From 9acaa0bda8498346b772362dff0ba3f348f8fc5f Mon Sep 17 00:00:00 2001
From: Achazwl <323163497@qq.com>
Date: Thu, 1 Feb 2024 16:28:59 +0800
Subject: [PATCH] update README-en
---
README-en.md | 444 +++++++++++++++++++++-
README.md | 3 -
assets/en.code.case1.gif | Bin 0 -> 5164288 bytes
assets/en.creation.case1.png | Bin 0 -> 311672 bytes
assets/en.creation.case2.png | Bin 0 -> 570606 bytes
assets/en.instruction_following.case1.png | Bin 0 -> 361229 bytes
assets/en.math.case1.png | Bin 0 -> 166772 bytes
assets/en.math.case2.png | Bin 0 -> 70158 bytes
assets/en.special_char.case1.png | Bin 0 -> 285644 bytes
assets/en.special_char.case2.png | Bin 0 -> 89205 bytes
assets/en.translation.case1.png | Bin 0 -> 72571 bytes
11 files changed, 443 insertions(+), 4 deletions(-)
create mode 100644 assets/en.code.case1.gif
create mode 100644 assets/en.creation.case1.png
create mode 100644 assets/en.creation.case2.png
create mode 100644 assets/en.instruction_following.case1.png
create mode 100644 assets/en.math.case1.png
create mode 100644 assets/en.math.case2.png
create mode 100644 assets/en.special_char.case1.png
create mode 100644 assets/en.special_char.case2.png
create mode 100644 assets/en.translation.case1.png
diff --git a/README-en.md b/README-en.md
index 68b37bc..296c70c 100644
--- a/README-en.md
+++ b/README-en.md
@@ -1 +1,443 @@
-Update soon.
\ No newline at end of file
+
+
+ MiniCPM
+
+
+
+
+
+ 中文 | English
+
+
+
+
+Technical Report |
+Multi-modal Model OmniLMM |
+CPM-C 100B Model Trial |
+Welcome to our community
+
+
+MiniCPM is an End-Side LLM developed by ModelBest Inc. and TsinghuaNLP, with only 2.4B parameters excluding embeddings.
+
+- MiniCPM has very close performance compared with Mistral-7B on open-sourced general benchmarks with better ability on Chinese, Mathmetics and Coding after SFT. The overall performance exceeds Llama2-13B, MPT-30B, Falcon-40B, etc.
+- After DPO, MiniCPM outperforms Llama2-70B-Chat, Vicuna-33B, Mistral-7B-Instruct-v0.1, Zephyr-7B-alpha, etc. on MTBench.
+- MiniCPM-V, based on MiniCPM-2B, achieves the best overall performance among multimodel models of the same scale, surpassing existing multimodal large models built on Phi-2 and achieving performance comparable to or even better than 9.6B Qwen-VL-Chat on some tasks.
+- MiniCPM can be deployed and infer on smartphones, and the speed of streaming output is relatively higher than human verbal speed. MiniCPM-V is the first multi-modal models that can be deployed on smartphones.
+- The cost of developing based on MiniCPM is low. Parameter efficient finetuning can be conducted with a single 1080/2080 GPU and full parameter finetuning can be conducted with a 3090/4090 GPU.
+
+We release all model parameters for research and limited commercial use. We also release all the checkpoint during training and most public training data for research on model mechanism.
+
+- SFT and DPO version based on MiniCPM-2B and human preference: **MiniCPM-2B-SFT/DPO**
+- The multi-modal model **MiniCPM-V** based on MiniCPM-2B, which outperforms models with similar size, i.e., Phi-2
+- The INT4 quantized version **MiniCPM-2B-SFT/DPO-Int4** based on MiniCPM-2B-SFT/DPO
+- Mobile phone application based on MLC-LLM and LLMFarm. Both language model and multimodel model can conduct inference on smartphones.
+
+### Limitations
+
+- Due to limitations in model size, the model may experience hallucinatory issues. As DPO model tend to generate longer response, hallucinations are more likely to occur. We will also continue to iterate and improve the MiniCPM model.
+- To ensure the generality of the model for academic research purposes, we have not subject it to any identity-specific training. Meanwhile, as we use ShareGPT open-source corpus as part of the training data, the model may output identity-related information similar to the GPT series models.
+- Due to the limitation of model size, the output of the model is greatly influenced by prompts, which may result in inconsistent results from multiple attempts.
+- Due to limited model capacity, the model's knowledge recall may not be accurate. In the future, we will combine the RAG method to enhance the model's knowledge retention ability.
+
+## Quick Links
+
+- [Downloading](#1)
+- [Quick Start](#2)
+- [Benchmark](#3)
+- [Deployment on Mobile Phones](#4)
+- [Demo & API](#5)
+- [Fine-tuning Models](#6)
+- [LICENSE](#7)
+- [Citation](#8)
+- [Show Cases](#9)
+
+
+
+## Downloading
+
+* Language Model
+
+ | HuggingFace | ModelScope | WiseModel |
+ |-------------|------------|-----------|
+ |[MiniCPM-2B-sft-bf16](https://huggingface.co/openbmb/MiniCPM-2B-sft-bf16)|[MiniCPM-2B-sft-bf16](https://modelscope.cn/models/OpenBMB/miniCPM-bf16)|[MiniCPM-2B-sft-bf16](https://wisemodel.cn/models/OpenBMB/miniCPM-bf16)
+ |[MiniCPM-2B-sft-fp32](https://huggingface.co/openbmb/MiniCPM-2B-sft-fp32)|[MiniCPM-2B-sft-fp32](https://modelscope.cn/models/OpenBMB/MiniCPM-2B-sft-fp32)|[MiniCPM-2B-sft-fp32](https://wisemodel.cn/models/OpenBMB/miniCPM-dpo-fp32)
+ |[MiniCPM-2B-dpo-bf16](https://huggingface.co/openbmb/MiniCPM-2B-dpo-bf16)|[MiniCPM-2B-dpo-bf16](https://modelscope.cn/models/OpenBMB/MiniCPM-2B-dpo-bf16/summary)|[MiniCPM-2B-dpo-bf16](https://wisemodel.cn/models/OpenBMB/MiniCPM-2B-dpo-bf16)
+ |[MiniCPM-2B-dpo-fp16](https://huggingface.co/openbmb/MiniCPM-2B-dpo-fp16)|[MiniCPM-2B-dpo-fp16](https://modelscope.cn/models/OpenBMB/MiniCPM-2B-dpo-fp16/)|[MiniCPM-2B-dpo-fp16](https://wisemodel.cn/models/OpenBMB/MiniCPM-2B-dpo-fp16)
+ |[MiniCPM-2B-dpo-fp32](https://huggingface.co/openbmb/MiniCPM-2B-dpo-fp32)|[MiniCPM-2B-dpo-fp32](https://modelscope.cn/models/OpenBMB/MiniCPM-2B-dpo-fp32)|[MiniCPM-2B-dpo-fp32](https://wisemodel.cn/models/OpenBMB/miniCPM-dpo-fp32)
+
+* Multimodel Model
+
+ | HuggingFace | ModelScope | WiseModel |
+ |-------------|------------|-----------|
+ | [MiniCPM-V](https://huggingface.co/openbmb/MiniCPM-V) | [MiniCPM-V](https://modelscope.cn/models/OpenBMB/MiniCPM-V/) | [MiniCPM-V](https://wisemodel.cn/models/OpenBMB/MiniCPM-V) |
+ | [OmniLMM](https://huggingface.co/openbmb/OmniLMM-12B) | [OmniLMM](https://modelscope.cn/models/OpenBMB/OmniLMM-12B) | [OmniLMM](https://wisemodel.cn/models/OpenBMB/OmniLMM-12B) |
+
+
+
+
+
+## Quick Start
+
+#### vLLM
+
+* Install vLLM supporting MiniCPM.
+ - MiniCPM adopts the MUP program, which introduces some extra scaling operations to make the training process stable. And the MUP structure is a little different from the structure used by Llama and other LLMs.
+ - vLLM 0.2.2 is adapted to MiniCPM in the folder [inference](https://github.com/OpenBMB/MiniCPM/tree/main/inference). More vLLM versions will be supported in the future.
+
+```shell
+pip install inference/vllm
+```
+
+* Transfer Huggingface Transformers repo to vLLM-MiniCPM repo, where ``, `` are local paths.
+
+```shell
+python inference/convert_hf_to_vllmcpm.py --load --save
+```
+
+* Examples
+
+```shell
+cd inference/vllm/examples/infer_cpm
+python inference.py --model_path --prompt_path prompts/prompt_final.txt
+```
+
+#### Huggingface
+
+* Install `transformers>=4.36.0` and `accelerate`，run the following python code.
