`IjjF@NJ8I_pp)pXZ{NWVTpwoMDTf^yO_9!n^~ibd}xAgV+Zb8Q<0vc^Q{7KUT)p>-@G~;I!?f
z$R*F-jk*F%Hd;t(5v1#P^GW#B;@(y_Ai`$;Z7Z0;0}HB3usjQL(b7YB1`w#imzu<#FoJ
z5Y|^~I?;5Qc}ge1-(h+!a;!hY!ZXFkU2Tv;8!(l+NHohJ&#UkBtVlgYlU*<7nbID@
z7e?|re}S)*bWO2AYTrwjeO7006Y@W|-@&IISDJnjKZ+`mb{)=Vz%Vg}AJY_Eu(OHI
z-3@6e=~KET2ao#!kuVUGL1l=VOWa0K9@f1B9+GO8)o~P3&dZJpo8)OO5ovuR&B-GH
z3f30gP`91p{d`=W0{A%*O<0s`OQkWU)5~&Z2fujEwpW00cS`qefEIX
zW6T9Y=$brSWJ}&UHKsswl@&P_Bm3j~yAW4#bGPeh;F9kT)b6j5sNl3C6)KfJi}I88
zWGQfJLe~-)6l8mVoP40oo<5Wr`~!!xLEk^4^oys@h8?PMjCOJyPbm8F0Wny!3QfEM
zX%W4xhXp~rU||G5(P{gBLAxv8VVxW-1h_lzZFThDC|0GJyEcH=Y-k2CStRV$^DW81
zTZAVp+V<9D@JWXkOTN)o-VP%sz-AO*G#y4(34j+IgaPlR;_d#lFO)m}Ru>?dK_pCVl6}?&7w$ts7%M9QY@H3!k54
zhe-@Y#xlk(cI|Q!E`9YQiJ{&F5e$CB3+2RTv+mwiyAejX1S(9ycX|3DH%Lx;&S)Py
zI~~CLq>`kv+97^>xC*g$w-z*0N`aM8wDSz6-P=jI<~gEvs5*f4DL#2hywTqo6UHTg
z2CYWLk9?E`9nU?9&&b?2F
zdD?*!o;_fE(Bk*?@Q>iaBLh~Ul5Qr$A#syNlp_?A&>#zNdSYG$=5K|P7hSD_^$`&_
zI>=UIQC@wWRaea^GOZ5JNBGQTnsl^#2)$UD;P(lp>
zP;=r>oK+|<#QOv(W8la`@iR~8PG|Hf%ID60OI^|xTJ}`!ul|G`0u3HK>z+5PowfHJ
zINAvmU#f`Bmj-YRbnc6#$-|LDlFq2(TWnSlJ(d&EIP?eAe6gD4(s9D&MAm2(0tKh^
zSBIa#vXdNUV{sN~Buhj=&J?H*%mSkSrZ*7C(=I`Zh#U1)58rdx&MlUy{{4}|ZwxOf=6B>Da@Q`wOCSid;2AUZCCwiZdz}A
z=W^RS!?tWfyuaku+@T=P?*F;y_U5HWYWh^J?rviANv7i6O|IIL|0~y&aF)d1?*Dm}
zX=j0Yh*zSCUadF?%6oBZ094p*&U=kMi~5A<8QFnGP0=zr?A_=!iE{3V
zv*sJ;v7^7X4|K$Cl=G&dCUAC7T6p>j9)nGdJno=8@p%bpLRM^erb3wran)Hr
zTe#1Au1m$pt7_Qi@2w>GSnm8_!V9zyO9KEVDae(=iy>;5ZM=_fC
zocVK%IBNr&UV2sr7L++^wAMv%HPZt2tG&yF_
z>S&YbOMRfiV=gUdEAOP5E2-?@`Nl3Va_&#RYqK3$n%
zUT$|Q1;^h!y)-3sBmsC&eKhh8;|nLl
z1auELLPm?*|+p9ey
zu9aNQu8Ve3+|HvzDhBx63|sEHruZdn>7s
zC+i~niZa)%_%cN832m-VX&@MSPvrDLdmi*IVD%hRl}>uN>Q05r*tC9n_}s=UCe4Nv
zS2aVEn(&^ru9VH6B=C7|tuIwao1o2QEj-(#96Hl$qNhM=10Ef*1kA5DugTlDQPzrX
z<9A784djLfPkvV^
z{6`oJr7IRLw0MpN)R9axv~^2Cb(m-fa(Z1S*v&tjiRf
ziazY9#C-&u4)j<`og876mVpi+>`!!@ZRc@IW&ZFQk)D=z9;~b*DeP+ME%JEMU$d!?
