From b26a4734ca7616fa04f6f76d1c9d4e2363b50c90 Mon Sep 17 00:00:00 2001
From: Robert Muchsel <16564465+rotx-maxim@users.noreply.github.com>
Date: Fri, 5 Jul 2024 08:54:37 -0500
Subject: [PATCH] Remove librosa, merge branch 'develop' (#325)
* Update KWS, MSNoise, Signalmixer Data Loaders & Evaluation Notebook, Add New Scripts for Mixed Signals (#299)
* Remove librosa
* Update README
* Remove .keys() from kws20.py
* Remove noise_type argument from kws20 get_datasets()
---------
Co-authored-by: Merve Eyuboglu <141146522+EyubogluMerve@users.noreply.github.com>
---
KNOWN_ISSUES.txt | 10 +-
README.md | 44 +-
README.pdf | Bin 6551023 -> 6550371 bytes
datasets/kws20.py | 502 +++++++++++++++-----
datasets/mixedkws.py | 493 -------------------
datasets/msnoise.py | 179 +++----
datasets/signalmixer.py | 154 +++---
datasets/speechcom.py | 483 -------------------
notebooks/KWS_Noise_Evaluation.ipynb | 242 ++++------
requirements.txt | 6 +-
scripts/evaluate_kws12_nas_benchmark.sh | 2 +
scripts/evaluate_kws20_nas_msnoise_mixed.sh | 2 +
scripts/evaluate_kws20_v3_msnoise_mixed.sh | 2 +
scripts/train_kws12_nas_benchmark.sh | 2 +
scripts/train_kws20_nas_msnoise_mixed.sh | 2 +
scripts/train_kws20_v3_msnoise_mixed.sh | 2 +
16 files changed, 681 insertions(+), 1444 deletions(-)
delete mode 100644 datasets/mixedkws.py
delete mode 100644 datasets/speechcom.py
create mode 100755 scripts/evaluate_kws12_nas_benchmark.sh
create mode 100755 scripts/evaluate_kws20_nas_msnoise_mixed.sh
create mode 100755 scripts/evaluate_kws20_v3_msnoise_mixed.sh
create mode 100755 scripts/train_kws12_nas_benchmark.sh
create mode 100755 scripts/train_kws20_nas_msnoise_mixed.sh
create mode 100755 scripts/train_kws20_v3_msnoise_mixed.sh
diff --git a/KNOWN_ISSUES.txt b/KNOWN_ISSUES.txt
index 67df63d6c..cf930d7d2 100644
--- a/KNOWN_ISSUES.txt
+++ b/KNOWN_ISSUES.txt
@@ -1,11 +1,5 @@
-KNOWN ISSUES for ai8x-training with PyToch 2.3.0
+KNOWN ISSUES for ai8x-training with PyToch 2.3.1
================================================
-* As of PyTorch 2.3.0, torch.compile() requires CUDA.
+* As of PyTorch 2.3.1, torch.compile() requires CUDA.
-* Using CUDA, certain models cause PyTorch 2.3.0 to print the following warning during training:
-
- .../graph.py:744: UserWarning: Plan failed with a cudnnException: CUDNN_BACKEND_EXECUTION_PLAN_DESCRIPTOR: cudnnFinalize Descriptor Failed cudnn_status:
- CUDNN_STATUS_NOT_SUPPORTED (Triggered internally at ../aten/src/ATen/native/cudnn/Conv_v8.cpp:919.)
-
- This should be addressed in PyTorch 2.3.1 (see https://github.com/pytorch/pytorch/issues/121834)
diff --git a/README.md b/README.md
index ddba559ec..cf18db5e2 100644
--- a/README.md
+++ b/README.md
@@ -1,8 +1,8 @@
# ADI MAX78000/MAX78002 Model Training and Synthesis
-June 25, 2024
+July 5, 2024
-**Note: The pytorch-2 branch is in development. Please see [KNOWN_ISSUES](KNOWN_ISSUES.txt).**
+**Note: This branch requires PyTorch 2. Please see the archive-1.8 branch for PyTorch 1.8 support. [KNOWN_ISSUES](KNOWN_ISSUES.txt) contains a list of known issues.**
ADI’s MAX78000/MAX78002 project is comprised of five repositories:
@@ -11,9 +11,9 @@ ADI’s MAX78000/MAX78002 project is comprised of five repositories:
2. The software development kit (MSDK), which contains drivers and example programs ready to run on the evaluation kits (EVkit and Feather):
[Analog Devices MSDK](https://github.com/analogdevicesinc/msdk)
3. The training repository, which is used for deep learning *model development and training*:
- [ai8x-training](https://github.com/analogdevicesinc/ai8x-training/tree/pytorch-2) **(described in this document)**
+ [ai8x-training](https://github.com/analogdevicesinc/ai8x-training/tree/develop) **(described in this document)**
4. The synthesis repository, which is used to *convert a trained model into C code* using the “izer” tool:
- [ai8x-synthesis](https://github.com/analogdevicesinc/ai8x-synthesis/tree/pytorch-2) **(described in this document)**
+ [ai8x-synthesis](https://github.com/analogdevicesinc/ai8x-synthesis/tree/develop) **(described in this document)**
5. The reference design repository, which contains host applications and sample applications for reference designs such as [MAXREFDES178 (Cube Camera)](https://www.analog.com/en/design-center/reference-designs/maxrefdes178.html):
[refdes](https://github.com/analogdevicesinc/MAX78xxx-RefDes)
*Note: Examples for EVkits and Feather boards are part of the MSDK*
@@ -75,15 +75,15 @@ Limited support and advice for using other hardware and software combinations is
**The only officially supported platforms for model training** are Ubuntu Linux 20.04 LTS and 22.04 LTS on amd64/x86_64, either the desktop or the [server version](https://ubuntu.com/download/server).
-*Note that hardware acceleration using CUDA is not available in PyTorch for Raspberry Pi 4 and other aarch64/arm64 devices, even those running Ubuntu Linux 20.04/22.04. See also [Development on Raspberry Pi 4 and 400](https://github.com/analogdevicesinc/ai8x-synthesis/blob/pytorch-2/docs/RaspberryPi.md) (unsupported).*
+*Note that hardware acceleration using CUDA is not available in PyTorch for Raspberry Pi 4 and other aarch64/arm64 devices, even those running Ubuntu Linux 20.04/22.04. See also [Development on Raspberry Pi 4 and 400](https://github.com/analogdevicesinc/ai8x-synthesis/blob/develop/docs/RaspberryPi.md) (unsupported).*
This document also provides instructions for installing on RedHat Enterprise Linux / CentOS 8 with limited support.
##### Windows
-On Windows 10 version 21H2 or newer, and Windows 11, after installing the Windows Subsystem for Linux (WSL2), Ubuntu Linux 20.04 or 22.04 can be used inside Windows with full CUDA acceleration, please see *[Windows Subsystem for Linux](https://github.com/analogdevicesinc/ai8x-synthesis/blob/pytorch-2/docs/WSL2.md).* For the remainder of this document, follow the steps for Ubuntu Linux.
+On Windows 10 version 21H2 or newer, and Windows 11, after installing the Windows Subsystem for Linux (WSL2), Ubuntu Linux 20.04 or 22.04 can be used inside Windows with full CUDA acceleration, please see *[Windows Subsystem for Linux](https://github.com/analogdevicesinc/ai8x-synthesis/blob/develop/docs/WSL2.md).* For the remainder of this document, follow the steps for Ubuntu Linux.
-If WSL2 is not available, it is also possible (but not recommended due to inherent compatibility issues and slightly degraded performance) to run this software natively on Windows. Please see *[Native Windows Installation](https://github.com/analogdevicesinc/ai8x-synthesis/blob/pytorch-2/docs/Windows.md)*.
+If WSL2 is not available, it is also possible (but not recommended due to inherent compatibility issues and slightly degraded performance) to run this software natively on Windows. Please see *[Native Windows Installation](https://github.com/analogdevicesinc/ai8x-synthesis/blob/develop/docs/Windows.md)*.
##### macOS
@@ -317,8 +317,8 @@ Change to the project root and run the following commands. Use your GitHub crede
```shell
$ cd
-$ git clone --recursive -b pytorch-2 https://github.com/analogdevicesinc/ai8x-training.git
-$ git clone --recursive -b pytorch-2 https://github.com/analogdevicesinc/ai8x-synthesis.git
+$ git clone --recursive https://github.com/analogdevicesinc/ai8x-training.git
+$ git clone --recursive https://github.com/analogdevicesinc/ai8x-synthesis.git
```
#### Creating the Virtual Environment
@@ -329,7 +329,8 @@ To create the virtual environment and install basic wheels:
$ cd ai8x-training
```
-Using the instructions above checks out the `pytorch-2` branch which supports PyTorch 2.3. For PyTorch 1.8 support, use the `develop` or `main` branches. To switch, use `git checkout`, for example `git checkout main`.
+Using the instructions above checks out the `develop` branch which supports PyTorch 2.3. The `main` branch is updated less frequently, but possibly more stable. To change branches, use the command `git checkout`, for example `git checkout main`.
+For PyTorch 1.8 support, use the archive.
If using pyenv, set the local directory to use Python 3.11.8.
@@ -395,7 +396,8 @@ For all other systems, including macOS:
##### Repository Branches
-When following these instructions, the `pytorch-2` branch is checked out. For PyTorch 1.8 support, use either the `develop` branch (the most frequently updated branch which it contains the latest improvements to the project) or the `main` branch (updated less frequently, but possibly more stable). To change branches, use the command `git checkout`, for example `git checkout main`.
+Using the instructions above checks out the `develop` branch which supports PyTorch 2.3. The `main` branch is updated less frequently, but possibly more stable. To change branches, use the command `git checkout`, for example `git checkout main`.
+For PyTorch 1.8 support, use the archive.
###### TensorFlow / Keras
@@ -411,7 +413,7 @@ After a small delay of typically a day, a “Release” tag is created on GitHub
In addition to code updated in the repository itself, **submodules and Python libraries may have been updated as well**.
-Major upgrades (such as updating from PyTorch 1.8 to PyTorch 2.0) are best done by removing all installed wheels. This can be achieved most easily by creating a new folder and starting from scratch at [Upstream Code](#upstream-code). Starting from scratch is also recommended when upgrading the Python version.
+Major upgrades (such as updating from PyTorch 1.8 to PyTorch 2.3) are best done by removing all installed wheels. This can be achieved most easily by creating a new folder and starting from scratch at [Upstream Code](#upstream-code). Starting from scratch is also recommended when upgrading the Python version.
For minor updates, pull the latest code and install the updated wheels:
@@ -587,7 +589,7 @@ The MSDK is also available as a [git repository](https://github.com/analogdevice
$ pacman -S --needed base filesystem msys2-runtime make
```
-5. Install packages for OpenOCD. OpenOCD binaries are available in the “openocd” sub-folder of the ai8x-synthesis repository. However, some additional dependencies are required on most systems. See [openocd/README.md](https://github.com/analogdevicesinc/ai8x-synthesis/blob/pytorch-2/openocd/README.md) for a list of packages to install, then return here to continue.
+5. Install packages for OpenOCD. OpenOCD binaries are available in the “openocd” sub-folder of the ai8x-synthesis repository. However, some additional dependencies are required on most systems. See [openocd/README.md](https://github.com/analogdevicesinc/ai8x-synthesis/blob/develop/openocd/README.md) for a list of packages to install, then return here to continue.
6. Add the location of the toolchain binaries to the system path.
@@ -1075,12 +1077,12 @@ The MAX78000 hardware does not support arbitrary network parameters. Specificall
* The *final* streaming layer must use padding.
* Layers that use 1×1 kernels without padding are automatically replaced with equivalent layers that use 3×3 kernels with padding.
-* The weight memory supports up to 768 * 64 3×3 Q7 kernels (see [Number Format](#number-format)), for a total of [432 KiB of kernel memory](https://github.com/analogdevicesinc/ai8x-synthesis/blob/pytorch-2/docs/AHBAddresses.md).
+* The weight memory supports up to 768 * 64 3×3 Q7 kernels (see [Number Format](#number-format)), for a total of [432 KiB of kernel memory](https://github.com/analogdevicesinc/ai8x-synthesis/blob/develop/docs/AHBAddresses.md).
When using 1-, 2- or 4-bit weights, the capacity increases accordingly.
When using more than 64 input or output channels, weight memory is shared, and effective capacity decreases proportionally (for example, 128 input channels require twice as much space as 64 input channels, and a layer with both 128 input and 128 output channels requires four times as much space as a layer with only 64 input channels and 64 output channels).
Weights must be arranged according to specific rules detailed in [Layers and Weight Memory](#layers-and-weight-memory).
-* There are 16 instances of 32 KiB data memory ([for a total of 512 KiB](https://github.com/analogdevicesinc/ai8x-synthesis/blob/pytorch-2/docs/AHBAddresses.md)). When not using streaming mode, any data channel (input, intermediate, or output) must completely fit into one memory instance. This limits the first-layer input to 32,768 pixels per channel in the CHW format (181×181 when width = height). However, when using more than one input channel, the HWC format may be preferred, and all layer outputs are in HWC format as well. In those cases, it is required that four channels fit into a single memory instance — or 8192 pixels per channel (approximately 90×90 when width = height).