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import torch
+torch.manual_seed(0)
+
+path = 'openbmb/MiniCPM-2B-dpo-bf16'
+tokenizer = AutoTokenizer.from_pretrained(path)
+model = AutoModelForCausalLM.from_pretrained(path, torch_dtype=torch.bfloat16, device_map='cuda', trust_remote_code=True)
+
+responds, history = model.chat(tokenizer, "Which city is the capital of China?", temperature=0.8, top_p=0.8)
+print(responds)
+```
+
+* Examples
+
+```shell
+The capital city of China is Beijing. Beijing is not only the political center of China but also a cultural and economic hub. It is known for its rich history and numerous landmarks, such as the Great Wall, the Forbidden City, and the Temple of Heaven. The city is also home to the National Stadium, also known as the "Bird's Nest," and the National Aquatics Center, or "Water Cube." Beijing is a significant city in China, with a population of over 21 million people.
+```
+
+
+
+Update soon.
+
+
+
+
+## Evaluation results
+
+#### Evaluation Settings
+
+* Since it is difficult to standardize the evaluation of LLMs and there is no public prompt and test code for a large number of evaluations, we can only try our best to make it suitable for all types of models in terms of specific evaluation methods.
+* Overall, we use a unified prompt input for testing, and adjust the input according to the corresponding template for each model.
+* **The evaluation scripts and prompts have been open-sourced in our Github repository, and we welcome more developers to continuously improve our evaluation methods. **
+ * For the text evaluation part, we use our open source large model capability evaluation framework [UltraEval](https://github.com/OpenBMB/UltraEval). The following is the open source model reproduction process:
+ * install UltraEval
+ ```shell
+ git clone https://github.com/OpenBMB/UltraEval.git
+ cd UltraEval
+ pip install -e .
+ ```
+ * Download the relevant data and unzip it for processing
+ ```shell
+ wget -O RawData.zip "https://cloud.tsinghua.edu.cn/f/71b5232264ae4833a4d0/?dl=1"
+ unzip RawData.zip
+ python data_process.py
+ ```
+ * Execute evaluation scripts (templates are provided and can be customized)
+ ```shell
+ bash run_eval.sh
+ ```
+
+#### Deployment mode
+
+* Because MiniCPM uses the structure of Mup, which is slightly different from existing models in terms of specific computations, we have based the implementation of our model on the vllm=0.2.2 version.
+* **For non-MiniCPM models, we directly sampled the latest version of vllm=0.2.7 for inference.**
+
+#### Evaluation method
+
+* For the QA task (multiple-choice task), we chose to test in two ways:
+ * PPL: The options are used as a continuation of the question generation and the answer selection is based on the PPL of each option;
+ * The second is to generate the answer options directly.
+* For different models, the results obtained by these two approaches vary widely. the results on both MiniCPM models are closer, while models such as Mistral-7B-v0.1 perform better on PPL and worse on direct generation.
+* In the specific evaluation, we take the higher score of the two evaluation methods as the final result, so as to ensure the fairness of the comparison (* in the following table indicates the PPL).
+
+#### Text evaluation
+
+|Model|Average Score|Average Score in English|Average Score in Chinese|C-Eval|CMMLU|MMLU|HumanEval|MBPP|GSM8K|MATH|BBH|ARC-E|ARC-C|HellaSwag|
+|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|
+|Llama2-7B|35.40|36.21|31.765|32.42|31.11|44.32|12.2|27.17|13.57|1.8|33.23|75.25|42.75|75.62*|
+|Qwen-7B|49.46|47.19|59.655|58.96|60.35|57.65|17.07|42.15|41.24|5.34|37.75|83.42|64.76|75.32*|
+|Deepseek-7B|39.96|39.15|43.635|42.82|44.45|47.82|20.12|41.45|15.85|1.53|33.38|74.58*|42.15*|75.45*|
+|Mistral-7B|48.97|49.96|44.54|46.12|42.96|62.69|27.44|45.2|33.13|5.0|41.06|83.92|70.73|80.43*|
+|Llama2-13B|41.48|42.44|37.19|37.32|37.06|54.71|17.07|32.55|21.15|2.25|37.92|78.87*|58.19|79.23*|
+|MPT-30B|38.17|39.82|30.715|29.34|32.09|46.56|21.95|35.36|10.31|1.56|38.22|78.66*|46.08*|79.72*|
+|Falcon-40B|43.62|44.21|40.93|40.29|41.57|53.53|24.39|36.53|22.44|1.92|36.24|81.94*|57.68|83.26*|
+|MiniCPM-2B|52.33|52.6|51.1|51.13|51.07|53.46|50.00|47.31|53.83|10.24|36.87|85.44|68.00|68.25|
+
+|Model|Average Score|Average Score in English|Average Score in Chinese|C-Eval|CMMLU|MMLU|HumanEval|MBPP|GSM8K|MATH|BBH|ARC-E|ARC-C|HellaSwag|
+|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|
+|TinyLlama-1.1B|25.36|25.55|24.525|25.02|24.03|24.3|6.71|19.91|2.27|0.74|28.78|60.77*|28.15*|58.33*|Qwen-1.8B|34.72|31.87|47.565|49.81|45.32|43.37|7.93|17.8|19.26|2.42|29.07|63.97*|43.69|59.28*|
+|Qwen-1.8B|34.72|31.87|47.565|49.81|45.32|43.37|7.93|17.8|19.26|2.42|29.07|63.97*|43.69|59.28*|