zbBP>-7>6zua1!c;^J8Lh@X^f`=*SD&BhuJGvalzoiLwksmO)==4YWY0DdQ;CjGtRL
zprNTVZ@m-w$!qEH#Sb>t;`1bs6%n>trDAa9&X_~1ALYIYZX>UHsfEcZ9rst
z&&DEb`(@_HrVU*gTW!Qow6aXTS7R?X>LqSka<%#oMSo7?6Dyzz(1!-vqYWYX#$9XX
zOL8tlCD%@Am!>c7mSCL{(#!?s+Phdc(3b%L4C4!f*E6%bA&tw!EO4l>1JypeWr7xJ
z+1@TZCHX)wy?nX6Ly7fz=DhL4O%8BQ5f(Qudjv+nC_~Q}w3xGs1L1*)VeScEzEr=3
zydhBPtY2}R75;vEv_zx>>O9o{BnVV_!T9{}@)D{z!D&Wz
z+y;_+HyAh=K;Tzr4={p&;iDey*q{JvIWG|p>W_ebV;ajrrCY-00t#b&fJyked!`wa
z@j=0qvwevBY2G%!1h6s2-6u`V&K&&B
z;QJ%EY%kB=hiGfvVYNGQM?Y9Y@^=m?efal1-2VXv=JtzmuLidu7vp7BS2HB0(>Lnu+^58b7h%l>5vb78
zgeszWn*Y`t@4@U(nsmqmbtvl8VKl$tM7n=Y>->sYv0v0os~<$5lBJR*vQ!~OStXZ_
zU98!x6&3iZb{vx%5zd#<3=u_b)zom_UVLjMo9jwsLt*~arm*}?)pJ2ZS4ICUGa9@p
zm%scJ>Ggl?y=7FE>-IOQ2udg-h={a=NJ@7o(j{F|QUVf6BlReV2}lW2A|c(~CMW^|
z(rM5oQqpJMJb>t4?|S$9Kj*_4XPggv4|TzP)$IA3bDGjmLCskFlERRiws_a8(m8|J
zY)A-7U3SK2LO_&5N6%8ZQBdejXzSFtSFbPOB{jPYc)S1urR77QX!HFcW?ZaanYdnl
z5+5c?h^)ug-W{PG`QFBB*PS70$JUaURKEaSpsfN4>D8TQ@&t!V5;R8U1)1`-A|pns
zEx#qpeCo;?%GC32j=dRv>Ah-_qy1alv>WQgVg1aWUFXYrCwoV#ISsQLpOP)djGhoK
z9?0dfq9nqJjGJ!?QMVnq5#QUYzf!)0udyYMR~XE%0)SO*Uy_)qoh*%;p659@b>EscG!L5dpM%1-eGF@oK@IvGo_GB
zf~6v+q5PSH85K+tv3Y?XQlQWhtHONpwB=Q3P=nfQ=+$KVhzPYa=~B(Oh~HYD6o}J&
zYh!%t8|iO6KLR!IRjK(P(tolrCo_2RdO@7u^_Ol%41lQV|72LS@mxf1^5s(NS!Xjr
zNET~rnzkxOsh25OhIMi+NMha2ppUEy_v~7nD2f*KH*NG@)h@UEf+YB|#Wq#DJU^Up
zJIbd9UBXIQTEw81fSU};*@RT@36%9g$`fW}7(5MZoZkwoN+Pq2G+o?oC610Lb~z*k
z)Q_j}^Y007VgDlTlvJI&h1>cvj+S(lF^W!vZj3=z?3X^lA9H5+p~#Q9D!1gSSbJVp
zjUw0a&SHn;WY!Gz#1=@L@h67c@w_au%E-4rueyK-d7TTIZKng2ukO0g!yhBN
z!3uA6bxhu3G)5+vu9aWb`iNf9LFc7kaBThfhd(!ld5n?mJ=5ueP)D;Ml!_;-_PTH4
z``}^O6_IGqbuV$uSl=-=6VkD69xn
z*>hCC2`6S2%D}h*x>(x@{j^e{&ks(+dup^__#9@wS*$<(l{47vBEpR8ow8z;b
zU_6BpC{uQNT>XixVA<7ZFUCfI3o;g=B+YHu%aKAPcS8@DxpC;!H3C;l%}xQ~}#iZhosfRCpIiPgFeH>&pCbLC6}*r|q95
z4h1g;2*$?;-}zoIJ6z)iauB6>hmS4`WhY;M6>~A(bMV2)_`=5$=