+* There are 16 instances of 32 KiB data memory ([for a total of 512 KiB](https://github.com/analogdevicesinc/ai8x-synthesis/blob/develop/docs/AHBAddresses.md)). When not using streaming mode, any data channel (input, intermediate, or output) must completely fit into one memory instance. This limits the first-layer input to 32,768 pixels per channel in the CHW format (181×181 when width = height). However, when using more than one input channel, the HWC format may be preferred, and all layer outputs are in HWC format as well. In those cases, it is required that four channels fit into a single memory instance — or 8192 pixels per channel (approximately 90×90 when width = height).
Note that the first layer commonly creates a wide expansion (i.e., a large number of output channels) that needs to fit into data memory, so the input size limit is mostly theoretical. In many cases, [Data Folding](#data-folding) (distributing the input data across multiple channels) can effectively increase both the input dimensions as well as improve model performance.
* The hardware supports 1D and 2D convolution layers, 2D transposed convolution layers (upsampling), element-wise addition, subtraction, binary OR, binary XOR as well as fully connected layers (`Linear`), which are implemented using 1×1 convolutions on 1×1 data:
@@ -1171,12 +1173,12 @@ The MAX78002 hardware does not support arbitrary network parameters. Specificall
* Layers that use 1×1 kernels without padding are automatically replaced with equivalent layers that use 3×3 kernels with padding.
* Streaming layers must use convolution (i.e., the `Conv1d`, `Conv2d`, or `ConvTranspose2d` [operators](#operation)).
-* The weight memory of processors 0, 16, 32, and 48 supports up to 5,120 3×3 Q7 kernels (see [Number Format](#number-format)), all other processors support up to 4,096 3×3 Q7 kernels, for a total of [2,340 KiB of kernel memory](https://github.com/analogdevicesinc/ai8x-synthesis/blob/pytorch-2/docs/AHBAddresses.md).
+* The weight memory of processors 0, 16, 32, and 48 supports up to 5,120 3×3 Q7 kernels (see [Number Format](#number-format)), all other processors support up to 4,096 3×3 Q7 kernels, for a total of [2,340 KiB of kernel memory](https://github.com/analogdevicesinc/ai8x-synthesis/blob/develop/docs/AHBAddresses.md).
When using 1-, 2- or 4-bit weights, the capacity increases accordingly. The hardware supports two different flavors of 1-bit weights, either 0/–1 or +1/–1.
When using more than 64 input or output channels, weight memory is shared, and effective capacity decreases.
Weights must be arranged according to specific rules detailed in [Layers and Weight Memory](#layers-and-weight-memory).
-* The total of [1,280 KiB of data memory](https://github.com/analogdevicesinc/ai8x-synthesis/blob/pytorch-2/docs/AHBAddresses.md) is split into 16 sections of 80 KiB each. When not using streaming mode, any data channel (input, intermediate, or output) must completely fit into one memory instance. This limits the first-layer input to 81,920 pixels per channel in CHW format (286×286 when height = width). However, when using more than one input channel, the HWC format may be preferred, and all layer outputs are in HWC format as well. In those cases, it is required that four channels fit into a single memory section — or 20,480 pixels per channel (143×143 when height = width).
+* The total of [1,280 KiB of data memory](https://github.com/analogdevicesinc/ai8x-synthesis/blob/develop/docs/AHBAddresses.md) is split into 16 sections of 80 KiB each. When not using streaming mode, any data channel (input, intermediate, or output) must completely fit into one memory instance. This limits the first-layer input to 81,920 pixels per channel in CHW format (286×286 when height = width). However, when using more than one input channel, the HWC format may be preferred, and all layer outputs are in HWC format as well. In those cases, it is required that four channels fit into a single memory section — or 20,480 pixels per channel (143×143 when height = width).
Note that the first layer commonly creates a wide expansion (i.e., a large number of output channels) that needs to fit into data memory, so the input size limit is mostly theoretical. In many cases, [Data Folding](#data-folding) (distributing the input data across multiple channels) can effectively increase both the input dimensions as well as improve model performance.
* The hardware supports 1D and 2D convolution layers, 2D transposed convolution layers (upsampling), element-wise addition, subtraction, binary OR, binary XOR as well as fully connected layers (`Linear`), which are implemented using 1×1 convolutions on 1×1 data:
@@ -1319,7 +1321,7 @@ Since training can take a significant amount of time, the training script does n
Training might succeed after reducing the batch size, reducing image dimensions, or pruning the dataset. Unfortunately, the only real fix for this issue is more system RAM. In the example, `kinetics_get_datasets()` from `datasets/kinetics.py` states “The current implementation of using 2000 training and 150 test examples per class at 240×240 resolution and 5 frames per second requires around 50 GB of RAM.”
-6. On CUDA-capable machines, the training script by default uses PyTorch 2.0’s [`torch.compile()` feature](https://pytorch.org/docs/stable/generated/torch.compile.html) which improves execution speed. However, some models may not support this feature. It can be disabled using the command line option
+6. On CUDA-capable machines, the training script by default uses PyTorch 2’s [`torch.compile()` feature](https://pytorch.org/docs/stable/generated/torch.compile.html) which improves execution speed. However, some models may not support this feature. It can be disabled using the command line option
`--compiler-mode none`
Disabling `torch.compile()` may also be necessary when using AMD ROCm acceleration.
@@ -2083,7 +2085,7 @@ The only model architecture implemented in this repository is the sequential mod
-All required elastic search strategies are implemented in this [model file](https://github.com/analogdevicesinc/ai8x-training/blob/pytorch-2/models/ai85nasnet-sequential.py).
+All required elastic search strategies are implemented in this [model file](https://github.com/analogdevicesinc/ai8x-training/blob/develop/models/ai85nasnet-sequential.py).
A new model architecture can be implemented by implementing the `OnceForAllModel` interface. The new model class must implement the following:
@@ -3305,9 +3307,9 @@ Assuming the network is compatible with the new deployment target, changing the