+|Gemini Nano-3B|-|-|-|-|-|-|-|27.2(report)|22.8(report)|-|42.4(report)|-|-|-|
+|StableLM-Zephyr-3B|43.46|46.31|30.615|30.34|30.89|45.9|35.37|31.85|52.54|12.49|37.68|73.78|55.38|71.87*|
+|Phi-2-2B|48.84|54.41|23.775|23.37|24.18|52.66|47.56|55.04|57.16|3.5|43.39|86.11|71.25|73.07*|
+|MiniCPM-2B|52.33|52.6|51.1|51.13|51.07|53.46|50.00|47.31|53.83|10.24|36.87|85.44|68.00|68.25|
+
+|Model|Average Score|Average Score in English|Average Score in Chinese|C-Eval|CMMLU|MMLU|HumanEval|MBPP|GSM8K|MATH|BBH|ARC-E|ARC-C|HellaSwag|
+|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|
+|ChatGLM2-6B|37.98|35.17|50.63|52.05|49.21|45.77|10.37|9.38|22.74|5.96|32.6|74.45|56.82|58.48*|
+|Mistral-7B-Instruct-v0.1|44.36|45.89|37.51|38.06|36.96|53.56|29.27|39.34|28.73|3.48|39.52|81.61|63.99|73.47*|
+|Mistral-7B-Instruct-v0.2|50.91|52.83|42.235|42.55|41.92|60.51|36.59|48.95|40.49|4.95|39.81|86.28|73.38|84.55*|
+|Qwen-7B-Chat|44.93|42.05|57.9|58.57|57.23|56.03|15.85|40.52|42.23|8.3|37.34|64.44*|39.25*|74.52*|
+|Yi-6B-Chat|50.46|45.89|70.995|70.88|71.11|62.95|14.02|28.34|36.54|3.88|37.43|84.89|70.39|74.6*|
+|Baichuan2-7B-Chat|44.68|42.74|53.39|53.28|53.5|53|21.34|32.32|25.25|6.32|37.46|79.63|60.15|69.23*|
+|Deepseek-7B-chat|49.34|49.56|48.335|46.95|49.72|51.67|40.85|48.48|48.52|4.26|35.7|76.85|63.05|76.68*|
+|Llama2-7B-Chat|38.16|39.17|33.59|34.54|32.64|47.64|14.02|27.4|21.15|2.08|35.54|74.28|54.78|75.65*|
+|MiniCPM-2B|52.33|52.6|51.1|51.13|51.07|53.46|50.00|47.31|53.83|10.24|36.87|85.44|68.00|68.25|
+
+#### Multimodal evaluation
+
+
+
+
+
+
+ | Model |
+ Size |
+ MME |
+ MMB dev (en) |
+ MMB dev (zh) |
+ MMMU val |
+ CMMMU val |
+
+
+
+
+ | LLaVA-Phi |
+ 3B |
+ 1335 |
+ 59.8 |
+ - |
+ - |
+ - |
+
+
+ | MobileVLM |
+ 3B |
+ 1289 |
+ 59.6 |
+ - |
+ - |
+ - |
+
+
+ | Imp-v1 |
+ 3B |
+ 1434 |
+ 66.5 |
+ - |
+ - |
+ - |
+
+
+ | Qwen-VL-Chat |
+ 9.6B |
+ 1487 |
+ 60.6 |
+ 56.7 |
+ 35.9 |
+ 30.7 |
+
+
+ | CogVLM |
+ 17.4B |
+ 1438 |
+ 63.7 |
+ 53.8 |
+ 32.1 |
+ - |
+
+
+ | OmniLMM-3B |
+ 3B |
+ 1452 |
+ 67.3 |
+ 61.9 |
+ 34.7 |
+ 32.1 |
+
+
+
+
+
crtE#4T1N17$>6C+{aLudP2kk@`#~QARd$LUvdB
zWXt-|_s`@&Ir$9R$Ey=^g44&(*eNL3KfSQ&`&C9L!VZ==?XoEWD^*Zy*;Ixdk_7*2
zNw%i-WGD5vCiP~g^|v8*t)RC&rOUJN+q8P!!v2Pl?R5`3(?|vL06WW!4a)>O>rw>^
za2m$X79&kaLsZE|VGCDfV`t^y;IWOgDQ5_-;7ntuFSVi9s^l^{}{@lyeYp
zekI}*{ax{u7Ws`8DQarNYi7$6Y%BVmlSk?NyZ0H79H+P|hlCVTOsSIb>XhZK0`i38
zN7^|g5Ba0e?ne!$WGSa~1E*Ay?Uy$k-RLWi<<$&|g-_FF$IDdKCNywl7
zB(0L5<&rl-{-&yu=ehW0M9*hAe}Bp$aAYfZ#i5AiP_ndF^sG`0uTqY)S5B)^%CZ;s
zwpXpWP|b>=dQzo6!li*G|2D&=xmBfUoAOHm(m?LgK&skM3E(z#yfif8Hj+sFYJ*XTFPx!Q{9Ne
zr6W)cqq>AWbAU~7TcEl1mZ~in9V{puED1!RYP3!D7b2AxYDrbrj`lW%7uHc#wj5Ws
zjE=UfJT`KUc3L%dmNgC{#WD@tPX0Ac-%%RjHO^@~E>D!Tj<{X#9IOBby$&ABCp_*w
zSFTHr?jtqsh3xv?)t;^n#?P)jJ-Po7)czsj^?Fn5MRDy_d2al|&PAfuM~?Sv0k^Lf
zub|FPCy>49}W&yu~Mqj94acn-b
z&a$7+Wh`1S<5Cc%+mM0Zlu6=w*jATF@0nC`mB=cT9psT6)Q}@1l%yb(!`u|s`9A!S
zM-<_8KAB(vrAGmMW8tYlVZu#8MpHq-O`&a5QHfAdnBG6Krc#7ZndC^!j8OS@Q+c&~
zoQ6=Yicp146BO)Gfo-aw6t1L)RZ-t+WL&4sG^Q*x*2oCgO2TTzh3iyqYi+@GI^f{X
zk?H8BhA81im1mg=!c7IRCU9t$xMvkN_7>KAo85WaJP6Kt4~B2w!Z~3OfTx*FcuBQT
zF|;YV)w7MVxsBedkXodj)2oBut3(ph#x2qaCgt+Krf*h>XYf$HIs8?hPAlllAnKpn-`-n~l!K3&Zj0R(U~gbP=vM5}VCzkf+
zUeNt+O3VGF_*cyBrNT}$7HQ0>C{psUEJ--l;C8+X!dvCcf84GKQ;1I5~Q_+
zrivzuHM2}>lh3g>I^}@+z$siMWwwfi@J;O{(B7cVdbKA`IKjaP=61X_1xo+|T74ei
zQ-~z~GU*H=W!7#^bTsRYVvva!Npi9nOyIJbZccKxLS#q;QHUnHSWgtF6==hgLAEoc
zOFiDKV3h4bz2o*YJjD%&Z1u&b6ian?TPkmpT!9;b3RKD(QbVHM#5G4qKlOC16SuuH4!x_QVny*ZXyH!<8ETJ
zT*n>~TWr^H&kGBev0gggnqwITzez{g_hB2+eT>oO>itYvUh48>QK?^KSlx2J$g!tw
ze39p@EPXk|1^vK1_+?B)Ly2p`+mKkGinB#wb}qsx7MsNfbl}x%#Hm*$Kt=LsYO%(-`vlXxQjWPN<3`?
z%lf>!1?zVBg@I!u`;38eCrI1Kc4nBu$O#d_Z{)fQ(lK`5uH{EME^O8t`{tau}_
zj912fm+V)7fJf5qD*+^d0;|E)@mH&%Tu*e(0{QbB)?yDIOcb9qLuY`^O79qt&s<59ID()rTWSx
zs4jh#vd-rD{nFNDp@Z^v4gTPfEKAeOea|~m~-kyw!10ct7$+7uq8Iged>s=`F{`MFFz}{V>kyxxu
zHTTtnzrXQ#RvwThe#PVX>r3Ra@`$0YZQ@+cCA0bm9@lImq0MK%-_&XN{7*ZGO(Xpn
zUu6-_ZwX=alpef%J%L=_e(
zbz>(mlq_}K#ZqugP8pQ$oHnY+EWElq;_Bd>zVzQbUk>3lU!A_XSRnpLN67LS59Gig
zjasw~PwofWA{nJ*_1N;7
zTsBc_T#l=r#4-tCr~2{5Fn$`I8uQJ%E)J)KNppV71Dhg$8#z{D*Y8hYs
z@4rnxq5g1Sd{qWJyEQxO{7AZMO_`Csjo$72z{qq>MJ>C%TA0C1aeRFx3T1CwVlek4
z;S+zro?~>1;U-5y<1pyZ*+_2ENcvgRRP78cT3A@);BRb?IJC$}2_+IU3SU)P
zcSj^H^QPCsH`CCb$Q0yF&t=Qq67V+(28l=6(Ecd>*B|=S)n_tS9kdQ5mr9f(-E~*p
zT;6rA3)ceh=Dy{BlYD2JPOq&my{?0{>8B|ZUfX=w
zI>hksX~w?SjwrL|sBih9`|{k<(4=!o7Ormry_wMDtSu}U*(=pC+SgVA&9skzDgmN<
z#=T8*gLJM{KnUum)$L3#@nv0!5%v%yF1&=MyKF|JRYKD_1|HD6x09qF1@L1m#73$+
zIn$3L+%W45^jBsXhdc88H=AMU9=SSn+lk2LZFcDOsE_ek8Yg^IF5_klx_nj)6x~-N
zxy@~DS*rDM>vAM8JJCe?Dbg2eNM5+MGX*pYenKSGDv@R7=^vtj>6tdV9##Q
zKKBoPVKIpFtEm6st6n@jvPm4dNG<#Il!`MV#lm+esk4M+H9Q2;aBj
zXIE6ypNgjuM^B?a6U3aU<9}#ws
z-sk>(UOoC8{(`7;^jmUdW-*$mh?BSgNSp;FUj0jq9wpvXB{}{@a(YB^_xZ)`uNT)9
zFCIj{AUR$skNV*9i-7i+l!5CdqvK0Lv!|bqUvhDg@j8+Ta&gY0|2OsPIicz+3CCBG
z$FG#Q$Q2yP2UW=pqRA1XkK*}oJ<&UQ5>)xxxAKW!
