, where the audio name is without the extension and the normalized sentence contains the conversion from numbers to words. Take a look at the original transcript.csv and you'll understand it easily. Then, copy your audio samples inside the wavs folder. If you want to make the data generation process less painful, I suggest writing the transcript file first, then record the sentences using record.py.
+
+[pretrained model]: https://www.dropbox.com/s/1oyipstjxh2n5wo/LJ_logdir.tar?dl=0
+[this]: https://www.caito.de/2019/01/the-m-ailabs-speech-dataset/
+[here]:
-I implement yet another text-to-speech model, dc-tts, introduced in [Efficiently Trainable Text-to-Speech System Based on Deep Convolutional Networks with Guided Attention](https://arxiv.org/abs/1710.08969). My goal, however, is not just replicating the paper. Rather, I'd like to gain insights about various sound projects.
+## Training
-## Requirements
- * NumPy >= 1.11.1
- * TensorFlow >= 1.3 (Note that the API of `tf.contrib.layers.layer_norm` has changed since 1.3)
- * librosa
- * tqdm
- * matplotlib
- * scipy
+If you want to understand how the model works, you should read [this]. Otherwise, treat it as a black box and mechanically follow my steps.
-## Data
+1. Edit hyperparams.py and make sure that prepro is set to True. Also, edit the data path to match the correct location inside your local pc.
+2. Run prepo.py
+3. Run 'python train.py 1'. This is going to take a different amount of steps for each voice, but usually after 10k steps the result should already be decent.
+4. Run 'python train.py 2'. You have to train it at least 2k steps, otherwise the voice will not sound human.
-
-
-
-
+[this]: https://arxiv.org/abs/1710.08969
-I train English models and an Korean model on four different speech datasets. 1. [LJ Speech Dataset](https://keithito.com/LJ-Speech-Dataset/)
2. [Nick Offerman's Audiobooks](https://www.audible.com.au/search?searchNarrator=Nick+Offerman)
3. [Kate Winslet's Audiobook](https://www.audible.com.au/pd/Classics/Therese-Raquin-Audiobook/B00FF0SLW4/ref=a_search_c4_1_3_srTtl?qid=1516854754&sr=1-3)
4. [KSS Dataset](https://kaggle.com/bryanpark/korean-single-speaker-speech-dataset)
+## Testing
-LJ Speech Dataset is recently widely used as a benchmark dataset in the TTS task because it is publicly available, and it has 24 hours of reasonable quality samples.
-Nick's and Kate's audiobooks are additionally used to see if the model can learn even with less data, variable speech samples. They are 18 hours and 5 hours long, respectively. Finally, KSS Dataset is a Korean single speaker speech dataset that lasts more than 12 hours.
+Open harvard_sentences.txt and edit the lines as you desire. Then, run 'python synthesize.py'. If everything is correct, a 'samples' directory should appear.
+
+## Creating the android app
+
+As you can see, it's not very comfortable to generate the sentences. That's why I decided to make this process more user-friendly.
+The android app is basically just a wrapper that let you generate the audios, save them locally on the phone and share them.
+When you write something and press the play button in the app, the message is sent to the server.py, that launches synthesize.py and then sends the audio back to the android application.
+If you want to use the application outside your local network, make sure to set up the port forwarding, opening the access to the port written in the server.py. The default port is '1234'. You can change it if you want, but remember to change also the port in the MainActivity.java
+
+## Using the app
+
+Run 'python server.py' on your local pc. Then leave it on for as long as you need.
+
+I have to add images and/or gifs now
+
+
+
+
+## License
+
+Distributed under the MIT License. See `LICENSE` for more information.
-## Training
- * STEP 0. Download [LJ Speech Dataset](https://keithito.com/LJ-Speech-Dataset/) or prepare your own data.
- * STEP 1. Adjust hyper parameters in `hyperparams.py`. (If you want to do preprocessing, set prepro True`.
- * STEP 2. Run `python train.py 1` for training Text2Mel. (If you set prepro True, run python prepro.py first)
- * STEP 3. Run `python train.py 2` for training SSRN.
-You can do STEP 2 and 3 at the same time, if you have more than one gpu card.
+
+## Contact
-## Training Curves
+Your Name - [@twitter_handle](https://twitter.com/twitter_handle) - email
-
+Project Link: [https://github.com/github_username/repo_name](https://github.com/github_username/repo_name)
-## Attention Plot
-
-## Sample Synthesis
-I generate speech samples based on [Harvard Sentences](http://www.cs.columbia.edu/~hgs/audio/harvard.html) as the original paper does. It is already included in the repo.
- * Run `synthesize.py` and check the files in `samples`.