Additional information about the evaluation kits, and the software development kit (MSDK) is available on the web at .
-[AHB Addresses for MAX78000 and MAX78002](https://github.com/analogdevicesinc/ai8x-synthesis/blob/pytorch-2/docs/AHBAddresses.md)
+[AHB Addresses for MAX78000 and MAX78002](https://github.com/analogdevicesinc/ai8x-synthesis/blob/develop/docs/AHBAddresses.md)
-[Facial Recognition System](https://github.com/analogdevicesinc/ai8x-training/blob/pytorch-2/docs/FacialRecognitionSystem.md)
+[Facial Recognition System](https://github.com/analogdevicesinc/ai8x-training/blob/develop/docs/FacialRecognitionSystem.md)
---
diff --git a/README.pdf b/README.pdf
index ebd3e3d6120a9b6b95700b0b3fafaa8cae336909..efa543414b5a8a2b4a039e27b0a3b38fd8403be5 100644
GIT binary patch
delta 132795
zcmY(qV{~9a)2BE&n-e>k*mg2W&b-g_uJe88=UvtPr&sr0
z*S@Rj`aB+d|0o_jK@ec(U?mn1fOBzmHaE6|^8~HY(|5t=boaY9ob$pfsa}5Y@Yu#5
zm+rEe+{Ps6{H{>a_V9a=p+Z}=+5Kl;rrM_>Qh*QA?^qIU(Q$?)gt@-q;Sa9oJuv>>
z6o49_`Sk*nXFKO)-??l08}`r|cKI5nV_#3Me|>$LY#q-d+XCT@=T-EdRg*m%Y5T1-
zS7eU)b=enVop}nopN%Zf^XkgM6Bz{Bdz7=Iu2{b|;P}c%HVBWWNza
z*`GfUWJKB7Gdi9WnT&T&n?D65ncvgqaYia|f6@G`IbLElu&si^7!ZsJ!v7fb#_wv*
z9Tn{FTa7&&mIq#EEzMaq<%zsQ>a@IcbGA|#XJ&HRW*H{Gs5$cl!P`4FySz1mlvM3|
z9t?l0bWr_xS@*6#MlJ%BY$?y{yN~Z7X|_ZQ*x-gOFLCgr*%0zpXfVf&_EguGGvYECpNLy%}A^f}Ue
z0Bk8!A*$_6%Y|Nq^9u3^>$u=EfSv8#jYW!q1Mn5I26sQY&QUYaWVUFLt)wjZWl1%+
z=Tf10rTCybG$c0sUBa50%@LL6TdD(@3d{+|Yp2^3Wpt1k`VMi6q?~NG
zVoe*?Q6&+wcE=+zUVK@Z(@~j?;<g&mV(cM2s%&2Q$7%
zG_d1iJ~(X%?3+b1?qtUpeqj^&oqpiQZ?BU1`&)|k7%B24REHK)ky;J7cH339gN7_a
zdzGb54w)f|P-YGW{`hf(iFlU-0SO95^@u6_w3OqFC-4eiQkc7G_+L+O|J
zxbSf)-4GZn=|p6OQT4*s38F$Q+W0br0Kk@ip2?~{e`RJ3qh>yNA$h;a!R84%`jMEL
z^r%|A&MZ^+-jEKAQK^#VBvX)z2B<4-f4$;#5sq}iTLd|HTp$#@ORfzpx&fj{;KpEh
zFqXj{(h~=RdxVq?Fa>{k5$HLH+z-SquG#{+Gxo$hd>%~}6iNer`*+ZgTj`v+0rZVg
zrKj|67JDk1Mrqpf=rJ{RXobu~fN3f5XcVfld`i-M^dDHYFx4y7w7hCB!YX&%Y#!sp
z^Dwjav(ds)rOr#tRvJG;O;OHe$U@2U{GLd=2n^|s)~q)R30)QqJI%w04-I~@E^>_X
z_L{6#ho!JN>1+uq(Nq@%1{xcBeH9nR_OJ{&vDGOYsj6x3-X*
z5=b9R?kN$J>KpQ%sV8-{ZY&1eb#lA(}+Nt#>tYzw)#xuA3Wam*ulqaI7mn6k>SDZcL*ke
z7u6N3X2Ehv{^6Hil$a`8A6pt_xs<~uq75oD^Pof#$^4wB*5eiC)fLwYKgV+v*8?-n
z?(AnBeqms%AoUVPMH7%~100*muyG&gV84wcGIx0fnf9Mh#`IB*+ROIQE$GLzv=nMaesMKqd+1F*|6AGsY8%F>)HpW+W;2iIFg7$7v_
zQTDl&6+dw=luIkDCCDeNI}F(DI%^e7W6qD(qt5Rtb2_cFl&vF!fd)~WH@Mkkj-ia^
zX=;X%sJ%Jf6gC|37_0LkkSOtCqyQmp*+MpGZkUoU>*v5X4y6E6mSLiS;vWMupa0H(
z%n0j4$m~Z0EP@+jDb}U5({BE(nHW(FoQH)swe{~28fqDj56=GZ#?)J#QXN~uDK%IQ
z*e>WFF|8fe)pnv9_X@6GzFTJJ}S-
z?XcpWn$n74<|sknhKffLWDE2W?Bop-qtAvl-i$e(d&s59n&&o<1X;@>rMw`^BhfTr
z+IidTmzsleuXE^2eO$3`g2p^i%U^Mp?%JO#5s;5q*+^3jd_!x0+g`1nKlh$4^V|hf
z9d=SZG#z%*0l`Z~|L(Q!P8X8YjTT89F0dx$Al(x3>?`zqcZDb6O^zS593iUTc_=YO
z1Ur&(Pof4Gjr99Q6$3Xz;Yf&LHv{nZn7y9bt=P6C*|RpRSrIYgct;!ESU6>)9C0(C
zirc4{^Yy|Ke?9*(pv$Zp{Gbkp6Bk-QAcbs$M&$>g0}h1`qqL{1$v-JLm7J`y?haU_
z16`JeZk$w)y6BK6EITg0_DiL2@*EMFxVv`GF&yUwdL}Dx5_&W}`RjRXL
zW4XLF0X(#`e;04&R7;tUbmkx}KBQxV@}?x@EkX*sDp(Ol1gTF!C?P{#RrMSCM=EUd
zJ0IoZLH|5KG~1;!iF*>>gC@#&poY{s&;p~
z?(g`IRH0del=YNvM){G;m7Yx~!0##x;nkVxz@C`^aR>DZ;mGGvh$0b4
zrmO-_JqNM}p^~7^R6())q>p49wCJwNMGDgDosd}*QKbp4g@`WiM0s03GMf}G?gbcS
zL}N$;xv;SaEBZC?Mrgscd(SjvjL{!Lucha%WYH0KAqaiDM!k1*T%Z&$ZPPNX!mHyt
z0M4128XHQb__~zU5B8a#SM-luebu^~R?a1#SdfGslJE@`6;6>hXZ4Xr8!hW__$F$I
z3WS@9Likm!v^fro>KlI+6^n$O&Z>+r(lKi2OEO~QX@X)2<>+Zz+`%>Lz>r^HreKSlgWN78w(Cm298ENYz1@>7{ZU7
z5uWF$>BM(a%Pg5}ogBG*n6*^;E44o(5VeA@lZdzDBNx-8$^@-oH)v;{xQ-qHKU|aD)i_;~NZZI!oj89A
z_9%vQunQ)MUDY|M1&g%xGymrWYEpf7bg=@4>p&VTZu1zQI8dCAj+&LPI+sUzA{L9W
zk_`DXYH9J?VJh5uMfnyFP$XMY&Y>%;RPreLYEQ^BhJeJ{83Io0!MV#+mRwkM(Lf>9
zu(8~a;>ValYio3(Z({tB6Y17H8MMx!%tBWLTFds6((
z(+GZg>2`wSmqhK@dcB0qK)KHbtTQq4w%*qqN#iF0dXFj_-1~9|fO{F#@Xld$&eOI*
z0n4!4lhbRhw*?X&Db))Sxp@u}zg4LC8wm`w+SX}~oCN6ubC;4okfSag_7Nw?#3?oy
z=8&uNu`T|OsS=@R)|lgOic|9f)`y4|sbL-@bEVxKxKUrn$pGsH_+uTu!ifkPxii=b
zYOGo=x=nXe99J3+pjKLpy9U=@uqdK!ikw-gPL%~wX7_?~QVrAhxzeaHe$v|X#rMbP
zUYvXf)OA#!;543t(xTA%&Yr8b-HMfNtcwlclLjstTvxOUg3HdTw#$Zb;7wll7(jqp
z2GYKYH9p)xBW_mX5-T4h5xf_0uBIOi<|M-0Djwx@5o8h%U>lcG40CrmHsycYK#Ri7
zR^s!1ZlarN#Wd@3R9kvK&~j$?OWYO()3qOnl`EEiGx6JB3my^Et)QbRq2F&@8Zg63
z@E}Jakn4Mcg<3^Y)jDm=ijPHmLD@E*LBB=13tl>~gm5@8eFHnzqPyAF@4Am9^q>)0
zt#o2KWogX@ocsJeaR1$8TTyoI)<=@y4u#P$j%yq);@pv%CZ0HnXy2#q`UWO*pySeoN+8d(ZN-Keco=h>19{^6g$sBW=
zj%$L^#%a}|ONt;@omESqPdk|RcqmlXOm?2Gq89%-PnJ&Dta7$V>#^%zQa(~xU0@k~2m$>LS;C6LN
zDVGp(^m4A2xSQZ&1-(7Q3so88)zDFi_|F3Wn~kDR)a?(
zPYav)ak=gsV=oBAr45#&oN#ZoX%jLX=ljHhAXA6289sSg`6E?i&Es3xlBN!TW{N{9
zc0xnRMZ?Vj+os?=y_?~mz)N`zl2Gp%RM$dQQXI?vdI&LEjz$&Z`2~5!$%z7*8~x#V
z^bif``Q(Rxnec(?lp`s#g1F@v_zMblZD2rNG}usN{5Uu_O%P8X2=rb2lj`EOaXx8
zdV$TH4N51Kt!S%
zXx0)(?j8{^bI2ed{lP{oy2(?~?P3!P*`*>Lqj7yynU`c;FSz1H=6e5T6VQK``%Pig
z%`VU>?c!vnA5RzTo>3PLc-fR5WecjaMJ-{#zqUf|>qTE4VCu&3)`PbfI+97x93lJi
z)Tes|MXf`v)yy&;{MosBKUZD2t*+R1D$qfstg|WYC`d`JKyz9CSC-0N%c|=$Mfx?~
z_2b3b+jnQvn=Cv3k_6|0HqHLd$8<5TgKl)BTI)!p#ToQ3IfV`)@J=%NEO!Fe?&_Ll
zjQ=Ddz45UvYOxOieut7Mb)G8Kb_25Kw5vBe0bj2;)g9oMZWCW;FR|nw)iV3ipV;nJ
z8kY2<=cVq)iXKk;z&Qz#-C>rG$SVow{k8n5ESAZ^QtcjFJS!ed=;y$P>pU7Isb*V4
z^PaVfi~lVmhO28k@Wq5SOrtcdBMlczrl?dou?UYdBV}
zCTy%~I20~oCSn!h{~KmY-?cI#Np~cnf(1Ca{vTViMQ0;zaRk$^uC9DEt0hCCr1LM-
z2s!!c>hF+h%;Ufqb|i%`q*7LePXWowCJ(o&MQmmQEjsp&rS^-a*^-CxPb0xKxzA;R
zU-Ic2hCc5z{>|~dpBIa|debW_>r(5rz(aYz$@7U-506U${
z);le2T{2Pc-l(D`V8M60(a-B$m7aTYYvH@BJ4*B+7%scrmV13Dvw^(Ne$GKucVT$>
zhWwQ6=C}e&KRx|ZzU87)eF?MB;tUEMa;Yp>8UP!dCikFZm#
zEMx3R-r?1efC5xMVJcC5d@ti!G1+=xK-W%2dluC~tKM(P7aIp%H{1B#3Eqqg%6m)}yE30LiB_!$?oXpZ^A{MAnP}L7(=&0ey&(cq$WS6J{JBji%
zxBm-q3}>!4wm4qPxUXL%h;^53F2`43zd}huj{ou7dB2ol6|Il=+0u}$6!#^I>ntAP
zYoDb&kPJZ`jD*@DUT6*#uo?>24p+)Fa}LWbu-U?EJXgX8)+;2Fb>+XdKFYo7BM9qV
zfd}eDEbL^q2_d<>wU1R8&S54j)c*@(GbZ#sQsySa#*bMka*fur7kr9(hW{0PY#)|O
zdm+jRg72Pyg(OaNk?!`r>oRkxJH66hr>-j$*e~L$jKx^T$XyP_il|3iH8KWeVf+q0
zzG787lElU_#M)3Ev;;~MF_}Tu{ueffIWVOQp<3Z`7rE<_PYN2=u;3KrNthiJ>_kJd5xZOUL^L@g96
zax4IIKLUp&U$#=OYzuWrpwPJ%&=M?XIh6Fwsb+n)YuI4E(XR)M>#E0Ym91vYqa!0;
zZw@rwh3tYtl;4C6@VsG(2skXlN?9H!KHHr?TBdx=#9+(~C8y*A=_yUp@I%~Ggh_ID_i<*}eIC{NlQUFoN63~5_o
zz@eC#-B2Y*naUg45O&Hjbl0Jpu<<6pkx@LrQ#X}4^S
zqz(uUe%8l6p{vyfCMJ4rgNvqs$zNQYG%vTP#7t@x7XuN>S0VDm(?}@J`Uo;|5^Q}7
zx*?_-_z@IIlL!)kR584P(q9&hQMj*|`H)_9pL9yeKgj;J^mmqaIN8iTzlas`s5_fb
z>Ik6}+V8NCJqQx|em=25$cVARfC<4x2ee!0C&wbFX>zWuY-B;8Sy51mJl5-2gsBQU;isnc8_U@l
zApp-(W|nDd-N43mU%d~$Fo&Pden@AoF&U1%q1qz+uqoW@Dn4yoH7LMU3E#S4ztG&#
zR816*G9sP~RD`|O4JwM=A_mX;-X#Wa({Z;#PPJ5>0fQ1xC9h)Al(9s)scAPs0C8%|
zdOo%wkX8%e&d{;;vE>+krHzT5fr+Oyj)rHHSRK*8Y=$a(4fjoJDkW6mbWIXp4{f$1NQ6
z4k;T>h}0IT8@(73JsBhNUAB$_x(C=#MvuP3g#bqq7;KM9H%W;q7Rj
z<7ESPUbejcpquNYv>N8RX&lgBYz3|^+?`clplC;5Zs~#1h#2W}`Y43SdKjpvY|Q^*
zGDcAcdslONR~O=RAwql>6qf%){}qb=m*CQs9gOhPQS89^k_Bn8|C9b7u+t@GWBs3%