z^plX|*9mH`lg3`BalMHneSO}_3@h$QtB8-YB(uXH`f8Cyie$uW{it=
z%8^!K-0zaw*e&s+7L`{X^>mk)w3Kgg(mz|At{PCh6&F=|yY
z;yaQGXOr$bGFewK*`F{44A43{F?&=oLr$2TUn1W8V<30>7*X{x`pY}oaf&GQPa)%<
zQvQ94QD@0=Vqy5lQhCBs`(@Zw?IU>XL1!fu%#p10Vq^;9s6TPV+2=uz-msWYaK
zTe#L)xUpJTbe+s-f~S*=S5HG^Fh&HxEjp|rI_508R4uy7&8?y#)*B;+m=Gf$Wu+RW
z#H$v6bSjSe^7Z=VSA5N{mnUEE{(U8V^_hrAf;v`$kmnnv%V!3cZ+E8>pLirWHNVli
z=szRdXy*PdxlQ<%2;J64THO3Oux{Pc^=spv78G=N9UibvYoMa-5*#?eJ)Uqc2Q
zD}$bp@p~mJ$IVBrDj5ArFoIh!jz=!(OfG>(K21|TrA9u>MZR)UzUWN8c9Jn^;$3Qt
zRJV&_e~n^-D)TUp(v*wRoC{gIv#bw~@~VsS+=MhrQ+EHAibbsKrKa*}tO{^ahww92Y;~+c5lEp7qz|oCsIAqntTm|THE29Hfb$ymlN%0$u3KW?b)0F>
z)fz3G8$k(`Rzb$QwZ^Ea-sKv-^IF}cI6%fZz=QWk4EYay?H_4VKQ7O8acgkGIupo-
z>LnSQAi1Ur$W&%Z^An#LCxw>aw5dX^ndG^dD1{lntGPs-sbsvl6rY8PwwcZayB4`H
z8f0l*XK8;CZ+>oxYxQ~5S?PV!q2)8V)iNQS286}`^qc+}Lt#x-XC2^bl5z1f$kiq*
zUMzd6eMHN!w$8AZ*S68s7S3k}bG2)aw`;wy>#nn#qOix~GNLv$g77(Px;pH}|J8lLcC30olTPVuLqaE9=SP^*EHkS4T=gCq6fi
zdN;_Wo8F_JWI9fy_3l&&?os?sarM8FD4kN=+!JOz@+dtsE>)ne?6a=+6MW!CH!yt0
zu2l!z?&j59?*&WnLdX(9&OA+&>&YookZwlgeGAvFj*qXAK-{hd)WbeO$Vz-5Ic0HkJ
z{p>q^DDRsJd|@a73`!_mVMQ5-btu8n84Tdr??P3dU{&l3{i^e2_WETDZ_4I8s|ST@
zCiH8>4XTwEs)bXk^AZalyeTpQ7o`c+0boT|26ZN|I>*~OP)eQKLR|@daqD$e0JtF}
zrLnxJAzHXG?zXonxkh85Q370(2dhyNZYnaURiTEJ2*a{cU~0@sk8Yqu;AZlaX0qF6
zQepVDC%hZhyhIJ3GpLw?!B=8-n`u=T<-zygups^7dyeXol+uQSh>_rnZ+)JRC!;6K*Hv3
zBTP47x531#$GW-4s;H515ymRgOP$(F@73#@+7aN@5n|XIwAhzKQ68mPzE9nX5^f=(
z9vHrD$-(rOE%sM74^(>%~RIf=qYcO?IX&w%;xG(oRnpP3Q1+vM!AwjmB5+#y8>1>vzk$w8(7;5(Nq8hF8{$
zuHbpEFh80kxL+X^TV?N@oLX9(flt%lFT8JAq;6T8TUujSUgZ@_2h!T6q|Ic|&c-aw
z%6V@n-LFg0ZH9<$=+JE#iEZi`&jDICQ{FX@-)}jluRuk&JzBPbEep(aOZ@k1A!6%X
z#%n>#J7ILYQR!<5#;i-C$h>sqIUHH!y;ml-U*f%AnZ6Ib-zPXkUM_6+w;a?jtqhBy
zrs%e*?so%Pc9F{{Kkvhs^xe&tLu{Jy?)mazWAS<=f3vmNkpt5TuxU&-a9$z~2x;OeX9j(-Qrsi{|0yx*mI9yFn
z$e}xc8}GwfE-a82*4PV>_@#%>1%&?6?AyUG<}{@B>W&%}A$}d_)3cC%riZ=8eT4u3
zXF1605}$Ku>v<7!^-k>S6ZW_ja6&1LVHQ7TXvMH1G5tQ6sn)v*>|HzZE(>##(0Y%;
zRtjS8&wa28LT3-wGQNTz5)$+1{(alw^Nf^_=j!w7w?MM@bcR2p-}Qvj@O!$5got~IQBvKW0|DpB$5s--JH_;~d-
ze`@AwWXMa|#IZ~l8`gVT%#$DJ6qxi!26$c+&(>PbWrIrU$`j
ztqyRy-Qoe6GPS_5NlVxrAD^M_FZ1pgdIf2_q%T_osr+EeMXEEa-YmJOK(CvsovD01
z$qN}$K8ZQ(&xw1j7xiD4ut9Q#~VWl!kolL9%ow<849h+xNgY7
zVm%m2Yz)3STyG5%SD}MnpB>D=I7t8y%*82kuvLY=^$vr%yoHivdOx_}U-5ZJC;Rr#
z6Q;hEKg1MIu~~me`2<$|UK-0X1X5Vy98)^&r`drFt^#X8@6u)8g)-&$t%Y*LKg|i{
zsuNfb`?4(iK9cp#8J&cfdhFQuZ^Uxu(b9Aw=Kth5`psjM#Rx27)s^Ke;is(|D2g+ekH!K
z$_k_l{h1xgG4L}dQtX*^?muOD>%4g5Q0x3;hXLz?bZ?ezyGm57UQu4wR~B=$;rv(EhqY{74^?jSSw-ao_5fl`%t#(0X`V6QCSKF=*
zF01V{g<;F;x}9#z>iaDf9O^d%!G{f7_iCm#;z`c#ir%pSb!9W;n+a`
z!oGX;+368I+pmr%tV1*%gyM^zJMm1Rn8`ZCP1@SJ<(%S`GQ1ox#E_oyHt+<8=-9_2hP%__<>t{}d|Ga>aK
zW~$vAUOS>k2Ctpa^nuh)n>vfu&6vbQ)Xf8{qAs~Z^AI&lEVTK|FY9e8Xwa6
z62iaYLD=H9;zNJvwid`LCa@MJzAUgFAVYk;5&!7pPE@ADh{txi53%P?c9gQ`Zhl52
z)^o2II^wxsZkBwVSt*I}MAac9!G}%g5iq(1OYC*jL89Vy3`#HKv|k(F=QuqU5{K3d
zDr9h+O=}=UE)cf#q76$b=)3b(XHK!JEgco_n{}v<_r*aDw&fNDq!+(C^&!FDZEho4
zF?S1Cam+RF_y0a*{tw^>o&Rve$7_=rNt2lKbYAw)i_}a!*5G!0o4UreqH`?Z>~lO{IiJ<420
zA{3Ve^Te_!eNq5;j>iwY%Ku5nG!RhJU6TWh80kM&&CRt1#V+lCp;?_Jc?FWKAb
zP$#PyyB6KvfQ5qDoABMc#qlDc`0llvYSGLxiUE|m*d`ZGr9<#c)
zKX>>=MW(P4BIw@ns=OkAGko>At$U~2z(+@4!nL^*zWVnUWwr|8YsgUDZn)BMh@PV9
z>lA@5_V~&$YnpY+{nx#?8FwUPXr20zM<4HWVRXn~TT=(9U*6%1aS{dxx`GB&^Q&S@
z!Z%rM1qTf4oD!-rIM38$=m%4E;$Zj|cZuMTU0sQBcc%{VpvQQ%nNb6vjre2XFDnQ%D`p&
z=I_CC%?_>=ai{}R$=kW1eAk*Nh`p)V!u(>qYb`tt*Zu{YJvyw;LPQ+8lz^8G9O_}+
zT~-q8XG>@0^>B0q+G`)yL8rV@tVC<0j*por;J9jgkZ$A0j9DqBxay$4M+cJOvePK{
zZdwP&P+OX{*XXOR#y{pdDUHL{OxMXWkthEOFdHmD!9jq-Ng~v6i=#kr$SU$Qc^k7O
zgcd~jIGkpZ91m}VoVG+op5?Mab`^n5RO{F!sR9Mao+h+u3K|JQGTf}uBLsS3hmQVW
z$bm^h!|dS5d9@z|WdnrGA=6zOfT;&|3AZy#xWIHqv!SeotrP~g|Mkdwag760gBNwx
z#+rT<1ZZAuMP2o9)1AbS)NjxOJqFCYk5X`{r=ZXEsIBo?5~t{nOqAyYxaIVMt-iUM
z_;z;C`=SO4N5LM0h9L(R&1g8xWOi9(Y;XEats!;$x44c9wl}T
z#m&pO>2KyUye}F_?)Pvj8t==-(c#w9(?_Cr2QsaFD>g1uc3$6|V0
zv6on%8_I9rFKMKiz3_vvzX=L|6ZZWkhzHA!{U-hBOD5|}{?qqQEojLKf@Jny4T3Dp
z_%dAiF5~}s-{8yG=gU;+%ev$HX$Oaz`hEE5higNy^tf$Y+Dt@y78H0!*ZY;oc#3@V
z|9a&oHtUxm<1f?aDf`o3a@JpN$6xVdfP!p*#M1zsD}OzK0PVs6EnR5PTBD8pW(e9KTur-
z|I-Z~+zF4q3Xh)+AC-+5ZwQX*3!f2)n3fHheI2o^8?n+CF(=E)zhfiULu=9(@IxSQ
zZzmG38`uAow>Op@L_T%3tdVjA3{^A$?eR^$Y
zQ3&q)7)>S@O>P~nKpB4O8cp3FO}iUSFK6XhD6)PPvG6gH+4|q7q6pTx2$s2jY;uuY
zdjCEP#&BB4e4&cu2#NWy3rbA@0z@er>pAsNN3EmzFUe^hJREfnVk)2YcKJ|YIcB6=P6C((c9ydnCG$zLG{*4Ps
zN+?Q-`jnK^pY)t6DRV9)P91rA-S{(>$F~vw4t>aC9OE!