zSD7qAx3o07fhtRWCbg}C-s2iD#X@Ib*@V7wB=5d$@O1skxb}Z6yXTpxg9IVEFIB4G
z5}mlD=5o^fKCO7&Q+f1wW)wM$E{-81gRu2`&!_y?pF^-rC&u;bJ;6br1M`s1zxw5s
zpKkuU*c0Q|kfA5|f}_vq0zYgIjNO0{cog^3zgWtWNM@R5Bre8lQ#{z)$p4RSWypU}
zr9CI5TAq>``}i{>Pb=30WjKsz{OI=mYGVP-i12lHGHI&47tGKyxYOZ;G*Z0P?xhP2
zvAC~(q88I1+oCTYA3ONfs??UQW>JNLqr@(@59ned8gFy*?IqFQ#Jc_7#wZuM7y>5+
zW$MbJ$_Uyfr#%$A_eI|0OtOth%ifeyRMTmB@m{fE)Cbt`z|O2v
zR1VH}@Y7WKbyX{KY;FZE@FO;asU(S|7Z!vUk!+QoZK))5&U`vxIfkALvUB>rSuR|3ggN5b|Q%9}HJtlO``
zSRy9x06(BkMhsR4~TNYH(OC?kXK4k>?{r%{F
zZ0H_dFI%ee=O0n@>&&xfNh6Ul&bw7%;@ZIA{$^*cWW!)q+exBybTYR_4fc4IEr207
zSt`>!Jn>+>4h`O)A=W>9&;6K$B(5qgr7EwZ2|jSk{xG_uF*OTZ{OdfT
z+>Y~gf_#hjG|+|mgJ+={ft#e3Cb;UZZi7}ZK7IqP&h{kxf4<+KlA0&MOc#%2Gg>-*
zRpaD54^M;hisSIsL!okDST=sW(i;QNRlS{OH1BPZqz^LS@RB~cH66wA!sqKS*)2^Y
z@8mxd44QPPBWkLGs~()n(^OqHYW*Z63AWD_Qe4$E!d
zXe}b=wsfOlExS^9AX@y0!jIs$*&48py-XoqaKJI}w`ykjGxz$NnYaFQm(~b1eoVzv
zP;TYu4V_~k%!r-w(Mno;9P_7$=@g?KtEIvCFHT;funaD#1{8XGGJ!w@^Cb-Y*37a*
z3Gez(kF8*inEZi``J0Id4x*yy^+^hY+Jj@UUw^$ncnZc|y;K|t-Llbs%2N+WvtHEJWsdHbjrlKLsl-tAMG6gxoI6`5_q@_$*z580
zoJY`a(S=NOW?eR_g=sz?AM~SN*H;Nyq$m*~Ax&;s)%r`B!EgUO!R`0}qa7^39+E(Q
z=p|-s=;zZ`b)><=2cz!vpG8(UEt8C9aB4k-E}Sc<3huW-Ka}>cx7W9=(q}bMh?$DiTUT1^oYXvbuzz!rMW>PFYL=54LF-AK|1j#x$7~
zek;inWuhGF>85+?xc@o|{8juMs5qSz_YvpBMBb(&Pa8a0jFrmpwh(*`Vub53<*L0Ioz27X
zpSKfl#Ou;O_*bcV3+Gx;f%4~|N;(@{-XoQ2$gF9}AO$qfixdmBsqNg!}_KHmGQHhIewdy0{x?rVC%(L
z&b|qdt`p?olR_K41g8-4?7BaN_eJ8}CmS5RKM_rTY#v32^=pGxw&wUz=x=`IQt~z(
zv?6rC8rwMw`hWNxBl>nPfQc%|t(WUb<573AjMezxo-HVG-+(tAPQ=$QpqCDRa716E
zsV;Tkg?6WesrzQ9xE|F4xLWEm+6NM+CPJ2sCm%ftAHVyc5&aAQK@G@T1|7ig*gTe{
z2=l|T){S`3KZ$t)VbpJ)(5<`X$$m+9
zaqFv(X#V;t_yGXmO>{$r<%#*owRa61&_vrV;;QiNkgxK!Mxv$){!9*gI{W-#+`#_C
zd68R=T4^+a<~1E&Ij7JUih&0vrl4WJFkv3Phb7lGn
z*y>JIPM}}_wY8Y`wm3!s5g%fx|F92y6629)=nG0j#PK+b8aELYXY#J(`Ac_gt3qz%
zDxPNEWr4A-C5yM)(9b}ijl7;if^wBMr3ND%09jkiw|VOeBA&9=>k|H(EyJ!7Ttv>%