z$%eENFmUW_|JjvXLn{sg$HD$z@%BH0ljFZalJ1}2;{2~jp(FeM5RscLorHvh3Si-6
z`F{)4Te|u#o2{6Bqjh~`OBFb1e`m5y$_)40;<^bfOHW<#Q)L!)bRl_}iXRn?8oXYNRdpo0F
zGbV5Q3zyAPE$)C`Jx%k{?5T1;bh?<9X#KN7h3wjt!~4dG`AR!C^=7Ng4ZO
zW44_=fpA8zhC`<90XZV3{#e7Oyie8l*>wC{A_ob#a^_06B}J|
zU7r)#kgUMzD!G3g1@c;C89muDUiDLBEB!=73v_}yY@#20Ie5M+@-9jt?{g0nKX4!6
z%cY9uLedosbe`)$xceh)n!voi|4BuxLX-4
zAzuH;bT1A+c1FJXJD9Rct$4DcaV0D4SY^n*V>#|3)-kKi$s
z2>K!eV2NqJY$)YsxM8Sui^gRq+Kl>MtWM)w#cXh-ow?jXoEDw%%0uS4VL;~9)7{Rg
zLdyBYQ2J#T`&^jBgoU&zDlDNDhFom
zurUKPdzKxg4P6z^L)y&J*->L0lwjY}uS+ox9MR7{lEEnG5=|^rZOVuc&$FyomJa?x
zgC_s2e`s!eRqA~=V!kY9dqq+#*zt?x2*o7e41<0lu_$t4ZoS)=jSS_z7|(z6Qk_Hz?@ukhSVZL#dkkZ)iQdMcF=_CTR}xGY$N
zEaaZzGq+dY{X~)E6gSlQF8ZAJ(c058fRsb%LL#PI9RnOd+QyRl_YOPvz7V+UB}Syi
zTAcieU*5?)$&z_XtL9iYdzEuLL;=}rDN&sfO6?0rJo|_2A-@$vtUrIj>t{qE?}5@q
zwe&oK6hh4sqCgIyeLa$K13^2hQh9N>6!T_~vasz2+5^r8s^O)!NtE`!6~$+SdsnT1KVnsN6EA;{FU6KVvu(
zG9VtC3b8`!IQTLiN$ICDCEl>=YC-VMqMS6dql;;9(sd-v6{a3w2C{j1lfVnE*=
zA_o$Q&m|zYH^Ak)BPmu~RGW6?+c+B6I;)K92M2%-!6pec1UwT%W1rq$IC_*Whx;2^
z+TXEkYke#ouhYKM0Dg!9Ykk5!=~0pQ7-Xbqfh+|J1k1`^UaIV{2ACf9jxZHRIBa%t
zSHhHkVx;<#1$!j2zmR@y=c~YR+DX8uO<7B@MRZknP&z^r*k$Q2rJW(A1qryit|m#>
z6UObah62iJ%y>MZqXh3)Wk#Cf6oK5^$X_56QdPMPaI1~dKnmd*&BEY;xEbCbkW^!A
zH=`3PR^{i9G$kY+L3S@p=Ct7)%w>!wYlf~r20?9ii+Zhgoiwt!$paTYKDM_<;{&KK
zkH;Y4u8nD;lWQnB4Exo9Fmx+Ss-og4?aEB30O@P;-kbh?0-3)L_v
zRSdkij)brhNKd#mhcOkGgc3g^tmKD{ey=1
za9Xe!cz(jStjJ;E7=t1&I624S#P7G+qu@PB%{~^08~x>A4L1|`Q$1t`En2mF
z=t|_ffQS36YMj(*A$G<^O{h}}N(Qh`X
z#vO+Q@JEc%pTx3H3ZwBD&%iiEi^1>2sia!z2}%I4iL5?GOYa2z*KX1?#G((R8Y4oo
z87Wg&kup$D=E)(r2Pl45Mb=hP4Estn#vzVgpn+uGXt92HDXM-<8w!-9n+P-Q!xG0_
zI&trnK&pSZ4DLe3s_zJME$dhZRQkC
z&9Cbc8nK?p)KliiZDmc013t5F1CaJEz?~10_ySW=pJhrsgGPJCJm1q0EbT>3o~dsS
z5$m^7p6K`>VmKxn7{TTOGjtP~W`yoB+k~u$6)vVH05AUU
zU(2)G&~IIN`$7BQ*k^nQBYJw+sj!B0Etm#J5K3MD_W1q
z_jj$B(-WiwTY^n#Nu+*k7u@L;$>DhKTT&$)h&N;9`iPG1ymx52_8iiyuAmn!nI5M0FE9Oxz;&aAw4hMDAX{4Op0_|+JQ{k=R9!W!g6oM
zVb2O8r92-Nlp*Hkeppehs;#aadcB_aYp2EL@2|@@_pJA)Ex-54C%gB<_-&8N=x002
zuWrCe>j)^}=d#_HZqwNZ^m*-^ceDdTdv!{5Pl035iDRGv1tr%B(s!o)%UM!|Dey+!mIt--F-Kudh7zfA@GU
z{JqiMaH|@ZT`lv{i@gUS_;BZKp51OX6!2LdyZ|g2H@>W|KOf^mP-SNd>Wowh5
zSB!mmH7>+=9LDLj~k;%VOzyUyY^w>}{zlm#2I__{ZQ@
z{eH5FK~g!GTg76%0pf?z5}YMXr?-uxnHkI-Iqr!_|1v%{XSKK6@P(Qs_3>%s(pgjzQN))Z;A?R9lss`8UJ2
zEzpG9PGuwn2{Z;vqxwe(1;Qk9y*2G;u#zByzL%VJPWo^Yuhy-gyM{W$Sj-oLj{&$+
zq#)X=PA{jl%QI=k4#lNDR=Xj}IRNSAV9>>_{ES;!`uMh2P}#3Ed}MR|DgC%0m7ln4
zK+)qbg$D+iX6leTp!IVfW#-lbqJi)$xuPD!9gm%o2S~xPt(C2gI}H9KiJU`F#6p
z^Eem|EFDT2ay{WwE4FX<<(9-7g3OtE!`?$c!!!X&yJ;)v&kIFc!31K}Rt(a5?7nqp
zfB0pq;U$-aJDG&Qh&~PM)499uNtB9$_Pw|2uyeCe2<8Mku_b2t1oWu|JT{IIR#oa@
zsWBwFkP;S{&}*}{JowAtDuB#R4XY+DU+`UKPGpTtdo5=I9a5cSjZwl1N<6%Xt!Q@M
zErvCKE8a!*5iO|F4^)dQ{o>m(ZYtvdPQcI6H3>SKA+hMe2b>)EcXX~nk=H-_Tnt!fO3W5z4=LnkQB;|mrp`15CSfUsgniXovH4K
z+iP))n4&F0Z2B~K8XtWEDeK1V;`3KE)xqQs1U(UMh}x?sSQUB&;E^_pU5g2)HQ}c0
zn@{T%Xpwh?4c@q%(58>SW(D~G;A>YkW;p|mw0zJgk@0mi_)!(ZhQRZJL$#A&*P7L$
zu~=*ByM7pG0Me=bInX$5-=gc0VUWMeDbbKUK8DbNfEc;D|QP2p?}r^!00zWqo05QYk9cnMYihF
zgp$}DiwvP9&FSP>=_if)Op0OgbRYBPhdA5*2KKCVZ-}LbzRul*LQ#%XBiMnPasYhf
z`=LEyC#AVqtzk_kd%4_U1O*v}R53q=?B6bovld4seprl7#CKWiaI;aEK(u90jxaIm
z85n7D7rZTibR6qM%4fGlXwG->6ocIRePV&|Cx@hJUhOE@Dw`FkOjGW!*&U;!eC}D;
z?+XmFLgEyIIs|y=jw!Z%GE;r17&XEE8%DJ_TSWBCbwa1Apo&fLG(zd3~Fg@)MS
zrzS_pUtsmkJs9b?Qd
zDig5kQ_WA|g|TKGaF-`)7WfIK-8vpd%P0#VtPxtP+cyocFm-`^xMsDUtU4AWriT--
z`%1wA^rfJkgdK$##^8`O&^@Y(l)DJ?21%iYQ9)*boWcn!^*Zqx1ZnPCGfN%S;B?55
zLKjMm<0pBIQjH?eXe+r~B@TbIBTO=%Avo4%gP7b<&>M$@3xunEXV`PcXkdzn=r0dcSN{8s;E-$YCaj}pmWx83hu_N#U29I$
zn(E&tdX+$mbRtFfw4#>@o$S6Nms$M)P(cVnL<)1*z4ehJ(%71}J^xAxHR7MZJv7}k
zx)J#ac1Dp~yUZt;Z7V5=lAvE#>$B^OOG!ACVWAtT?Js=dRK0&Yzo7!B-DhKh!}^mS
zV{b8aoTOoAi-5zDMGiNRgqWxIZ5$HzM|dI&-gQr|e;l)2YvUP2Y4mqkdJhE+Krd+k
zqXX;}Jp-<6k{L4l=OD&J?I1?SjB$t98gor;%aCj{x@RI$(_nwO7w+!EnmkDZtdE<8
zbO^Q=mDxzk9ziPYcHmG8)ouB7x$F@7O!*HOc>PW*6Fr7up#_gY1m2?VgfXI6LQ#qP
z9G7Z>zvFCNEsW%c@yDlOyhV8?z~uO;g%(k@QMbS6Cp{-NVWyYUsR{Re)MtoZ&{BYc
zC0UGF=bGuT_zDOTOfpljx+Q@}GLH=ef6@U=a4loztvHIMvWz!ILk~#7d)P0qernAeM}s!HWw4wuathewcK@CG)yiRh2MFEENA!rW4phapz%QEV}`JZwn|DZ$n2eE@@&5Gb3srE=2Bljg1yaAK@Vm
z{Lng%f3EYt-xy9g=XHfHm2$$#Y}cV;%~rUuabH!+>O{m!bALq&;Y?gBt{FDmvI7;dO8j5_3iL%UQzm!#AU+#
z?IT5aFSrpi;$v5L9VOQ1D5q9c(EPdPkc(0fgD?8bDtzb1k+P|3c|IS0u_ih9BV{Kf
z5{rR$VbN}<5tyY2#a5@7Lnm@wbSM2*SyE(KKhzLP&y|x^^^qZ-daT=FBIE{Tp90G+B`M~(Ugd`c5j&b|-5
zTzzt0U+we)!))TwQv#%?Mf+nTkC*w%EcT)BmkPKev5
z$qMBN{-El_qKT4^@G%Z`FJX7EHS$ra1iO*41jwXxQdtXMh&AzSjD&(th`4HWGTUDg
z25+VlYEBM3jH9V`W@lQawLau>COecMcqci$F2oXRz*A~LF;imlR`H;kT9P9WM1Q8s
zV?055;78m$e~QAN^`c!`=R<0M9~2U;u1fN2}bPCS67NRYF#QMA}^IWaNG*Bzsi
znpVaJbbj<^cMdjp`6?k(?bI=6%ZyTeD1l_Be%y+?Ubii}pwcIAO=)dRW5y@oZPR|I>1mzj=r
z!Au2z2f#YvzL*;sXb0~Bk;DlLAKC)eSd~)1qGPIE?KfS0j=o_R(5eD|o1H&$6QaxN1fcwfSsV`tMh9qDB}R;ud^AH<-z_`K3s2QD
z!K=Pkd3}<|jVG<0Zl*jbbl(p8`G&f%NWQVczhM_#3*cRLSrC9vST$w+6)XS%fl
zM?o!4e18a46M8UN{}2);y>VX
zR^+b^7Z?K;*xd-_9sh=xdqnO3=oE4W{El`Q*^RcUE0Cuvi@UC3*=z{Lv*CUmaz3-v
z6Msa%?s)o&?Cqev*w)(4oo}8#%!P&W+QrXu5~yaT;_^5!xgF`|lCI4Ls?+#o
z3C>2J8zoO*I6k^V4$zsT*B6X={H4gLP!N2&n1`ZCkCD%WX{5$J`mWF?`^&~_iJXR<
zW7+cqvki{VJim6C@U5?q1`AlJocK`EHVy*HHU6Fj6eM*llJl^;d|8D}Ah3oZStcW2
zPJgBxU=U?>UU1yuJkn)8wOCgJVA8XV?0cbEI^`7icH*HYQAC1C^fALRZ@*W+7O4x~
zypXN0deWqw=k|{r;5clm1
zqo^8X$~}SokgN|mun(COVKN8@r%zj8d%_j%qL%>Ybv9~%h%FjH(eCOY_JV6Jm@gqw
z-}#aZkUbV$@K>+*n-6q=nOb
zfoVx9QOyNTwWh~K5taj7v{j#Qwt+cj(M}Gb*!0Q(8tR*&9i%ON#o>!_|l?g~mY2-u^d|%0(`sNl;
z3m?zFoE*o-8LCQ2s~wG(aOAG1L-XJuw(d$%`7fN|AiUlR%kB;PftMJ@VAcmpP!{J0
zxUKc{hW(qw^u?q6j8@+@k%A5?%79O(6iK?0(NL$sWza~ienCUi%tfdSGYgIoPQS`(
zbZXf!g70k}9RBO7_xpQY9gB2F)s_=l3BFj-l5c%|;$4y$Z|N_84TCM~d*!aTbIn=i
z!c3O&O^oq#YZ;U@UhA+HW=%nhhw_8z!
z@wq2xp{0|URzH*-G_Za6eqrZ;wA$bPolnK%=Dd8=D;ufPXG?EEZ6z7>2wts;I^Yl7V*!M2}&VJk6n=GxLhS0
z#KG|;WbXu@(t1oH7dik-7aK}wj>`CME}lVsI1w!J?^(-Wdm2laVb)#OO+yqS<@$Xf!QJ)pEy!8_(9L~R}ROoC57qFe5#+u<9+Q;mscEcbH8?t
zyFKd7+fYcPs8xo4lU3B*PG+L6_hz=KYty|FR@qU~bloSZb@=DI$hae{V|73pf@=WD
zj(u5kw5g9?{@!ttqYETsT&hqLivd1urZO>X{G5b9hFBzWZ5O!#T?tT?My{o9j4M%vMP$+QR*->NE5V@K~J=tiog;{mja$l
z8{SoA8+U;UL%S|SHZ@&HuD*}Eu&~N$TUP`oM}7s7OEtan`-#KQn&$tZ>KvmZ?Uuhi
z6Wg|J+fK)}Z96?NC-%g)lZkCiY
zacxHx)H+YrW3_Fqq7eR@1hT0O_mnkQ49ZDlS?K+ZriYdLb2R232ZiAf&Ii(2Pr#lF
z<8rYQS%YLakB{9mNYxPg7EA?lp&qPcbSP^
zgBa6l#MY*4x@h@VmrWB3MZ=S*@bc)@PW;&S@rpAg}L5bs^Y|@f&E&D3R*Wt)R>~?=%wR6
zE^~Vm+|^cfw0~6*EHrBCV}sx3H*YOfzQ9P$JDWE(J)D>2r2)gTQnKr|yFtmo>CV??
zr}tgrVmlmH*MlRE&Co)J|VaY|PVS*6f~B0jA8w(n{Y@Tmf6EVh7&<+*t0
ziQ8R8HAONyxQHL5;xdOAJmNAGwu&-iL8N)7(~N?&lJ*D=89wiC3XhkKV=q=*vzymo
zlKMP9hX;q#hKBJ+N<3<^H-Q_xofkhk<$2Xu;JC)U(*DG+peL)94Zs#&ius3~sF)>P
zJJ~#l>PTCT_+Ql-2@1d7bhaVj_OEfNho{GxiMev^n^q#$z^6|_In6-EE)47ATLgdb
zLZyAR{d*6?DI)wM?jYM`C`3&dR*c;>6(V=JADz$1BX|@w;1D*zWPstIp;`#MyQEWC
z&il*?jROvSZj%Cm0qMNMAkx?