+
+## Acknowledgements
-## Generated Samples
+* []()
+* []()
+* []()
-| Dataset | Samples |
-| :----- |:-------------|
-| LJ | [50k](https://soundcloud.com/kyubyong-park/sets/dc_tts) [200k](https://soundcloud.com/kyubyong-park/sets/dc_tts_lj_200k) [310k](https://soundcloud.com/kyubyong-park/sets/dc_tts_lj_310k) [800k](https://soundcloud.com/kyubyong-park/sets/dc_tts_lj_800k)|
-| Nick | [40k](https://soundcloud.com/kyubyong-park/sets/dc_tts_nick_40k) [170k](https://soundcloud.com/kyubyong-park/sets/dc_tts_nick_170k) [300k](https://soundcloud.com/kyubyong-park/sets/dc_tts_nick_300k) [800k](https://soundcloud.com/kyubyong-park/sets/dc_tts_nick_800k)|
-| Kate| [40k](https://soundcloud.com/kyubyong-park/sets/dc_tts_kate_40k) [160k](https://soundcloud.com/kyubyong-park/sets/dc_tts_kate_160k) [300k](https://soundcloud.com/kyubyong-park/sets/dc_tts_kate_300k) [800k](https://soundcloud.com/kyubyong-park/sets/dc_tts_kate_800k) |
-| KSS| [400k](https://soundcloud.com/kyubyong-park/sets/dc_tts_ko_400k) |
-## Pretrained Model for LJ
-Download [this](https://www.dropbox.com/s/1oyipstjxh2n5wo/LJ_logdir.tar?dl=0).
-## Notes
- * The paper didn't mention normalization, but without normalization I couldn't get it to work. So I added layer normalization.
- * The paper fixed the learning rate to 0.001, but it didn't work for me. So I decayed it.
- * I tried to train Text2Mel and SSRN simultaneously, but it didn't work. I guess separating those two networks mitigates the burden of training.
- * The authors claimed that the model can be trained within a day, but unfortunately the luck was not mine. However obviously this is much fater than Tacotron as it uses only convolution layers.
- * Thanks to the guided attention, the attention plot looks monotonic almost from the beginning. I guess this seems to hold the aligment tight so it won't lose track.
- * The paper didn't mention dropouts. I applied them as I believe it helps for regularization.
- * Check also other TTS models such as [Tacotron](https://github.com/kyubyong/tacotron) and [Deep Voice 3](https://github.com/kyubyong/deepvoice3).
-
+
+
+[contributors-shield]: https://img.shields.io/github/contributors/github_username/repo.svg?style=for-the-badge
+[contributors-url]: https://github.com/github_username/repo/graphs/contributors
+[forks-shield]: https://img.shields.io/github/forks/github_username/repo.svg?style=for-the-badge
+[forks-url]: https://github.com/github_username/repo/network/members
+[stars-shield]: https://img.shields.io/github/stars/github_username/repo.svg?style=for-the-badge
+[stars-url]: https://github.com/github_username/repo/stargazers
+[issues-shield]: https://img.shields.io/github/issues/github_username/repo.svg?style=for-the-badge
+[issues-url]: https://github.com/github_username/repo/issues
+[license-shield]: https://img.shields.io/github/license/github_username/repo.svg?style=for-the-badge
+[license-url]: https://github.com/github_username/repo/blob/master/LICENSE.txt
+[linkedin-shield]: https://img.shields.io/badge/-LinkedIn-black.svg?style=for-the-badge&logo=linkedin&colorB=555
+[linkedin-url]: https://linkedin.com/in/github_username
From d14441a68bb00b2932dd4d121f5205d6408eacb1 Mon Sep 17 00:00:00 2001
From: Simone Sasso <65923639+simsax@users.noreply.github.com>
Date: Wed, 17 Mar 2021 02:07:20 +0100
Subject: [PATCH 02/29] Update readme.md
---
README.md | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/README.md b/README.md
index 11d626f..a4e40d4 100644
--- a/README.md
+++ b/README.md
@@ -10,12 +10,13 @@
Getting Started
+ Training
+ Testing
+ Creating the android app
Usage
- Roadmap
- Contributing
License
Contact
Acknowledgements
@@ -81,7 +82,7 @@ The android app is basically just a wrapper that let you generate the audios, sa
When you write something and press the play button in the app, the message is sent to the server.py, that launches synthesize.py and then sends the audio back to the android application.
If you want to use the application outside your local network, make sure to set up the port forwarding, opening the access to the port written in the server.py. The default port is '1234'. You can change it if you want, but remember to change also the port in the MainActivity.java
-## Using the app
+## Usage
Run 'python server.py' on your local pc. Then leave it on for as long as you need.
From ce94968894972e4ea229a20a82e249da91b6d87e Mon Sep 17 00:00:00 2001
From: Simone Sasso <65923639+simsax@users.noreply.github.com>
Date: Sun, 21 Mar 2021 23:53:21 +0100
Subject: [PATCH 03/29] Update README.md
---
README.md | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/README.md b/README.md
index a4e40d4..6a1f03e 100644
--- a/README.md
+++ b/README.md
@@ -10,7 +10,7 @@
Getting Started
Training
@@ -49,7 +49,7 @@ This is the most important part. If you want this to work, make sure you have th
* []() An android phone (?)