zxiq9)+^R?1ieUQMZrYYy`nF#Beo;CqDIFb>e%zm~njbf*7T+3@ab1*g+n-_oC*_VH
z^8ridBl*nxIjq~wLlQn*vXkLRQU;R>us@
z;=-_j!U&-t>~3K+OHqt`QOj(+WpRFDa(;?X{?GX$i@hSdH^s>}MOM#>3v7yUzkkj^
zG4MgQKp{3!yg29Ue9kwU5^njDMxhcIO9@;)w_%{9^-XTuUP-+WE<-8lVkzwpEp5^-
z9eDFYTEA#QzHG{-OlLcHrnqcjplrzosI6aI5?Z`gT%5{MylzvT>sh|}rhIq4_@JqL
zg9UmRS}wGe|4yjTSHAGJ7e@guDI8)cw}4gxTJz$u!7sEur0c9nWd6U
zp^|*?$6uRDs*+0T!Afe^BD4?`#Zq-F4?QX_XC|y-q^|lPT=mfa%0^htVO#a-s`y=j
z-%x1vNK>hhZK)u2spw$!C`--oO||5HHA22d7W^M&DmPdoPhF!-SfeByI6hyg6IQEN
zf~_TrD>K-y1>gc(u=L8C>Q4(*mIig4tktZrI+p!9tCG5(3w6JQs~v6YLAIus$%#(~
z>+$#N2^Z>r-_{ceH~1+H075xx$HMp!kTi!#CZ+8Ya41b2x~Q|Ya4BA;DpW1
ztj#cmW)<7!CUA4BaC1Lv?SMgZA0d2Xu(^dbS$n^6CJa7T0*@N6UD$^s30qbdObn_)
zHLS1-VOT*}%l<;s?rjrlu%(o>b!)$6o3QnAU(eH4J|d+718a!7ZM`FGyH~)rJ+fehEO9dAWinG`#i
zF)b{1t!##^oFW|`!&&OUnUzBD_S@!hYWVPA=j48;fLG@rtP>}v2-$UsKJSuD?UGUK
z5)AK>!gSS{!)X<}wd}g*7T`Li-3CM5aSC-DG@Y!e9cFBuriMN2ianO4JwLsAJ{xw}
z9`yWZCcjx|!*6bLJ!ta??*$w7LcDOby7VC~{y!@*$t^Px_WK^U98hGKtz>Cm7^rryvH;`#6xCIx`$FL*wX
zZ#dvoIy{v+ycj-=Y#z>eKC((hMo>jh@vK<
zE~7dAv~k|Dxp>jJl!!UOw7FE;`3$AGH1_$-
zi21y`C}H?a8EPIn>@f0t0T-jzivE;~n2m(bHkHl7mu7uX3(a>6jYNx`MzcMlV}`@i
zgZ9(I@ad7I>7l#D3EHJm_N6JMrJ1rNN$2n1>=%$k%c~F*H}>UC`{iwRD+qj{n|%>A
zyo^>_JccYDvLk~;kw*|*PB!~@IPI^npf|cN5}n$DZcRrIkDzZjj{b*6
zk3QWii#)2h-y=lt5oH`b#~u+wk6w}-lL2C7Id;ybjDVTa+k@HL*ajn#Poen%proRzWxlsVzq+zd%#BcHuH-*rNqv&IQ&f9A2
zs;0_qeblXd`gu6;A_jdMjl60{-nNNj;((ZrsH-mMg&W{<7=b}#T#iLuPN4svWlC~4
z1-~Rsy>7G=
zOY#Ey82b#vJNiSy?=cyRVesdyPCr7d1dVL;s$>WSv*pHY#e3-}28qbr^*1=VlzJ-vqy>Qt<@
z|0bhkYOZ$#bO)ac7$6M==Ai
z%NUQd-6^OQa`_H?b-alv03h!nnCo-&Hnxe&(no;sMYI1?p1wBU7q4U`Mkq{=BtqW0
zeE#-~xu&A+85fSgd@i{7xt&-7^Q--Z98IMJ5w6fLMXJX$CiO}uwz8WFP|79!#<@mK
zhQ{rzQsynAzgi!4$i`S7ef0CHE=Cp26FH_VmoM@x1vO56ES1eweQb@3Cw(04m@oaG
z`)R7>*eA}^mAK}3CX{$)VylM*k)_wvDk7IS$MkESzs9IA!RExMI4M!hkQ74m
zOih{r#G@|HTAl{kQrl-D)EohdCt1~ay*Ug9ta
zSMQJdJJ6ynVhOYa-oA62173<=ErrmhyRJm@i+S{00*&3b(h}%yHcNEwZ??mFy**LY
z;>))?MMz_>qcTn(%wZ$e`0i)|=PaDA%J|$}?z>^VZ;pZDI7IJK-23*NU8pGvF#Ai;_>E+F#l#mxRl89w_DB38S|P}
zlBri{>p0T2Yn|?VR_5!8+Fu`6lI|9PhG$me8bzo`F1g%p0yah#L6{yQu&z1>kKI}v
ztUng7U0cYDYna0HXL@i8@Bnsu0N4OKH?3XYP;2kt0W&~8sju&}Phg4amak&cX&kh0
za8A25+>zGt*HJ2VDFYiFSn9wQHXYpHVHpykRXb9po(P!nRf|r`
zE-xo!ZgBRtX{P0P3GKCDfW&T5wq0!kafs;;?7N0m`)5DhX=&KAMRjumBCdP3kO$_(xT~$PWSE)%3Kk8hU=8;v)#gxm62dC
zv$wdX79{=>E%isU)LW-t!O|%v4P*XMrm=VurL3LjWx-~pH)ycslmMsUseY!}W+qnQ
zO_chBoAqjl(AYU;3bV!2i}h%8V&YET_SzPzrW$eQ;Lbp$TJcXbB7QG(XW`LG4TL!*OjDKiwh1%U&)u6=t+`#x%r&n+w5juw~5P_({3$c
zLujAgh2@9%4b24JN09f6$R{f9os?p~f-Vi#9z+Rr(;NQ^1{kiASvB;qrfYoAnVEgh
z?9s2}?G#%;6M%(lv*2spJ0OMmV;9h)
zv>hQCT*-FCrH~r3s}5`$FS&Oqz7!lgbUiPIE!#kYXb;S+Ky$4Qu9Z1y2Ub?#`Gpo&
zoZWortOH&`I=D5iryaVHP?cIU+NR(eANetRtrGdT&sSiqf~~yPX|e8oGU>;Cb)J(k
zk>`UdqNh4D%@e|a>(U_a)A&n~Z6y`Y(J14yJQ>j~y;j^UPR8dsz{}~?r|2lF)b*zN
zt3$5)+lWi!E80uQQHZ$LGCknB-%9Mr8SND%(|U#A_dYElx!YBszn(?jpTkukN45Yq
z1}L^_W7x<0tN?JgA5^|ZkV|^e3%EZmz(LaD3aPN;vkEuJ!5z})f0*X~Ew;|Zh(5xI
zKT!PoP^9gFohY758{YS}r*>cO!EMhJnHSy{!Uu(0*1wW0etj`1OnT7v;zK(rrj3kE
z;uVhsg`otcro?MS2`anx*SYOf{*Iwv
zY9OcWVeHqZ;T?pr9n3D@p47Clbak+%er5OX_=M?T+xW(n+rj=plJmJF_p6$DvO+$V
zLNuB9K4I%YO6Orrr;xwop+SpCxTJ`eq|l_Kcyp)tMyJH0q~!DOUu!zQA4p1l_%7kq
zC52`C{@tZ(^SwC0;G5#mccsPe$_HK3Z@;mgepjQBQfHIW*l5R?n3R>x{T_YR_xD4;EL(r1qFkh*++Pv71RA+GJGn&j{-{#9*xde<)NmUt
zmyG9uTn0ui^SOMnlgt&Fe2$2GuA+QiY@VinUx15jA&+czY&t=e<)*br~rJ`TW79N>!r{buF&K%*yOJO8&YUq9BiB%Y}puu
zVFrIQ^ct@A=ZPqGDJpgw4&^$^^mr-ug)8=_D&|>tR4&R_O%7CH1_l?0DuxuRpDT?#
z9v86+c4tY1-(MJY0w|s{U?h;Irb%fhx(J>We(Z-sfsB+11EI)lO;@uMO2O
zH3*EC8fAnUHdPJB9^4%u-gJ*qbLHr$LK1DZ;M#9dBHMS5rcC(j5C}^j)ngLH8(O