zE;$!H8g>fuSk5v(N`RuFpfy)GHrnJjHLFy?PPuMJ}%TQ>l5Kr=C!R^af;7<&P}>6
zM+JLB2U<|6gNwXb%(FL_AMi!LogKs6NT$57Zs23-KJ@Xlwt!s{0iD_<0=_-H@9{Qh
z(n+Hhb3gb319Eg_M@dZ_;iGThyJ#1T1II9Q7DCh
zvosGUXA?tilObwnJU9!?6n$NAV~dGNa<3>}Mo4C8RWou?2{o(8vE=(QzVG2Q%h4+s
zQ!&Q2^^`}yfKCMp0uhK}skbQ!{4w@B>gbrZA@;)0P;JdRZpdxpW_JyL13{*Dazs+g
zvn4Rq5+NC84iY~K7-jdB4@VAi!fjs&BH9B+DbQk>Buh9NF)~4aC@S9JQkNp|{f$r|v+>uI-!qq9RdzrS
zW4{Nk0|z3)Q2`zarm8)vWk|)w93#qVW3|I^tZ2Q+KfX*s
zfDT8-84RgU1(
zWkVBTQ!JvOiBS=FDo1(ktWXV5o=b};OVM~KU%A~C)$D~MHLpDmq!X95jnDz9rKrET
z35!=&v*^>!ZC0JQ7f2>kUfYTDsw*pGW(Js~^6S&3L5e#RuV|8f5e()nduk5l-52fw
zc<8mTym0?W1ZIBD4E1n7-^pY?La{pnIrRDj@dR4dk#{2hN<&rMUjK6qz#Ts!jAbMPw!$6~oMEU?-F85SiWSBQhyj3l3v2TR&T;
zXg`&kwoJotH+w2sLzX9-jMD@Y=_RxOGhQeN%0uE1oOH=0Vv}j2pB6=`>%h(53T55E
z$14zz3jQaYE6Jf^O9@%7Ww^OHSMOt#tCErZT>@}vm^rM%1)r2k(|f~Z}xUfSQmEq9y)&$$VjMc4>hR0Y17>X9Lt^2?Pc1@rq#M%lWe0H
zeACF+nB7PkOyzj1uWRVE=7C?$0I>VWwtN=if}dYS?RxE1^15_8+Yw!ivJ>f!YpCP;
z8=gYgUXixxZUn2pJR)tCg$zy=wWaQ8`dz}Kiz$V140ko6IBoe&(r{wPupzp->!R`X
z^*u^(*qFNU2mT>}D$tN2=A?e3Iw@jXV20X!Ft4OU6>&Vfo*w6l*(=
zP#r#IZ0HGa;GQXfqV31XVlo*#8LxaKwIAi0#e-tD+9C9C?TecAKbjVxK|fTOj!fpl6-C`^r1h^5ZY
z;hW=pRs(?vBU;ETGD+j#!nIXWqM@a!FvH4WifKI6utK+ZYXT1+=S+_c$Jw-6ieP8e
zu$#)`ME@#Pn4&ioJBwG9i#(rL{M2@0e@CwJzrDI4_qV-6EgrQm?*7T9Rq1f)$nG80
z#0N+St7^3z0oJmQmqk=PR^UCr#E(RR+1o?Qv2u9TYCx?T%fnXN4GO;S;-jS&%cj^y
zB9oh~jLZIh>9jU%3xstvV&j-}R1#d5IdEUDALK!8ix}*kjv;I_y!hVG$s~B81SwrQ
zZbJjjKAMr_#Mu)sy+Iv;lN2P-3Pz+CP#=H%eT+je1$bthG+5NA7*3I@A90f)8}jw4
z5l(M=P~v#)v#1v4E=3a?;qL$yWKZuHSY;mTO^q8EJLc8KNWK1L;2m1228+G$W4zcZ
zn&y~vSrZ?0HNcGrMnrT+$MZP?@acyeB{r?kG?(R&AGg)#j5&j-7E@isau
z!}^eduZ}>;JDbATNm#Km7zq~VI!PGtn!?o6>I&U|=9`AZo<)kNvA!8zsKJmLwKHLX
z@`S!}if#A<;=!)uS&a*hX}J({2ROu10jbY7AQFQm+T$pZ=;EPVUZ#iZHZYS^9GxtW
zK~0V21ja`>*n^9uNEb!z%j3hz+>Q#R^JUc6#=9w+kRd7L>+)$+VaDozjhV#U0IpJT
zwP=!76|SN?0XUl>+y~lwc21k^l!D&Cc;wSsurH~<42*veI)H9`&Sj?*RaZ2@RrLd?
zuZNbuB2h>#LlW{U)E?QY#p)GA<#ZeHy=7m?6D85|IH?+UG%O9~=C@C##2T7VVtG1C
z*i^JMw%!mN!$v|W#6eL`qWE>Y(C`kGf6>)1!LAn!%v`@au$&clT)f|wkR8XzZu3)2
zy^_C{(`Ua+#&HiDie<1wr*e2$;FjY7FXETa%;8up+atP}L$~(fF~!%kXd=mu^SShC
zwLCFc3f#$M8N}-iEHP7!feR;JGY@dR*5*0&;RTB4W@z`fh)l(D7=f1FBuR!v9RF+J
z7Svcs=W40UtlX?H5z0lMk5!b)ve9OTibd4aH^W_TjzNz1=V;^gQLMG4&5?mG8boq>
zVN$qxn?GEcXkO%7tY0Bqg9INt(D-3p4kB!B9TnfszBB0c>28()%4K8}I~`*v%-9xE
zWSu%#a?pfR9+Ljv&mXX@LN$~B>WG)`V`8-6&aFfaOw>;y~CcCR*xZs(B?MVv1^|!x1q@J%mj1zWVu3}!+
z!A}s<-hS?x*AfcYxIe5}Z)3m7P{erbNR=()qoPN}b9xQ;euA9*~LBZ2Mrnl;eA^Z|D3gk2U}v+#+W6&yh~_1
z>pt8H>*`#M;A8kzcqPa*Jlik4>@#h
z+p8CULHm?F|8A_XcjpHEOU$)?qgK63tJ{0yCcfxS@_mD~_vS6^LtyXVy}exuM*F09
zE8)ZY?khylq<{O0+g8)EyVo+_x5|`ue@+JX{+9L2SEIU3g7qLVL!f`B*2BB?uQE&P
z4MxO@Ndab}>kIrFvmsbVL=3hfvD{d0c9n<#ByW`=PTncaRwA(G*@Yguu(7yV_Vp3(
zx&wip8i%YybeO!Yxb=I4q$=Jw@k1q&)Y%TtABPOm^jqR@&>2=Ibd$Ai;FIPjc#|6p
zqhkpe&VlCRJKPyYr$FHhCA#HfoT
z=vufb{pQR2iTGaY2vpiyoaUfb`6^{3JOyBU_$(40hO&#CylF}sJ+ahUy@?h9()6eH
zWd36KwNI}OQ0@NfcWq{q~In!xIc*7S9k5OA)
zLM*7om{*X+5DI5Nf`{>Irra3#J@a-XIyM*TBlH%5I%`>w9eZolW3jn_+u%kC5X
zx8YqxMqM&Ouj4+0$P9R`zE6KMVBKbwM*ph*{Bx8yaOP8Qovte|@qPDj%i(F*+v&FE
z`y0|1eTMpZ1-!5$`fWfTx6{@B%WD0oN#SH@*Q=!$1gkkCBvj62^a;`>1kMFx7u
zq#3kC|fWOe!}&UBgDAiWEaDdUi4{Z4`IVHkj%tkvh4v
z5_49p3|LtIb9P{H+<&mmL|OVYybAP}Qmpa|l9U9spBl~4v+fthRdn4J$nTF27})>q|p2w^!;tV=R^^*5#l=@E^uZ;6@eah{0&=7%G`}`r7eGqc%7)
zw(SG4av9AN#%KLbkl+apk`M66POeH=CMHdHx=^U3=HpeElKPS~B7zlM=bWrgZLb#}
zmkt8!N*U8GC$?B`jRXWdqdsurVKbXEc|gi%Jp2xD9?#PHzo`0ay|_xr%c!daNB~>$
zvfLOqoyG*Z7C_)wwjElhHsrmx`^Z#>jEJDGagfJ>7!)C?-Si%;CO
zNN9v@b)4WLVQ7(H@DF`P&b4g%!k?AMSMLe^{;@J$e8^m6l&B_$e6&nOl*t<``ar3c
zIJrYmz4n+O1AebSJ;yJYTaSlY=mk-!3MtAe-U^nq#0b0BBwD$B%MzBjzUnbzLDh^4GO#bNyDglgMbJfed6xslyhtH#W^;%e+n)_aynLt^QEaDlbPgD72cc(bB*8FHWnnr
zkfM3bYd-%8Ca5-_T37Vh;}o6DnKM+vjwmC#|C@6o^99PdHvEnX4_GqIHITs;a~mOz
zntat{HW{@jdTO9~{Ck5uMqL>C&r6SS`1aY&?70vjOWAB1eweS>D&s1Kad!lNA7Y!Jqnc8GxAP?H$p0BfH7U%_CTf(TNUi
z4rreXPDvbW|0*)L$82vmP#N*JXC7{<
z2ufbhuQJ6+JKQpO_a1=Ws)VCc3ioaj0$75RjAZf{=t2W&G&Y>|6{OwoZxu3w#}>&m
zT{)wFIWM=~@A|_m{<zzrCz$*BO{)b*!>RA28~3$=LOc`Z_6(8r>KZ9ealD`3n&fj7s&`$}mOrKoWRNGiBAGnKpc)D~sHy
zK|e1Bd7U4Uac+*^DZc*)!`k2T@uEGoW~<}Vk`J}#8G%((pZNnPkS3qgM
zX^@;z`H`ZdFbYtH$iAcO$vJ?y{15UFU$eKoN`z6yac!)Z;vdxLG|)Ys4)IbiB&g0=
zPyI5wu#61FwlP)~g}sqH3by?|U3uz!%7eyi*X)&N1weX7_|cFLvx4)qz22ltzd?&(
zR{8!NOg1j!z+0l1eHwIjl}@tBq8)C9r7f3!&VaRXng$RL8;!Z+_qPV>Sv*eGeOqTK
zJZDYv=hgus)e5GK0GZAxQLnGlih;*AozpB)&B_;v?V0L
z%<{f{`L&_=zx7Sts(eQ*vtViZ9%zI+%@1%pi?xpCF>E*e?KV^lYM!5%Kiu!dqpure
zOvb@DuYrJa?!_O-BpirZ$3&2S9kc^pCaBbxsC0$cSaHG^Zfg5Ba|hKGBS=y%
z);-5eTbRlY|L(e&%oln@X33o{1f2T-;Ov>wjem*ASmNcC-Km!G3GoZ|;j|`R8~5tf
zTKjuzSuOHbQDz-UHxvP8VWHiX<0sW;lQ&`G(SVK+Jhguj^u`;Sek9CouK=LiGmp?f
z{v<<|uq8jY23jolaSQ)%ONFaLL}75P^-`km4hj0
zLX#u?&YmrEwmoQsSC>A6uz}`X%xgr@u49aaI*pQUBgL*tQYmuA;Fs)+%xcv4oh-O-
z7z0v6?9JU@u4z0v_;+>wW-mTUnA_Niic^_D!k4IKHfkThVUn5kHbFu&Fp2}M
zLd#gHppxf9olDL;7a0dDUvceFRPLjV$J+B9u<0`Cl+W^eKj~r?
zR1GwzHTI-L)o+?WS@B9@nJp;$FPl34uOVy0wAl5Und!ac
zU-05EcguY7+@VbwVl0|76ST;;a_skU4@b&8$mRC!NHz54(tRtT*2Y+6mqkUjKY@RD
zV^}+)L&6_ndj$AyEbL)iKKNMsu%v_mX$(Z=V-*a$cP1fdpzjy6_~|;X4#grbdkViD
zF1OzHS{HMe6grQ6Y)1!vb$;-3sG7}s@cW}QR!&i)@g^Q;CBkC!O&^8S({Bl
zlYu5#m#aoAmX+1c4zOi^^`g?~sewn4ytj44Aac?`3IBVAdrK4H4GE6;NVz)LSnd=la+4YDUb>VXaaAMXvo98
z;F8e>YOkHR{^p;XgKf{^5@K)0sD6z!l~2ZxeEoWJs;s7nPwOMSLF7lk&$Ms-ZfzV~UbB%&HUV@jEx~
z(r_ki(K3vB)gf$qu$oCVUQVqwoBjk1U6w_Yoi&4D6KW9M9(RTM9OKtX)@B@auRw9)
zzox{gc6U3J4-AzO99C&;yF2(|iGvJ_rMsr-aTH*z$!5tXP=1rH)UoyYH)%B%OrdDS
z_EE;#a6~#H@UDnVC2LT4fC-S&jseE(
zbrkBTd(8@PgV(us&%L$kVZ1F12)IGPrj_+UxvJ5Omu3{IAQ2REwk5{KpdhqSS_zTO
z@icJh1BdT8QNTo@wb$JcDM*k}p`rqr12}K@VCJAd02~Fekf-qx>f)vn=a-$2_g^(R
zCVRg2Myo5l{?GJ)wBsTB9fJB$qP-Z!`Kz9`_=*x{zqDhf+YUiP9?PE#P-p!j3Nf_a
z&AruvhARZTr>LVeF%ZgSSlMUxBKIh?!hloi_?xc;0vPE*%E7RFLAI*gG|Z4p&*8(d
zSQi%LlN^vXq1Xl*P<4gg3=&i|5zoXNKxxMB6^DqOR~PkhMp5j26K=)ilzm+Nh7xwu
zg+*5k%iGJ=#D72zHv`NoPf5mN
zu#>Gz-8%?H@u4sgTXHrHx(gJDXIpvp#Mg`w>jPir;rt=kaLACoVw>DFvXAPu__y%@
z$X(8`v|h90HOs8_mdVK!NNYDGIvyVfiH@eE5)og3C##Fda*Uhe~cc$o4J#
zABA;fku1U5Jm_dbQljfOv}{IF!XMTTs{LAtp{e1zGd6}Q6-srWTBgZq$I_UKGRT(A
z^5xd4rH^Vyam88it6zHaX)-9VhdEML8_q=bq#&q7%Q*TI3yZR+wt
zM-iDciFY7bW(ZJ2g}YAKl+H94IYVDDFUUgWaR0;j>kFJr`G0aL3+L6La
z+|DLQT&~yOEBtZD4%p(N)nw9@<3&fyY7S-
zkfI9#@}Q%ZMVU0ok`+f%K&;!p&M*U2m@Ki2Q`99s(S9Y)-mkZuQ}iXCqnN}cXmq@>w9zJS$Qwcv)3(e~deNFSZaaN&@Y*+EEF!-x{gb7d@P6f@
z&~gv1=Hn`)4-n#gPp4NqzVA(awvxqad;fKoIm^}Dx*Yl)QLzmP3Mebo*P?;wc!S2$0@1_*Z1
z>;#7l?%9v?i&-TsNMM0x3x~s
z+xc&h{pDUsf)KBDZi|kd*>x~)bSC%*9GpIox@(JOQNedHGkYXtaY>4QO1+5zZNpGU
zwAU`epR{6uFhfGiMu2({6STV9i2FNl7P48}R<{WAuIy&AIOOPim)7AXbF(z;{bct^
z5kI1RU3;43!9a|Y>AR*X?B7V4mIcmhi&oiVdBbr(-o_=C*Pes38ugAIk3v&nV
zNw}ValxF8iRTsUDK@vY3DyjzRbeD8Y4y}-4w>%t6(~5Z_~Wm>H#3~sFDCxE6{8HY
zco?GD801qhW%2xTLDB4{*XvS|u-kfWxBIXb3WgEY`%N2J
z_s8@V=%^Cpus=;f{9XNe*}LFZDpomB
zKxe(i#B|EKbtW7JKt^GOroAuHrGi7FHlT;MWnp?