zOxRByO5YafIpUOl85ASHi$8{T%gbb*0M|IaXAowm;$%>$6-VFNAN0HPeAuZ#Z@$(h
zU@7jshiz?}YG9N!dGuMZ)8dj3G5Rn;q44d*Uo9LFM@d`bKScixS!WO_`N|bT1Y-d8MfhyMJ%rdB`IL#kEJ@7AnsQ4XBpbCjkqDLj_
zy3ZW%Lmvz~9X6ITt?>-z^i09pv4w
z8%8@iG7<3{%1nu3pByD6N;aYX7mehrzV%fo+WzCKz!c7aRN`jOQC&P#Qi;>Jk*2+W
z`!jURLCSo7^wy8ZW69ouVD>LxG8P*`t+s^6={(A;bPNceEnH{*CvN5sBRTH8)VY|IFppmsV)zlL-@N
z8}@1IZh5`16`ooQ#rFJ*}Tc;B^gFe7s#m^PpS&Nzu
zguX=HF_zuoOW5)Iz^1cT(-X>P{)vtrvFTZK|EM2nfLfr0K#A-@SX^OFkfjO3t$2?5y=qGKdg3F7emN
z-Aw6m3iytsMYe$2&hpeLaIS;I3443;?(B8vs(kek!B8cUadc;`x&ZOM(bq7dwo%Gn
zefLu&nDF!@&i8o!jot~6yM8XYThI7x6q~;vTFS8rMm0t5{^I|PikvYogMSQz*oVz5
zJjsc7b}LSO5g~%-4Mu1GKH45447k2?EahEnZZ+e!wn|=Q(R|uJ&(kTK1=t`
z8k-XC9Q&Uh@GM8_uLvl
z%nmKaiUl_D&#EzNW<+>~EA4Iw_JLX>fNidIN;hKcyq`?@KB~hnVj>Rgo=1QgJ}gk~
z#IhYOMcxAS3wZ4tna#}ky6$RXk)6%KN9y(?_TKMg0NC#*5qV=wqJ3cZfs_!w9I49L
z_B^+4L>2C*TZ|S8u{7t=0j#aZrapw&d53uK(eKi$F#*`icl3>p?3`7cACZI6tFO{_
z?iMy8b8jDf3kI*P>oxD0!;B(Fj!fQ37rxuo)ECKZopeU=>>l~caS+h<(|GB7oIh%M
z#5U!E*4%cI%6sT>%Raba6rz24_jm_^{){`Z+IMA@uNnxD*QM+kXy0|~u2G@jz%Dk)
z{D0sAP0v5Go#%iFdd4LDbUU{YL<@2tf|l$p=N{^ix#$yA=*9`&GW98!8vUPH6<0Z1LfyO4J=hVT8|(-`+)g!0C+1vK-Si^DCSE=!jS
zWqe%6dE*L{vqg6@oY?czK88S^KbR%Oc3$w=RU8<}0r@ie24-EhXx5$W`}jNnVSPhq
zmnOh3Cn|$j9C7
zYh%D%oPOaO&A&fdg7ZW4@x(6o0cQ!Hz&@2%1&uA*UMN#3KSpNbLc+-4zg+SSs_7lAW?{I;7kHUvha>l4*kV*De=pl3W~l?Q
zZb7dAw-E}YVQLoC9B-uLnbp6WB1y(a5<&gY`RanaReMGxgK}&Iva-Krtv257I{PqW
z@qkELd=+-t>i7{v!OgH
zzUr5HfM(*{)IsY#Av$WH+Rewu%<*@1$M-59aIS$q+UnQUz8(I3x8QV0yrbJsy2pqU
zkHA{HlL2!)(l%GCH(``>)Yn+0^_7JMXmareIVhO>1!NgmJHi&8`yoR4QV{~XX_25+
zj1zgudsM<810vW-eqSy2_gtTWpaqy_^EpF@Yk)9aE4v+VFO%ioMM7XEEaKaH?A!;u
z*S1kT;Ct19*Alx~@0
z>bDu`hbQ9!IBPzZaf>QE$jP2d>Nho02Yo}X?;Ba&&t0NdfQU?ED0l7dCErj6Vd!Aj
zq|LGCZL5&HVD-v4b4S`|WZ(%Db|*k_m_UFQbdQ3{+K)?ud3}!|knG~~ZC619F<(ct
z4ZpylUAY0+O-OW2lcn#%a5xdg7>(AanE{h9pFNjucIMXZ~Y+6$#N0Neul@caaAps=b&nz{OIyRcAjB441;BV3#$0C;@yb`;^{N?Ss<&d%r0d;&yt-d$GFh2IsNAM?eIH