z*`!>prrfZmJX}+GN>g!3Qw61|N;{>(uBE2+fk=sfxK=~kUP}iu6-%wAo2I2-re#1Y
zLe8iu1)3DI&@@6#eaD+Nvez`do07uiC$!pTn^WfG+E$Ez86uPx-)pe2YuM0EuoF$#
zENR=FkK5Hw12?s~@pORfI`*h(XIdRcxHd>i$4N=Y^{#Nwu-TwY2MpKo8ZH+Y*6~K^
z_}tBGc1)O_YyFYZwiKPkai;!S(|;^<1B|8vAlm)`x}o;EA!)3#-QlWBy6U1@QJcCN
zcym!G-Dq~bf1-1LrRJiQ^x}>567Ui=wSu+qbd$^UQs8>}*ne87OM2-jy$r^D!`fL(
z+W8Rj`D}=8PK0i@)_e{~w*WpHDy3fpna|6c{b|uD3z`8D>EKcu*SmT5P5lb;g$mvo
zH_?SM-i4X~oho~S+VlA;i-mgHg~kYjN8cts-0H!X3|deN1#>#BM279Oh8^8hz7hJx
z_WFfN`dvngMIfALs^4$FSZrZ92*D|)hTY^%;Sjwz-nn=@BLvYBg579bbZJb+O;IM4r5Q_(Bx?8Sg1i(ZA=R|;03&0sI@|+!c
zNe;NNM=qx5@Ddr{EdlOPfVuw}Vu^k{p!@NVE~2y6c+=h(Px{Ap-11|~6})m|f_&p=
zuDJZj___BA+kxRMXlWK=^y2;p>BWzihd;=OS6@o6lF_Y_E1OU_)Xc%ddDws8EEVc$
zz%tx~cG-lEVr8{#<$22wq9?Tw^9)>*_Zyj~PA#A+O-3
z*Tbe)-KLz&I5xwS>(KNw$2wOz^12qut!?(jc%8@FjDsU#9%Uj_ZYJ!l^}iM~(Pc9+
z@2N+TYwzMsnQ=ZUh51+QHHkpeZ_4J9SmO;zj*YL8=ET}2^4=y2t|p2OCQ9WSid!2>
zEgLF_=BgLw>RW@a(^oMlGc98aZ3+a9gN1ISg`VXEy|%f5!-hem`FC$~>1hkuWeXWx
zg7stzCy$tw^RYN|THBjhO0Nr8npqLATLxO%1e#e#T5_jb+KyP-xvpC+TLN3QZ0@&2
zyO4TfRxZj`c|&F(2P-%4Z8waWNsoo+#ilXSmS>L@7>A*zZ$mDuAjCUnhgO>H`Y3V>
ze`1ROj!l2dodBG0kK;>mh$+q-ZTT7GP53X##QpGR)csEZDT~M_*8e!HW0ZUKN49;R
z?8I?wfgP-Isz(yjPGbJ957yN>g<>Z`*^+3{*n!T<@rji^({?7)UIxXU1KnP>@m>z!
z-Y*B6eB7``+orJG2FAM`Gh$P=yw_)DReo<%L2R2GW0h*To5Z)D8n~NMZe8QBU*&CE
z7r0-WZkxt8E0?|#-eU`|+iBj~hxhEX9@@5D*mls_by6JE3U3Nh*!DQs_1;(Mc-!@-
z+YN9u#^&4BmfO}x+K%)bRL9#6KS2%Ap%8}$qkO1WtJda*Te%MQnRPbB={7Ucdo%H<
z`AGZ4bo+u4)WS5XaLfL`tAs94C2e~HVh$V1hv*ZVO$Uc-q@r|Wp9j~Py^P$#wwAbY_Hd+Q0uWz+4*0xiAs
z7f$EGV<65k`nY-coVgdcd1bgk>MuR5aF*#An4VAx>iUP%-498mD()Tta1R6!X&_ylFI_{T
z+(S9t!sb9
z0R_D`WivNLt)5U4_X_&kQs}VnwnvQ$SY5@v7WjXx-E~luVgELK1r;gjE(IkO1e7j8
zr4*#QrMqJZ1?leYX6ahkC6sQE?xj0-=>_(=yyyMRJ@d}{&OFb|^B*%ioUqq5%wC-5
z@%Znb4I)OFx*?=(i3)!YQ02_QuwqG_)K8kj!WM)Yuv%GGKyz3@d!Yr
z6sX#&yGj<+0_3i0#%F%i2M)i3AN#B@_^xu@M+<;P6n!@p?HUbyxBc#WqCnGFce^Zi
zQw(<#z4v>g_meZe$LM3MM(x&qzc*F+wN?9$o$s|)rO9workA?ye7WJ@4pW+$Lv6qw9T>KD=Ui^}LSam^
zUme2#xMGrAfGJOnI^F4zfw}_R>%9rDxO#L25t}2CqB{4lh2B|?m+IyxK!Sj0O((9pS@2#VMjhyiRwgGYx{X^n@;s
z=3AZ7oMIx^dn?1y(gym1t``eao$h}6zFs%`sqGL32DJD6`5kQI09urCE?NHe!{C1q%}u^s!Bol#S)N@Zbo&mf;tc0;t>R4gwbbG)5Mr)4+Yf`fBqxAayCgT5
zF|8yo9L+Odk{>NeU0M*YtX*1|Y@AkFlHDO=H_Q
zVZ2ab2KO(jfUG*<+BK|aSsyj*g$>#@Z`jdrLH9d^?P2Fo4*Rx?wsg)m^bO71Rt&hj
zLpvsuu0sbF2M{nN*vCETAfowl+I6da;Lt;8vdCS>F!O}R?t7fgatd?uYj{65QlydOdpUG|WAbQ8z5K4lt&om_Bu5A_RBNe?O39*8i1b0vOY;Qa<&QUli|Lrj*bs
znGI9gCV*qA7vKX-O-0|i&KbpLHqKe)1CFU}r%&Ue&FG!mlH)`sWXWqCa7=wLeIYB}
z1o!T%KS{EhR>PRun${zwe492R6z{KQ;dkYGsr^pN&D+@l^w6FBczx(@aefwb4_vnl
z-LJv|UzgQ(>cb8p>shcP7;+hQ+>Xi6a?(j+;C=9dsl5h4FJI1$Bop^LZJXd9iG(YyfoVI6ZyF*n8oZ@f}#Ci#cJboZwpvCWCv9Ac@nc!H(9kT9i3i*PUr#
zP(0-sRN@2-Q?rdgXTN(`BXo*!c7ubBI4m*}aXj5yn!swZq^UIgwsNGRNjCVO0bgUn6t;Pbp
z?4nQ~S*2+9#)2i^DHe&VoF61d?>hAftR7KX~}`H$znxMP);WmndUdRNG*$M#0P
zY&ug)iK-b7B
z-KVVe`)}0S#}ljkA0OyrFU-{4lhsh%Fw%8;{Oy(GtBuPku~
zK0>qqgd(X+kQSXNk`1S@!KS$z#km_w*fQb&o}IM=HKv
zC!X=><{ROTa6d4i!))&TkaZeupE?;)*;pDxS8j~PUD)U>dS5M{g@{k3HDxiN8VDV(
zO@thnvS|RGDLFE^M%V03faj2req9>S49{77J^Ve2H*;9m{M~}!sK+vIMr@xZ4oG0k
zpZ;AguI-K#6LlR)kqMaia7S87Xgn_KT_J~$@DzJP-$#0$sP{{j3J{M8N&1Eo`8!MH
z?D@%Dh4)~M!#$ezwBE9y_Z6l~d)gqO>E@pIidE@K53y3(f73Tshl?7TV4-FQ6!_{w
zTW!qwY36{D{Ce=6jkPptetv`xQrc>3?@qI@5@i0<)6GymUwCOxpC7hxc;wDx^|#EQ
zzjfdC7_=@7zj5OaamBaN&!2}s#JVcCeS<#nKb~KC)ZWlR0|YnG3#-6+$nRJBj$usR
zYt-#PhSPVW@LtH&JF6}WCE~<4NVVzCCpakU>%?ERxcNrIvme+bPiq(1d|M?rV%~<#
z02onzbn~dI-B}Ka&yMuA=Xl`JSrKx3G-8>53Ty|&qDA+$2?|>Ffx_be(F0=!;kiLO
zw^1YN@08|O^8_OoQSF)QPNP?g>k5~!1<_+WII82?&anm6dg#M&3qM!z_(QSW-1fw?