z9gNB2HF;c5spG!FyROvEl7IupsO$32rdbGY))-h9TmD6mq*QK);M&mFqA5L~@dODE
zT?Y+&u~0!`6svnd
zni@QY&W;xc-wEPlP1k#1j#}u+6n&VZ!EgH=jzg>69^@HBY_?f2L5Q5!F`OnSjWdX<
zBY)KpFMV~8qD-b{-bfU0`z!;DE&+P`=8?jaCi-_~vjcq_kiaPz{z^|kyIhiHA!D?<
zpW0;~#XsjdyHBk_i+@@++eBndKVDuHMaF(eK&(Qv3%71MIep$H+Ukk`d*y#S)CL3@
zVAamFF;$1GD)N6<^=TKyt|E+OF~*_Vmkwh_9aS(Tf=P-DD7z?OIwl4xk9wBpb>`$z
zkFLXc`AAcUSllT7rfRT=>e6cv$NqJM5>WMZXq2FAJIR?8YLB;$WG^Hb7SrY{eS}-%
z(tDUMbzgSdbda0$0pn*wR7^DW=D)TPfJX1mGWGlI{CrNFxx$KaBGL
zR3~O|si(O_at8dIF
zH@FtFkYg6x(oDH`4id+>`dn2wIJu#-YV6+2MWmS=
z|8_S^n8dX;ClI4&>-(nhFm`cxd2Q{NO+8uca?8lQQIvWOPPw<~l1Atu!#cMHLU_b?
zp}{rEeZi9D@-Ym&+xY}HJJ=*CebIZ9<$08uKj?qlUWlNUioUArc25PS$k6xwXu{Sb
z)X~%%)xz=)ard)jg8?nR@a>Pw*bpnUw!9BsakXJ9}Ldiw+Wo!n?P_JITn*@T3s(c1bBN>DYmTJdrTnIFb`ccJ@#
zP`m1*@MJyq>Om`Hi2_^DhniAcTKTPQ%ALZ^%hcp-tdtcOaO8Br=B@PEe!P3X!K&cC
z$K3OEMUHvL)V2E^I`r&DT*S)3ac(cy!6X5VjM@Y-M3wyi*7P4Ci?+PY*+1(bM|Rdk
zcWmg8}%qCz9zHfk3R^dzwOlQPBGKWsDsl{`mvC(8;<
zF`^zzr)yw4BlTy6z)$=nCe(3DT#HUVPjahN3#zLKJbhKpwQ1ES-t_pohql-uUT
zNc>stXi9u@>C4KTx0M{`&OsSHd)T3@Hg+pGjBkR1_WO{W!oSw;t)^KF3@UFe`RSlW
z|L)O;09;1QGX%ofMoatq$E40(uORSaA6(!wdeF4%AZnSWQhMTh1&^1PEx>m76_)#5
z`8XHC-^+I5DqmBy$=Ex?pqqIBqh;D?hpm~(ErV+%OY@ja#L(NP($^Am+_5!Xk89RF
z148HcKs+zp^w;hnM#tfc<}ESOs*FK040Ad6sW*#D)*>o2p>5MyJ}<_Jv;{DO6*q)?
zjv3LmYAa5+(7+;nG1_swG4%Tqi(T`)J$0%osL|-)9$;=U(`vlHSn>=L)vB{4VwOV=
z-1>$yYN?0f8VBpqV;Q_2OZ2)ucz=O0j(v*BhF8FN~X1YB?+~7PvY0FB2G$9aV)U&8Y=HU8-8zsdCN>rP%D|0Y%>DF)F!r+&fg-$yo
zo5~$d&U!ubgNt7Q)8_SfjT{MChlk9`v#|60Z4Xc7%o9^Eeg3Qw%|eZ*_3Xxj2++o_?)b)Qi&y~7&TvFxU(<_HQ{n?n1F>Pkb@Mcn<*z1_RMj8Xk)o|j
z7>F5HQR&{?t?xE)^}u1u;OFL1_d7Vyh6kxml_FYlxl_YFpDY^jRTK<7iM7MyTFBg1
z{*;^oNvB+BJ6$9anE$UCR*J@|e8dwO*Ll(oq2Wtq-hks;QQ>-bdAYT|)%kY#0Bw^D
z)pqR4ixJrA5XnnjjdfID{*Yr@gD2B3pNF6ceB9ZnTQrfWyg09PoeqrGiBht5r|$j1
zn$ZOdJ!qcei=Iz_I<-$xtR3@Q=E)?E5{~Oo&VLcP6i0c0p2|9L8G~Otzq*bOyZ2Bg
zr;6BtCt|;pSqyxF=41f{WIW4?M*MXPAG)a5x4iMlR#A`e?&p?|KLn94D83O1xy@__
z(&Ze@1H2N!VF@Q(NuBT<;c1SYhq>#OSrPOWRZa#EXwW=Ryj)J6qqcW(D?M;8?SW!Z_=0*bN2%ln_|%*E>I0
zgKpbRA?Zq1$x|D@%L9gw4Lb8#Fh&d^)h;@-mONK=84Zoz<%kaBB&ilLVrN{{e)d3v|n3P?DT`
zr)O70Q)sz6NOnsRt-jp5BdWwh6;~qcUPKHiOTV&J<76|a!e0_yNNS8ea6xbBlFObh
z=c;Lp_JxTUj_ZY#Ne$r92=;xowD`2*U{{s}-Cy#v`F8(lwzblBi{(|@Nq#_0jHQpy@;p8-@kI{{%B!zW8xhUuRl?$6m?@=0vJYv#7#Kl*l8E>;Kr%-
zOnd&ZmEhkF`B9N0z_nnF`ycZXijsEsPMxHyc#55SwVS8qoGhDB_84|MBrtwQFnil&
zO^02=Eqp^OzhkQ%$vcw#>b423d)DU~PiSV@9GBpA0!bws@a2HrKDU#b2htY
z@7GfaUFEwc#f
z=lF~IRa+9?=j8tQ-S-A|Y3KH}R*yANv>a{u%Fer;!?s@bu5Oj0Ig{js$?&Ji*$6K7
zbu;KNz~MDa?+LZvD9S9zIPb3g_
z;#)OG-|S*12RARKd5TdovZ`(0f_v1TV07Xmy;Ng@)vD*bZ-@x%+xQpR<`X3WK@qor
zr6TUZB~Son1!bE3YN%`)gvn8QO!JO%anw615OAvxIhZ$A7kAZ*1Boz6k;k)#5D=TpX`VN`?e_YBB`B+dJj#WWKMlawkqOeL2?Dp&^~cG6gNWZ`1U-Y5K
z#h3=F8KgOPDKa{7>e&QSl@A}M73zjcAt)4x7ab8Z<+MwIYw}*Bo-R9_07DMudZo+c83PYd12nY8AB)dmk}g))
zUUl^>SCy@L`7&uuQHAzjdoDV$1W0MrHgMp)m82`h>hi*eS4c~ww`WKtsW3qxYsHIr
z(A8p=s)asus#Zg5wzs+xOw!7
zPUi$J1EzRAK-<^YJQ6zn<*K)DB~nzOlNJD;Ibl)VO@(^dQy(f5?H%KoM|Edk2d@hb
zT|ba(#uuHP$KG*5SN+0TaC06#01#4x3+5b>?zdr|ZSNmpz{~24aGy%P=Ce50
z;J>aD&8l;!feR+e)t}UeH5!95k+and2x$t4AB1)EjGyEk?*(VHj+Ym1THi!kZeO83
zSOc}KfCfdWKK9tBc0{FAY|GY$=?@y<*s$PK
zu-tu18|_L^ggoeK;b%33fjV01b4E1Bd)_0B;rr-?#AQYHZ(E9fYBH1!)uRax9%_Oz
z$!zL*)#E!`0dG@P4OG1;xUZ*8LL4|`0}tdwydL)e=!dTTrN8M3AM<*n>1@mSafVKc
zfpIw%^Rc{+H*Z;y04+X@xX-*Nq*I!As5}+UKz!&W?n4&5m_RuPEg8`3HXGH0_#M!;
z%$&^DITCct=ncXq$nJ>kXU+pn;fg+QV`(%S;GroO!o+qLzpm&Bu%lb&&FtF!`gF@r
zlTqKVR+tlV^X;<*4dVCk`Sl3Qr-vpy){`RQANISK7V<@D%zxm@AzhBQPml@ER3ivk
zR8SVCFHHBp65;49O#iBZ{|_Z*VPgJ5iJ77Pqy7D*A($v+gpn#|FDLY`d|~e;8)_;8
z7$R)yb}lF^Dhty;_x;NfvoQTfBRo;+=NC)N!u;zVu@M4gunlJJWS%we=OGj
zTFf#3T>DpMoG39MMhNczlnhzeS^p~<&i*B^_{B5--&3s@w8xz{hLAV*Xw9M3#fg}D
zw&Symd7x2cQ^qyWH3qe);L@0o-NPq0F6Kl2H&lQ@KP#BYtKW17i2
zGHK(<1r>-9qD6wrHHO`w8dGU1f*fnOc}a{$ogB@SjEL|3X__s4a5OF-mzJ^A#t6uF
zPtFz1JNm0SW9d>Enq(lH@hUoah~L;|!A9%2)>@Z(ono80%^
zs_=sejG;&m*Wa@fQbaK*B_0BW$qC#Li{7S7fX(kqTyoubYECUQLX`%QJOM}IB$oAq
zF;!6OvqV#Z6nFlJ7v}~h~p`8%>nK)01-eZRKi_VVT
z6Xml>OXruo-UhII>pU2s+^Nnxs@_zRWJ69CME0>!#j|y8+q5CIczgkqI9QCzSH-D-aGsuh(OCB@W$kIwbxcGrp_&LNM-@YdF=J7}|
zF3uv^TW4?-uuA~w65v$|b;H(AR2vf{LfSt`(L!{nEW0yxT(oSr<@Lu0nPT|oiBb0Pq*Cgv;%Mdt1R?xkg4?qwl!`#
zwnZ92%+TY{F7P(-XtDB;t0Adm(*|yB06`5sB3z!e6hHQ{G#f6_9RTs|Cz>KT_(32V
zByjV4Z7No7=-x%RP9r6o~v@=4;0UdGVS^0Gz`IQuw7PI%G{sC%z?J17T;FN
z=$wtp?_^CEfe=rr`r*WB@Rdv;F7#PCaV`{o*}BZTT2j*!-wP~nuB*Ay`rw+Pw!8G7
zNP-h?wl_X|kof4X3i`d{mq@U#q9;J%nt<2RX@TZHinf0W1yXDT1=R^i)OUaI*uGyf%U_BcxYmZ3CUBwQbj+mSPbRB
zxxf-wj!E!^2R}pdW)mGKgFtw7(R9G}MTa2=!{4(*TvLT0jWNtxJ
znAker>+zWAq=}iLV^6vJqGR@{Ss!F`-V=JHeX=98+RgHw1(|3oh12xEa}6Nu_F9zD
zl{lqPyc4W5NbIxha}Wk^2A|E)DcRDPj8k#jJ{{IwY`?!2}3^?RTwC4+UG(l+vR
zbg=5R-t`y{+Xm@s8J4k-KnL*RW#teVzV~8T|1pY
zbVW$}0Gni#(1~n`58VnAzN*LC`e+{zV^s&lYU7wf`+mK}n6q6<91;ionUr2X39&cz
zh0-gZ)IZ5hY;HVTkG5p;XC@0HOJ9Dve&WquZ?uT(g0t0?*_Ep&R15=F_K&+-iYyzt
z9j+cEGl3TGJF*`FHJ8Bu$J93l=Mr??#44V#pL6iZ7@z7EyeX}1mqpV%^AcHLQ^D7DMQ
zkJ@b9yjiCCv2AE)DSH5&8}LUnI44ulMvNjB<6Wp7950F*`78K33Bb=vos*P}*w|)_
zWc6yZj-i%g)1qd)nN^9k4N;MdPdXyL5;_-4^G)puqUnGI6FGoVDUE=noSz>O0ScWS
zMbk3i019k4tN^^k#50k~<(bip$3%MAm^#?qSZxFibnSc9l*MRq{N(=b&_>F%@IgdQ
zW`rniiS;Dl>llF#56DOFCGRIzYjdqb_GJPVFHORT!16@=y+@_&Vpy+gsCjUSrq(t!