zyysl_PPiW16tRA}?l#2l`Ng^d@rs~@SAKX-P>~`O-WD%$XnjO%^WNGz%xG(aq3BRY
zE0V2~FZK2TiSl@X*XFv52+V~&#c%(N*XC!OU6x^W)g63_G&N
zzJ;ti{bSU_U0Q)9QI`;>Ev4x8@AStPVlV*uD~-)|Zd1GnF#vjmYG!xH4lw`{V2_0w
zu7GeXFu{V5iy9`H(sJ2iwE14^$PF1rndJq}ug
zl7uo7e8@gJa1!icSC3<0J$Q7DOeUD;W8n3ki0In6A(RIp$OeyY#m2p?&WO?m
zti3Xoea-%W@n$H9{6u2t5s%2C+G|ri#T_uyfT9X1rv&Wqlb~{FL2vaLj5e98YmK*&
zZh%nIX;*6qvGu2+I6EsL#N~wWzygKo`tEhOW0KDhK?wcOQ=Od)L3t1$Tk4c--2eMH
z`;WqwCZO9!s#$$_z~kG8j(?Hgv1p7=y>o=DjM5AmC^5m*=Vv0nKs{&x1RBaj
zxra-bG|~`gwk~vh>0KUaipz!+w@3^7y`Y82;NdqSCn@%9P@MWdebE0S?;%l_TC=X~
z{OyBC6PIs6$e0KX=umgT+b(|8d+i%lKw3w;<=U@VE0DO)l8%z|UxLXEmt@o=Z^Lf#N!0a2M1rPz!D+qV%f*=T328N$ZiSQgxeK9V!
zx!e>W3f%#++1`J!+N2yc{SyRhjKrMIP2*bc+K74VXcuOY@6L9d6f_n5nCF#~E`B_F
zBzn;4{ZL~Mo%l3zuSnLa0ANmfGY1fl?t`d9EZxs^U&yVcoao-{b!ciU
zuxv$SL(?)ciyIo9lx%psy-zF%Fm!S;A()ms59rF{E33fimOsD07uv|u&b=Z`+_J>h
za_skmqdp-zbhP*`)~-eHXOb0UYliAFS
zt%`Xq8p&hah9k4YRslDu4#Z$+Jq~0XkK5972nH_Fs@`)C6j{70q#Q^o!hPerjC3_J
z?*lZV!#V9IZHJn+xXTTRI=B_aMDOPQgG9jy8Ot{_
z-?=U74n!DrxCAZiQB(vHu)9j5ml!bsVhvD1@aqprxvseD2CYE<80UHDVe=HE{|Og%
zO7FGng_f=FnzOyuSKOf|^4EZYQV9@W3!l}LiBCBIc69=n&FLeI5hh@b+2L%oxwfQ+
z?Vz2nJat@uwy!+zh80|Y)iDvc1XK)`Gds4L-m{g_TVieW8b0U%4(c54MgNTAYhBNw
zS%^@$Zn6StreU=K6l+*~Z2bkQzzG-;p`PpWNb4vl4We}L$kZ(?Fnv!b9T&!|m4+Y@
zX`zuFuP(!u)~|b5V^toPIv6M-__Z3#yU1+#2RS18>yU&dc1O|UC7j=xD^>a+61TVt
z{(HA5h7`5Z*ldQ;1W$b#V|m_-{&L=|)@%QVvyVn0041q6C@!nnGItSr#ubU>wpvK`
zJjmZdN@7Qf!=%hKv0|&X54p?Qn7lh!0J(HqXzp;;Fr5k-aDn-tDLnwVXG18zR+;_!
zR(~Yd2T^Y0zall_wg&Q02sJA~oo1NHR@+wbyI5Y~rW#qS_BMgnBN#08?{?I;AXlk?
zNRH&UCRBm{89n{}R`&QLNT6u&6U2*wr`CB7Q@kzTwEiK>(Lz(RqnAu
z)eGxFOA!KdF$2){tF0}R1PD;Ou|OZxLr<;u>J(qh^*al3ZBWT-Io6VtwGvQ7Kr?Z5
zC}k_Y9&``WOYx{~(_?NiqDN8jh}fJZ99$1{eJ|C)<`y^Fo3fp+_=KejbV(Q@Bc|rs
zLOd@<`ffWb?ud~(??Eq!`rh`H!Q|)U7$h5Row3((PwAe#tS{*L+T