zsr}7Sudo;*GW)jag`(%6hCU*-y=9C0_
z%@#ehwnr`CmhGo^iXMZGQE)V-_*I7nJ0m86`}z#s^$+pe)6{J5Z7%WCG5(e_aP-wa
zKm2B{_a4~*_c@YgxXr%5T9fAE&EfUg-7)a{8(VpQQ4GIBM2p|eRh?Zgpm&dfC#BVF
zzmntZ`_*hT(R1{j!2^ty2h@fSaSYLKRxk%bCWjYJlnfthuRJz3e8QM=q}DZKxq{_y
z_%v+gY23>5LOq0y0gi7rP7njGa`qJ)92aVcH@=IRqCg_#HX;
z#QN9VWP2}i_g=l&=U&|tTqPFEC6+THku(DS((1VDhazWM5%%V#J6uW_4p^^;%<#
z%4171W^ZI@EQ%hs9lOa!3o
z0#R{9_r#Rr>y$w+sOC*X*Ve_BO~m)|#kWi(5bOUOQ;ZD>>hB*IHzesdMAKdj3HvTv$s`cBA)Ce|T=re4c0)d|K(=E;
zuIIaaTmF~njV~Bf%#2iSwv2C=jk$1MeY?ryykAqIHdRJ2Dm{Cp4EUwf#!6y^%AZYD
zluf_!n5s*@QnxEqRWDT6FVq;%etWbmowli!XR2LRnCCbx>A9laVX8Z@sXJbX%_q~o
zRPeQ^Kw;~Z!I7x~2J=^(BKarGzz2^3IkPeSmH{WT0$-8-UXlo-q7?f}Q|1j*p6{ks
z%x2nVW=>mXUS<|{$MI((xGl<2jve>;07}4JLgl>DZv4A?Zedf)5N6fv(Sr%FVDAnQz
z&dv{Fi^9b%9dZkQ_8os7i-39A_xk2epG%w-N?dWw0~NLdtt?#4SOLW}*lx!psN`1^
zYhV)VmCTk~nweGYmR0Rec%?-Iv?QX@A`;+C*GeMCcOv&Jq7XaLN38XtF9&an1E_an
z87<@1wrtrgFW9oE%W|HSW$zJc
zVSGxEv?`F>E0|yXsk~QcY*l2%jxEWW_iNa#t4D@oqGV(3tm@_V
z>y^P%+?EZ-){R#CjcH2>{$=G^>`k?0O>OMWJ$uC??9d=EG>QY31cqhpLpt`qR#-AW
z`s0>Z(
zWZh$R(BpIvpk>wTZ_^ic(3hoNJh0zi#4%6_9%$qkYy%JWSPu*w42;+eP9F>|+YAw#
zRV?qL+J{rdw$t|$4OX@@
zUWYUO6;htoBWc!iBl~k5<#R=E=HuQhj2|q_zg}D`U)-}<1m=sUto!c1cOfeN+*EYR
zz3#fV?Z)9;xu{rucGSggw@Q4}DP*@I1{A??ZYWo7a2|Q;+ss-WZ2=wEIHSE@b~|B5
zJ8^b3!%pQSrjG(SU@Npp-x!wta9aL?1wU5t@?Ju%$W~}Tly^b&a?JFZ}56d{O+bWO7
zj}DWnZnBP#inwmpDp5VhsDa~~M#E(u`Zk5iO`fAob-T3-`v=mT4+!3_QJ>rspP=ch
z(Fps8OeYw))fnO@k9glkf493bc6efa@&pTK$LRzsz~O25+w_de>$JD1thcxG_IHh3
zchHmDjuV{rYP?Df
zdWMG+j(C0S#EGos#H^DL;|6_nWTruKGx6H%JM!>4^9m#5-MBemj_fLR@7$d@Opxz%
zkgOWcdFybD*l%Ui(r065bs7rDB6xz2xf=scctssp*yx{Jd8867850d=N`dCvWr_p>{%
zw0XTW!}(_>S0%16y>(f8AhFu$0h-4zl*B8PDhw@|qT|&0@RPmEowScv$HJujYIZ6f9g$
z39Uihz+j*jUQjI)s)h^H3V@~ysOsY9MGJcaO+v+3Ev2|3Wi;Mk29Z*Z+cFA~4&mF5
zcA-v^TXYwJcek@lc@?O~`nJd2!BU0N-^m+r3Hl~vtANw|?AunbNPB)u+cs#h^LAhY
zHgt@_7lJ|7y&IiTBir7PeeY4^?I>>RD2dNlfLRtMbw9`5IPcv@2g3;|pGo=F%zD^h
zvB==W?V$PH)Tqcb$Y&G(&
z70kaH?suIoZoqPXax8XI)qZo?1_W49xaeDzw%Zx;YpeU+eZM^?R_!Zx6K<@Yt0Ii
z*04A78bB2G`mLbri*po(gW;b5YzjVaN=M`Fa5NRa+byNj
z_x^Y`jYe-OXS3mSLHCPWDi@2t`LfA;KGd#OQ(!HK+a0x=&0L-3U=aw$dDPV4d3kY1
zekigwTXkK5_Lu5b6-AfFlOnXTj
z3u60TL~hi`yo`RtBg>pjs@%_-nLs?iR#8qYPo6%6kmd}SM997^B}F6Ucv=_7hu*9$}EYam}V?6!4Vv>F^v
zT5xGq^Ega|=z9YjlLl_6!b`aKb)@?W2zNT!(2wMahtW?mzAK|&
zFU&lQgV~ZO*8|>1d8~&EH&Sl=_+;m?5mr2o-?^`Sl8HcIrD$x^)`hg^1V>3D4X*kDx0v&W7a)?p!$WyVG1HiPUf}8fVjdD&|d!