zxbt_s1v@vx6p3Xr%mi1SV3%0ngK8bmI6v4qorbWQk50xu1Lqk5%rzM_V*vlqnwfcs
zZuB#gtV4+KCUTJfY%2q+H=HEUl1OWu5O0EWp1dlWG5{`G9WeO}e9e@W2RhLYmo#!-
zFa)9>CT^Q212U+J6-%Noc?7O~duIuj_oJh8ZQ(`@VHd=)&KtJiBUivBa{kRU4K
z>JS)XdSziJc9K&^UpsBg1O&1%e7Ds?t&Uh*OZKKDSs&!#XwrJ@xoJ~&R{Z3t@-yhF
zN;S&ZoUyp;0iSQDDjYkzYi>^^re9Iuo3Qt^n6|2uZ+Fn(*}~1_8GZy(_!!@;P*|qp
zED_G_U+A+iNNG>hN(6I|rU>ccM+!dRb-V4dL75nKLU5jnNykuXRn%W1
zac>+cUr3Z+zX~2P54TtTzTC}JuU_0pxRZ;Z{WV|lmnheq2(P}^t)<+s)P{F&Gs-pG
z(YP>SEYJNq+?4sJ{rXwkCEc4Z4iPT1hwQ^Wvm
zy0Dxvdb*j20rLMVqhtB6rp^Bs`?9nC#J)_Z?EhEN=KodEv9PBTP3StL_ep`mrq9Vy
z!TwJOoJ^R7M)7~yzXb&T6QllDP>1Dz`u{g5X8#`^@MPtz{{+Px|JBOL$`S=;PiKZl
zWC}*e4oAoa{*90g(x3WMp%)kfPYDOZ$4A5{ZO2T+!p{C1gcR
z!zdKg7WaP}mj$5j#6Jz?HdX3$q#;l#0P9bY-v1vs&**U63T8Z)ng(#)$jN2O(8!^N;a*S7Ei5eZGIM#q
z(!MRebOpxx^tF#am5$S=KFU8x1V28irWNZR7C(L5SnwYw^|d?zUP~A*AF3yItaI;P
zsdm2_jBB>8)D1ku1=-p^i_TBQyR?q%dZ@CR^p3Wf(DXZy^6DSFn5XaUR>sSlZnJt*
z6w&KOlp@w%Vh!3Xqs4XM-@L?V--||$*R|y?qpB*2eGD<(meS@$bOYRAWz?q!c$5?#
zys5(ccMv2%8KpJ=tusaJItGUV3X^qeH^?JX3Bj0>&vvRQ8BupIEp3*a*2wLzcA7>a
z@(BY}EDouuY@Cw0$X3RG{HnL)4O4)fdz`kiU}(lHqY0YY
z{Is9UXmKv1QK}&(yS0|1C#lw@cTyhocd_BSv%Hk)vlLnY^_n6}T1jp0Xk+|E5ny|!
zDZRB(g(FQu!QFKZHd~ztHD*zEODV!YD*cmS&?+Kiy_N;0#$r9@`J$jjMKorrEnh&;<+
zF{G7Y45{A$kjXQsB_lB`%L2oC4B6H97|M?Sz`s7~j!==ZJ{RX?iny?l;!^4^SV>aF
z`H0=KaIuDFjktjqR(JHd>Hgx3z@()r@i|&
zzTUUVZB9PMtoK3s!u&!usGU}reFX4MPHh0<`_v5pvzU(*f;@AG@o{&6w}t8tz~lO>
zBL3-m=-15iZL{i|(#c1DNKFHjC2LH;=fnzQo^qnw(q{Hv)xC=_y%vS2z-3+&D5P8((s+OJBGW+*{J3FUvubo+nQ^b
z7hijTY}k-g%W^F84eF8!g+ftE`gXHJ`?xHYVFZ19en}uC1J|?JB@<}SM*;z@;h&PZ
zBosTe#Vs^$;yBh@`4i*;Stj?>W#g+^-S*iI&yn%M-?iWRhQD%J
z*79*@KK{pO!dLUhv(dl{{LW_wHtz(zqky2$GweW2fjr{Sl(phy3?J9xOi<4MOhiZ7
zbOmzWqry137Nam395}Z#-N?iXppo`rUG6BCa|vwhDnjy6@pSvcX*cK*Cq%SVVYt*A
z-DM%sSSwH<`Cd@kYEEYceF3f;(x~xT)
zJp&LH57%?mFvs(}LdXn;ypjlaLk4_-Rp8^|WW5CAYJ?p>R{{Tw>Uhj~ctSxRzs}jd
z?!S)S9i3c;1~vI7l6~eE+~5Y&IE%1xyLYFf|D6jhOV9>wI|&@KuGT40QN`O18C{y(p+`@)1@{IR#%7;r@s>X<`L>L6@?{q_+jSiNn`F(2Z1lSD
z&HlJ1(|_6+3__-v&R4tV0RM2&Lw!5KsV<{4;Ua9es)@3(9cU8qEHApj-8k4q;sE7g
zbNrobn_MLjUa+ipOkvJdT34mdi!o>eYL0k#zGs6w-F5A5(r4a-U-0Wl%lt1LxP
zlB(<@McJ^EN_^{Tty5C{vei%}P#Q)!7A;UjxHCo!9Tv#qW2=)UY9R_p#VqF@yw&4g68Oi~Lmo%uND_$P5}3H10xjMm--H9uugpb;c(w&=Uf
zpnQ{~G7*_*8jO4H7`t3E?`T>hVFUacwy{00
z4n5po(tbh(0?%3;6QeFEPP9IX{eeEhds7mi$)o{tu%-oC8;0iAKW@?!Bx~1aDuKOT
z756ZamXB1AcIk$t{YNq*8Azj3Uu~keN>UYGoIX{d+PgpkK#WtRx_n`ya`&-xfCfbB
z26Twaisw8xf4`~oEG0|#J0t=sweEsBw77rd7>^v2kO>{L_~m|EB{=AH*zt=n#SzW|
zp#;9bU2g0hociwdK|N$Spwf?8tk3yu!4M(-Mo{c`lzy}1p3>?_71;+%IqfFXD;L~
zc#7F*%GI#J%2n|r#Uyc)v__c|$ub*M9IV~8euV6F0hbhm{vEqV^BS~G{AnflI|p5L
znXw*u0JQlpP_cs;Q6Ow9ZE{G$yKf789H<3COlE?V!khz(LD6)jEfO+
ziq7_lct%j}KJL6MMB$7ruA5Qfh#fBrM0s)rKuxP+va#F_bufqZwiWb|k9Y?0
zE5v==tl{(bKlL}pV0b4-EHoL8iR%uu32{F1!W_H=81d?_>I>2}pq*5eVZs?1`N}Oi
z4L_CRkk2lksJg7#l_%iYW#yu2al3U1j!>8HEZdYc+$`&-C)>2TRxe`zRt`jX{56Fj
zzyP_VGDr&evZ6O^!{HH9Lk#OROaZoJW$P63GgXzyPK25qLY}Wdmbw(9o@S6)y+mXZ
zX0do?+XB=_zidZxQa0S)h+G#4ovUP_a+yB;WYJA_f|bPz$a&5(0^JgaOiB+H{bWn*
zs2py03NN)3p<0-78UJFDd9r$}3xyRQ;O;HacbIQK6DJ+S+CYbx&`BOVtp|tvrway)
z-Q6NBeLcrbpa}L>)L*%-3#Z(AAV)ooXVC@D84^6n3BGOWOngb~|2};-J-b=~?fhGA
z)3lv=X*M7|llO#_-6qH?>;Q9-Jhv#TQ(uZShF~WmFfuc|35nWCd)yMBTxQ2MMLoeE
zP1R{!G_ATU^vkS_z6BWaaK6jI5b537B=7(iPAalmMB*{+h{diL{Mhp(gP2_un!{`Z
zeYPNnwLz|{5=OdZlWS>qo+XuHO(K%@np$+(l23F$T#i@NW#|)Mf*ev1#GR=zi0JnZV57uo;4avOfi-XF6EnUicz?l&4{QdM
zO4RHvG&#qVDlPk0Jo5^6i^pE`uU3mV99ylkXLqXQH&1uy8JWDx=DQS!1tdF8Ygss=
zq&+rr^3o8>kfA==JVyCDbzU`y`
zugMHpG=*2h7|id<7HS+soGSttF&18VR=pwF3{ccA2KMubp?xF0eax7~U$AnK5>r-+
zzqC1fPo+O{#_+YZi|I*5@h!LU>~^93ef=Q^TlYe8iNoNw$=WbQjo*
z!M74quR-+v@vF)JHFxQ?X*e_Kxa5QQ5|8^boooZq;xg{OI35?ooY)fWze>aj?U-d-
zhGU%|UBi~bk#CUnR~M%(+lQ!tJp$6>VA?^uQzV{{+T*Dd!UA}iWcgJc2UdO|XR5Cj
z8P3vtm|f07{8RcFOh5i!X|Y&zR7s3E)lj=6TB>mn@313)l0J_42CGuLxVmB>*Wqud
zYrE*Q74X)+QW)%#NTPsfxR~B!xU$4(RWZrdfD^D;f=`A0&zu*JF%CO`+kF*_vYwCa
zP7uXLg1ARAQMs%24h9r88PW~Nh-#^M`0Gf1JG9KC;SZH3JFf6lJ_jkg~Okqs&~-~
zO=ImVjrA$;#15DrqG}O
z2T;Aj-5g_Nxw~wYS=}A**e|ze<>pL@(YS(5GW>f4qfRH{Py-D^aAB6onUk@-lidXe
zaSIQV!6oYywXh5o742a_HS5Gvpc>q))TADevP;3KWC$hObc<$*j
z#gKA<`!T^qx0!eVQg|;e<{3571k8}|n(NT$`+^qQR?aF3(){F(B~GncK@JjbfXT~Z
zVOm^l23YGcnyDd~Gbj}tzAzl_y~x;NOe?cZcRMtAoyGx-*s7D(0N2LyY%fhnMTinu
zPzaf%&kCoKz(gPFasHw;2~v0hQU1oegVZOIKwdvIvwpl48{nYInt|jAJZ{oB*k+2?
zh$eCzDLA1ldVdQv_##?Vc~R6&z>jr0Di~!5*|6B~Epgd?Y;|2^Wog@0HV}O&<)>eM
zUhoN@HHtTbz>Xu9BDGZODNLOkf0fF%LPFw$CAmtk6XFEJu+XsA$`wzwY(i+cNB#(w
zhQMawR+nNCDEw5FlGx*&Wxa5Jz=aGMNPM7Be%ngLk|5%Pu8j@RjqwSH;bNH1+9H)H
z5r3V|1c#yqzFp3xkp$?1ifwwFX{Lt)|6L9if{1uLRBg9tE57n&T)vtIMl?SQKs*hw
zloTr`yh}+T(!JolnW!|U*ZTcH**l8kX58W
z7J1FQeS#Qlo}9Lp?pGOrAU`~$Fh$%QjxL{B>drqV5p6{U;-$3i>gy$NAYoO(L1afm
zM-(mj1%^B8&|{R2tyFG(gnvYU#|B9BInFl8)3(Yi!lHHUW_q*Y2lIq$ki-NZop=`S
zz4W#>omdp>%)se&IOU)3Yfd9`6bE<6LEM>Wp{gv6qqGowYeNPQqoqgv9Sn7gFkp-4
zJ$;uD&Sw&vS7yI(8ZAs{J!Gs3`f38k#g{UC;cm$q7uYA7X}q(CY?p*ImZmWc|g>UhT%^Fyf
z#-lg`E-bv&v6CIZm;Itbe<(b4turgfP!&Cx{S}lJ>l&NZoZFJK30J@5JiqIvgOJsy
zFtz7an%E~BERyVXP{}&Xyely$M0E`P@S05*;aztZSPh7Dyh|EY>
z%D^@6WImB-bYqik?eErQJe1GO23Rhc*<;LMWgJH~CUZ_D|E_-J>N+?y;e|ZoqA7lg
z1B2qKPk#wVBU%>M-=G~shi7_
z`GNfw1K|PsyZ#{Z(cxPDJNnQT6JdnPLxpf10QdZ7KW=QY=nsD>yb6@}0f)vX-PtdS
z;CK`mHVKl&s3bKevXSaMd@W;jSDg2=Od^M%L((Uhfr?JL00oB4y7NIpg)5%G;&FG9
z6h#WbcxG0V1g6W6Tb4m~xAwdI?3IbVwcK^2aZdpX?&QxgNLZFU6TBDLBaGQ(ZZHM*
zRYdd^Hr08dD{6Ctla|XN%$#dQGFzC>+f0xt8M%jX!7CVg8fT*Sei$#cAep2)k`2hDMTi#>O%IIdkeVH;>)|ehiSm37%9Z{NSK^I~g
z8HH9>fT^HWy%}6r*C_uaAA#?&1pueJW{os&!E%^p>cIFZ*Y`-I&R^H)4n+$%t`JT7
zz*`FJFME(Z^W1mHawTwsH?r@igeXFfW>uc~#zjn6ZY(kO!N?yl1Z5%=30OtPW205xyyI*>}V?#?{flwubYZg`w~B6`|pG
zfqUfAsn4Tc-5u9B_L$y|7l1AAAHP>!Ek&@kwKB2A-Jo3CbaJpon<9D5?Ob%K=Rx5`
zqN&~DK=Hi^)!NI2LQ*IXAFm(wOVbcR%thC~Hygx36Us%D3REEegBjb2>H(q4^{-*`w++~pvSs=4d6UHqApgfhTZL$VwaYe4nUUz-p1kjv}Z8w
zg#vCUA72j_x6k*~qMfC?&`pPkYS(Y+&fYXvJA(VLw)7_cy)cEFw-3SIueZnBzW0mG
zdVo6s|LQ%9v^<#8^Z3A61#90cR_@N@V|diO@@S5XlOpdMPIs&V7~+oxO3?3HdXZv!