z@7xwlXXq}MoY6~{DQi^k4*IOIpbwL|gq|`V8K{Xuq*RTGl0CYZnPWqhfNAvH
zke44!hg~VW^=`$~Ya<--BIt2V`ef8U$ETi#8>=!=jn4MGt38d>KVYQZ-sutQiH$Nz
z$frj(^@!foL_4&sKfkuW{~z;uSDc^!Ll~
z5ho#OJ9Nsy@^FpwUL0RFkmz+ENQGud_^?+}xG=(0M{$ZI?LZ59!uh
zU`E@7o+Pzjm3gW%>61C5V~}D5Gg-0VCu?=0^r<1UXQll>0TfBb^5Pajr>ufw5=q*&
zYO&y4>WG`1Gk}b;h`{HDy?lAI&N7NcFGEK`QFYl5L>KomWWQa|k@_C0uu3k?j0G|$
z3RgJGBc8JO#6ABv(kvmL;oL$-7DqWAm{Vp{vy@euA5SbwDCBar{9?l1
zo?7peDmb;P;5o07eNS5Qfy|1>Gqg6B9*Cl;)3zk@90y&<^W@}c|q?OcoG_Zxns~zquHKDQEmRUN@o68vC
zpy8dn#R%V>vWu4dG+6q;h$J`#-Nh#67^>7_Oa`Isd7X3`Zhl}KIe+%D_wjL*=b{M%
znzHW$Igd(lc=`(ls(x8@L|p3PcNXdDYCN$M|Kiu1k91Po)TR-MkPK5!zqvu9Bd(O*
z*IN(Vn};mLcvEL=%pSo~2W))n(zX}PUJy@L2Yw~B1ubmDZv>mV?vOc{__|^wH)DZM
zT=J+Qb)FT+FCY`$T@||-j~%Hme5_}w$nQ0gUG%O*_^Z{IWUtB6XAPzABdx-y
zlA&xMM7gD&wGNzmrg{3kdDy|A)*_k10>Vd&H%OGgub|Fg6Y{g#G@9s$RW`5OuN@&O|INUr!BUxkG*DQm#+cD
zbhqNz7l~R%lY9KeK05ZtTwK8h5L2wmlOW#P6#@hW|uc}Q&|@^
zRdh#o=BR~D`pXJ<>w!zJ@cgCzVy>ph=1+>-6>I~KPKITk=M>ayRB(@8ae727g?Ft^
z@L3@P9rE*xH+uX{Hgd@M9{3<`+4&&uHJa|bQ>Y=jtL1ky1#3Gmjcy$`Lfb9$(yTp;
zzB+P+qZZNnYb{u9CjrN(H3IlmmxeDQHQIZdf#GJ*uMIGayyuj3U;aqGy=aE}9Q!fc
zj;XbHb{O~~vl;GoHT;w{I8pO3hWq+xzwK>?o6A84^x0_p-G21V9m)WW5re)zKEJ7Y
zfbDScsMR%7yyJ(3>7
zR5^LX?eO@+x5rA=k3YpcHi~&-KK4YfTGw#6WI7rPG=}9z_%wjfs;5fj*Vxl&LhSf&
z*vW4NS<%O-Prf}Qc#A_qhyx~kQ1K0?CI+V}2Dh#n7cz$1Nr=~ba$i)9*KmT@{uaMG
zMk-?@=ikVZ;@RSg%rmNciWqmX@
z0}(Nk5;03GF-gEv4k8j>B@+Hv65&&>d4h*CF?je6q*Ae@pNPm5mB_xvk_|hM>D7?^
zJs~sYCZBR3x2++ctG=hKu72`(Plk}v;N*!ZA(nMDWpK@-$XLo`ZmL*rN|S1;XeH{@
zQ>sL6>QW`j;!|o!4Rt0rO$`w>Oo^uZbj;=LvBPNSk0IjfAUcd?wjO7;CuRxQ
zqjQd90gbcx5wiy4kg_1i>1tn4jI+iQvn4AtGSsrBBG}RqtcA*KrNrz^JnVJE9FW=;
z5AIRgnDC}knrZ6U2;O3J
zE$@d3UitWUihyWZ#|{*xe64%08m1`ZpCkg-034WmY
zBs^Xq+F2l8MId=XpoWo|*O-}GMW{Gl2s|OAXu?vZA`FQahD`|L?r>C%zc+Cf=_cVD
zM!uhj7a5J`HLVjJM2gPEzn=kI(|EDt39-%|j?5#9W1hFzByS$Nyt(6%Ku<_uAti80
z|3#*0TqGV(e7KAMz@RGe>YXHZf+Wd1DV9k|wn^H^17RuEkJ1Srn~`)C;EyV*pEMFa
zc`Ea6^NO$6iJ6j$TfX~jH2K+?RN8$~)RR=qM^y}@D&yxO6Hx!jg5-62ykIQaSvFBc
zHakH!cTzS#L9TdGE}c}q@SQx^MZQW^UKk;mhFUmr+&hI+Cw5311f`zw$e>
zMye{9lgjMY%N!=iAYBx!>lLog6z)_NPhAx4&lIuWf6KHGr>Pd_u9u=oRAL~L2A#V1y6tMZf2MQ?$@Cs_eqqT|bDq+Jlj*O&hfb>M@1v*mkI4*tUMu9D
zYegp-0E%h4t5)WFLu_?Jpg-nL%_u+72-RSOs~$?MHcEe@`n5q-WeO-JGvOyQ5pFaV
zy)YJ6{|*qQ{7I@`)z##aOr*(8dEEda^!qVF%IA%N^|YBixp{b~x^t2_XxiNGBGGNi
zFxbs7oZKk%!Xnbmh%U+U1s{-%vdrYOBDpXsZM1BU6Yp2E?oG5Fbp?WR))Vh-+SP6T
zOxpm>+x_acy-Bu@9rXzE^nR+_Z7137Us&u+*&)^KFO%%SW(LLNmJc;7!F&#YU0T)X
zKy>L)IPH+z==gZXk&NF7q+s0cq@>bl`hm|`>%x?u!i*aYahBpYlcsQyoiY8Y;rt2W
zqH*b>#qX+=>}+)DZ0leqRcUWe;qI*Ao@!$5KI0DbNC&v{{c>}xNOF8d;S_)Am`vfB
zdg+lpIYfV&xe}B
z?3ehR+*rJBR`xRDwYs_MsI10kc1`H2s-oe}rCzP@w_H)FU*0HU!X369vL{*+A=(KEsz`
za}?nV$>9KG+CUjmrx^hgi0Jo-XupauSo6VR^!1sIgi}W8k@>BsMD5Q;IbQf>HihP<
zgcf^5$EHM|&qiZ4N8@V6;0wl(P{lk>@_!R&tR)bjL>Z?i5T_;(^a`*{o8wrhg7^jF
z1yg^CYbA)LCa}#V2wx|d&c1a&iPD)%G)VpEN||Jynq)mE^fM)fKr5L{D~!l97UY@i
zcO4pd9sM&kIe0E5GLZwn6#N&0-lob9Er_)ysd*=u$laWv
zM3pOb{SPw57Rtqk=6;~cW0=pSO3Ql%&10F*~Rb
z7(%lAlKJ~W`N!1xNUsDYuPQERl^}qc&R3p$*Y*tRY
zmSdro3pA~MQ!(XLynv>z6b4lYf{LVtD@1kb<%R3N>NL3S6lc#jWKuV#Yd7ZJG?>F0
zErlB$7l0kFMiA;o;Rz47PE&Atljm!2I8AfBPP5+uSc9fMJH4KNp*bH_SBk1H6mF=a
zfhvJuH6WRdH0?JjH3aF^6u>;cSIY#dF1=-6xNR5JwhB|F1vR4<
z+R>;IR+`^f8NYFFQ{KU#%?r?Coj){-FbZAR3jj5}{X>`0!6o9QlMdnch6w9I#1|o=
zB3&OcI%Rdc(tg{=dp*VRx(iAx$r
zFZ!2=uA4^hFJ0ze2EB2nOpsI*>ok0rER?yBm>di)G&PKP+ChM7LT+>R?
z&FAaQ{|p!|UYf6>TTqc2F}WL60UXoTg+G8}N;lo!y4amL)$cP6itT^l^B2Cfw2t-}
zU-enu%pBjnTRx3Zcg@o!*2-yL^qYP<}`d~4YD>(#Wp95
zq4fc)?pfQQHdj1feJ}l;VBZz@_YB;=>r;AwRIns0wwp>1aHi|&_q)Yn3nf*xdBDr#*#n{ZFCZ7(xx
z4=uKLslRt8c2bdb{1AP@kF|dnF!nd=V3Zy)A%@#WjnLNY!e<0Iy4fweZ!-avtg-P~>IsC#Ji0B$zI%i+H!!JP$SAGUp0ohl<
z@T+i!>u7`P`0VRs_;otNO}4>Je)df<{07W`sxm;;WuqW)6pZ1v-Qczpy=9%XWlq1P
zBX(yiwgrfYf8nB^k0ehepLGHWRz%I}SOE*Q2bgIJe|eszg0iYj>9k87M|`K#lV;P!hqg%-_I@t$ZpsjSz{PV0lQ+$z&H
zf_2-&*}?(zZ=mnCCvuhI)oq0u_NS_JI{tw8kHG4QP|}n92UU9)h2R@T>TrbVK3H-IzL|M%;strL!nMj
zCuZ!$fruaa9BbtV26-R+=R-!B8~`(=%#`E|$@xWYwUZmdT(*-J!MC@QAM=@gw;(~?YPT@OtZcU^!)+hx-l`1irT-*!WDJXX5oj8Q%)dT$cisVW#ej`a3y4`DEtVzDOsNraNM(y
z(}FjRd)#t0FMr&6vo(0!j(T!X_4|$-?*#q(Db1IYKR7HyC!M&`2i0Bp8USWWsq^Ku
zo7#Nnw3jCMpr(&58PCborl@>tfE_Y~80755Lk{uGd_fNLZATzSgf54WqheV2XJe8S
zU(f!2W{Er-m*XEkn-DSHV*akuE9+{m*6G519(&L-`UTk|luU#kF
zxM19C+c}FWZ3UKoVP-Y01Trmqz=NfZn%1JR=$kj=
zc@?g=(%gNo*RnO9d2Qu7>qEB-iaD?MQtHH@`+2=~(7lS8d+1?pKCm+j+m8|k-mx$0
zq&eG8IxTz5CK
zb>eLcdyO3EW+V_Zy}Q^R<-AAj8~C9g+-5(x=9fSdo&0$6#qa@9lf+~4>Ys$(Igh9W
zK42-I_>&JAKH~P^iH|K0cz&DnRQj_dLDJh_EP!XK5h6)mRvq}(I~PwyMv7wkZ4f`;
znYxEa(eG6UOWfuXIs+S~