zj8ng+YE^u`EF}64;`>CYhaNTR(plc`xEM6#VjV${8e0GB_EP4=YVyC6s2dmw5GBr8U7+Ub?1BapcJLzu}7ZOF?mA
z*0n`~3sVLd4!}|kj}Czlly+8EnxMT-7E1tK_ndK)AK}zuh_wk#u~E&(9$Z_@srB)Y
z?#JhoK_kW>*A5X0B23A_)(Ed4(w_^@4Eeph)7!7%dv$+rr&a+0-sr^iT$zF#vOZ2#
ztI&TO;HA<^w2ggtK#h;_+btvkL(;VU*TGU?{Q$|f6#7T
zd@uCy&41BRSsvWHGY%paPP9YDESzYFOiJ7-Zxx@u{gy?3-N0JGY&x1pidrx{#i@*T
zzSLu`cL7o+pyBg0y?s?ET30QX_z3hVMtV<}FIcEeTH=)^(cV4cl_$XrUw*D+M_~1W
zU&N*7@uoyN7xnffd6xeWpuGLKyEy6v%_5kJ7)1~mc}yIpT(3^DcO@NQ{R3>HzWDf{
zqM@;j%4L5?BI?Q_DQ|+<3sMW~!O9|)+r!JY0n|0&(@JO}_~_!7k8Ta(#*3TmdES`m
z)=dh4(}p3kLJ1k!T;v2};MEpcbN=Ac4OCmB5suFxlxdO8_m3$>e$2-o{;C*$9O#{R
zT3PpKZ~-R+ksCfU3he6nYo(_$fEo=(zQFQ6)@ZRQH1TA%H;;r9di!>LptkUmP1$bu
z3h=M8+MR5PyXyQ6@x&5)$(4U(xrNUSk`c=2eSHLFA9SXit?T``l|Gl5Cp9pG&(^-%
z{qYw~4;N_9>a_iJ7#bh1TR^ub<+T|>etKO9zlJ0PUh8D$JaSj^13gl?Z6JS04Nf!kr;DNTf83BtM-qn&8x_AKA<<0xBz52M7wr
z02!<9TM2m;q94-x18e0?CCxaBcDp%nnxXgqW!j1Rd({q_Y{U}zJ7Y?6baT1dU$46-
zouoeuzPT2P$y8JpUdNt?AExWdyInBQV6vuQfQ7S1hNTqRk1!C&*{`;?goam?0>~1k}T7)7o6G
zFO=QjPVd3%0`N2LIpal7ht|;yym)qG5gk+&o{QISZB;b^1~t3=qjm1gG(%X~CQ&Yz
zF!c8QSm=54gAT6k!LT107#rAke;@6D!z@HxnN6Aw)1rH=f_DdTZE`_tlRkDH?cJl(h0^$$UC`*+gg`N
zI{{o?!+lZ4lOwAxE~g#I!Ha#G0>#Wp5;+T$1dqQ{zf@t;0SRhitc<`sypBAbvZuhE=a^^v^Y;zfNb_MfhRe
zwc$n7Q^dEUjb~K>k8ZsTqwM~*IByy+`b+5#wIYUW;F0v5;kSNEX!tru4yaoCn0Ef@
zUd5d>lZBNFB@7p3^pCv^(JGy-e(E)WtHhjM{20Ib>?MQEJh;cN02}6(@M)_(Oh70;
zCW=HNKuFrjZS?;-YlYQ7E~D9{oTjgltRC*6k;9}ZjRy%kh^M*jNgu!p(6){oK$^-f
z`eS4mQP0)ILsk;bBDgzcjop}2;DO8*(rfSZFv*V~#yhbtfMH4#t|%OcHkl|xfYqBB
zstV7=9yU<*f`G@x07OB2z+x}~lg9K7d2&2wW2RAq>ONHvWorXbo<-P<{3CH4Fb?_xmPADjMh8J#h&AIwWe00M%yV?iGp{mtJ%8`T9pN}V
zRfdF0`WFGGr(-3=;Mdoo;4}}NEt8Xp?(m=92SH=>DYwnv>oc1loSNKE9Blm-siNcR
z*@itceeInn0AAl`p6^}X=V#b+!{h3iXr|GN9)J5GM6)*U2*ZaJ`s^F11&+6iyo5gM
zTRlMlX?`jDxW?*j#BrG6&K<0Df5dmX;ivA`JFOgIMDSOA=zCjDBbJWuWmjQ!|B2V)
zIOgKnKYyBdy76D&P4!>4hm&jjyW~TrE``5)VDME703zQUsa6SyB`D_mlBt?0hHV?d
zTu$Cp%F}}-;e{I`(*0`w^8(6)312ra;ut)3+;Pt&8#W4quNz+PPD8tJiewbJc5n}v
zrRISSvLx=qEXLxeoXKzay4uJj=Z)}>S-!zX)r|kRsqX24$Cj4WaG>SD
zYF6z+1(Bg*Y|aR=Fj6&xI;VkBbh6n{T0>CO%@z^8aY#1`#k%R*(LvPeGIl1#EttlO
z4C{R)5iUSxE7GqMh()52WDsBjiIioXMKhXgB2wSE+925$N}`mzhHC~XLb~@ZC}^1*
z0RPzEG>Cg~WIS7{m{noN=*!%nhfWSp>|xU-s=ugqBQ3$opnG|-2V%LeapB7|_b)NE
zU2EAo&DpJ%*Pks~J3^L5De;Z^jMvu(rs-?xdN>4sON`8JG(BntGPH;8)!ciDZ{GGK
zB@(}x9?3}4ix7rDKBMii;b=IMe~~?30${rk3C@8FaO}8;GuRs6p{F&pK~$%f)AX6$
z?Mllk=yl07BK$00akZw*w^z(xaVuyJbk01O%;4F0#D>gYp{3I?jkVI+y73^;OkQ0i
zW0`}tPtUXrVUP`^NsOUvQel5v%~Cg_f^?>Yv0i^EpYx=Pgw71gb<)sL>i5vR0iydz
z1w4>*5*Wy=MyJT7XKz@-69JO)9dBnLSBZ6G@R(Ao7F?!nX7Lpu(&W!7EV|+3-d5Mk
zQd1f8qQkhZK+2vr;r((F2;Sk+*0AOY$z)84;!qr7<^$kzL@21>dVhY{E*mN7TYo_=
z?t$h&gaV~6O=>7dC6W&6p$y+30dgniPGL5BgXu0J3(IIdef(f~eLDyPpu-ObBiCJuXj7<(;{TDR
zp8FPaqF)xtmQi^Yxa6T92eedVu{SUX6(eEsq3K+@hX5$xAP~`W+=6);29)wsosoi$
z6&g{4IPhr-zbOBOg{aK`Xw`|+UqEeV5QWtyu&AC4(f%yba)Ti~<8Y26dC)^q{4U4|
zVa=b7u#A(M-BCkvQ>&#%A*yK%hWVkt9>D8Zk7?^ZTYMM7mSj7w0g3gVM@tX#WJ!~n
zM=fcC^lqwV^Xa
z#Gyjvfk?v!612v93tTE%X;LI6ibp~?a<}of-=*yg$*v;llju!=k`h!eCjwMx1B!D~
z0_u86S)SBt!0ay)pIp1d7NH^cbvZ+Hj)heyQSAonKX&u0WmU*D0gpxq!D!jQ?rBIX
ztH`*f5W#WA?}z{%IoBF1Gr@AzD-zgHvVRv$tE*^XD8eV>x@zJ3K*gO9C{W67@Fgf1
zE9@*K4Fx7N5rql=L`iDJ!i)*$Yo0RAShqoVEK3VP0PHHA2?H_HOfQ+|e5MNPKW?w>
zU2T#ip1rf9RgV}8fMP$9nDDEr*a`OoQexR)#l-O&9boce(n`}dqtUFj4f?3!LLy57
z0gluqe#IOzXDb_9mLF{N%Y45yF6@(;wn?l%C)x}TwkT8CFC}Lz7axv3=%ns4P`Qto
z_vm*t;G(Jdn&E2YT<$Nyuavr@JR0gF(MA^QBApvDlYa@Drb(cqZGVm|VE(}>kOF@z
z06|6t4UDf!xvRb)-clR1b)(!I<`rG7D86Xrx0Sa8NPpui?I7TG)ZOe>*j5Uuark?$
z{mS%2L9UFUmeNH`yL*?bY%-(OHG(s1tNDd056CGf@s5FpfxaAN({STjH#fJas*9Z_
z%;>it&munTC2Py9WH*4yqzh;+?tx?HO^G|?@cv8lQbyZ#PDi@;q2U=^l}MB16)U7U
ziJR$Akp0yhCiJJps+Yyg@FU2&vj%b*K`qVpo{&di?D5zfh3)C;aknu3u<6-Ep9!}4
z7yx+?UrVNiww){JLf&}6p6;TjnDDfm=NtDNQeo7X*Q<8#oOq>YQT&8LeyH}CTovv2
zT7lYjYwtb5E~>}N2(Ha2#Z!Y}Juy`wHGp2K>S#h*ir1yJV&_C6g>We~Y1CjHBS<3P
zgcTHzoH(#{K&2~aZWn4@PF8YEbq*Z&gWU!zL{NB6A;$MMW{_nVcVP~b^~1?2OmTI`Xy
zCxcjVu^I!tP+9DG=j5PTQQtGX$sWV$sd9s^kRInq$(C=B11fU@Z73}E7CR8m0U+j7
zWNE0_C=v(&9%77e;~p)!YCr>t{lYRh1BR}u!zu1Vt<8&M279lY4DR>i}{fs
zLQUc{(t>@WP>*V;+Oe+@StufbQNI)z7t&$3s#^xofrO}_=|H0rkV+z;exlcb)%m;=RZ5FDsk`E9yOr3P)
zH!I(X*Fm6TuV!Vwr~Y;Bvp>#&s$a>&FzkRL)tP8Hw9^Ybwuk94wHLoO2y8q-`7k8K
z|B1GgO9*{?kyI?4)?Yhq1!9k;gzw`!RzfaiT`S;Z-Vu^1B3};sZjVd
znzd~#wl-hJ(PIrNFVNMp=!p>svhvCZtNlp;N-`64tVxelMRdw-HCpX!99cz+zi(?C
zrKp_QUC)UciEZ!|XD2a#u*iK2X1=1aEZ62MGPjAghx&_2vh!K3z=O_Nhp4Hl5LiAu
zY@8!6eyu^oz%NI-x<`vY1wa%@ii!Vd5}I6)>+{W2(jv%cWNviY`z#}^yYMR`9SymH
zuOXPuO-#R1gg#z}$h&zixl7^RP(7j2^D}w!vS#5Xod`Op>14vhY2&8L;N`E8abl#H
zSztiqbh4yLsonsAN4H^)v+5H+s1o9epE8?LaM35)k{}J>7@>Q90n8QEtDxWCn~wr{
z70?SEWU`(igut|BmhXohR&38xI|r^078Q)Yi-`AfH%!|7i4`RcbuJEja9|l&!CVj9
zo8iZea%qR2e1n^+W_+cu4@Mzbp$Mq_ocdJi!L@Li-wM5Co0yC
zx0bR~v1-mQ8wB%jD*zduGg~&^+RheRTS?etl;A%E;~>8d1K7c~qFkT~H^aYg*&I%T
z%Q6Vi_;4_(krI){3C_O73x;@WTV^Ol