* []() Some time to lose
-### Data preparation (aka the boring part)
+### Data preparation
In order to clone your voice you need around 200 samples of your voice, each one between 2-10 seconds. This means that you can clone anyone's voice with only 15-20 minutes of audio, thanks to transfer learning.
1. First, you need to download the [pretrained model] if you want to make an english voice. Otherwise, find an online text-to-speech dataset of the desired language and train the model from scratch. For example, I made an italian version of my voice, starting from [this] dataset. [Here] you can download the italian pre-trained model I generated.
From 9a29912ef2f42576bd8e8481f290d6e104565781 Mon Sep 17 00:00:00 2001
From: Simone Sasso
Date: Mon, 22 Mar 2021 00:19:59 +0100
Subject: [PATCH 04/29] Uploading application backend
---
.gitignore | 1 +
data_load.py | 4 +--
hyperparams.py | 12 +++++----
main.py | 34 ++++++++++++++++++++++++
server.py | 70 ++++++++++++++++++++++++++++++++++++++++++++++++++
synthesize.py | 11 ++++----
train.py | 7 +++--
7 files changed, 123 insertions(+), 16 deletions(-)
create mode 100644 .gitignore
create mode 100644 main.py
create mode 100644 server.py
diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..40cbb46
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1 @@
+/logdir
\ No newline at end of file
diff --git a/data_load.py b/data_load.py
index 745ef98..cf79981 100644
--- a/data_load.py
+++ b/data_load.py
@@ -104,8 +104,8 @@ def get_batch():
if hp.prepro:
def _load_spectrograms(fpath):
fname = os.path.basename(fpath)
- mel = "mels/{}".format(fname.replace("wav", "npy"))
- mag = "mags/{}".format(fname.replace("wav", "npy"))
+ mel = "mels/{}".format(fname.decode("utf-8").replace("wav", "npy"))
+ mag = "mags/{}".format(fname.decode("utf-8").replace("wav", "npy"))
return fname, np.load(mel), np.load(mag)
fname, mel, mag = tf.py_func(_load_spectrograms, [fpath], [tf.string, tf.float32, tf.float32])
diff --git a/hyperparams.py b/hyperparams.py
index 7cbffa0..0cd708b 100644
--- a/hyperparams.py
+++ b/hyperparams.py
@@ -32,16 +32,18 @@ class Hyperparams:
attention_win_size = 3
# data
- data = "/data/private/voice/LJSpeech-1.0"
+ data = r"C:\Users\Sax\Desktop\robe\Progetti\TTS\modello inglese\TTS\LJSpeech-1.1"
# data = "/data/private/voice/kate"
test_data = 'harvard_sentences.txt'
vocab = "PE abcdefghijklmnopqrstuvwxyz'.?" # P: Padding, E: EOS.
- max_N = 180 # Maximum number of characters.
- max_T = 210 # Maximum number of mel frames.
+ #max_N = 180 # Maximum number of characters.
+ #max_T = 210 # Maximum number of mel frames.
+ max_N = 143
+ max_T = 120
# training scheme
lr = 0.001 # Initial learning rate.
logdir = "logdir/LJ01"
sampledir = 'samples'
- B = 32 # batch size
- num_iterations = 2000000
+ B = 16 # batch size
+ num_iterations = 1000
diff --git a/main.py b/main.py
new file mode 100644
index 0000000..107f67f
--- /dev/null
+++ b/main.py
@@ -0,0 +1,34 @@
+# C:\Users\Sax\Desktop\RETI NEURALI E ALG GENETICI\Cazzeggio\TTS\dc_tts-master
+from hyperparams import Hyperparams as hp
+import numpy as np
+import tensorflow as tf
+from train import Graph
+from utils import *
+from data_load import load_data
+from scipy.io.wavfile import write
+from tqdm import tqdm
+
+
+if __name__ == '__main__':
+ # Load data
+ L = load_data("synthesize")
+
+ # Load graph
+ g = Graph(mode="synthesize"); print("Graph loaded")
+
+ with tf.Session() as sess:
+ sess.run(tf.global_variables_initializer())
+
+ # Restore parameters
+ var_list = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES, 'Text2Mel')
+ saver1 = tf.train.Saver(var_list=var_list)
+ saver1.restore(sess, tf.train.latest_checkpoint(hp.logdir + "-1"))
+ print("Text2Mel Restored!")
+ #print(model1.summary())
+
+ var_list = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES, 'SSRN') + tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, 'gs')
+ saver2 = tf.train.Saver(var_list=var_list)
+ saver2.restore(sess, tf.train.latest_checkpoint(hp.logdir + "-2"))
+ print("SSRN Restored!")
+
+ # intuisco che non devo più usare i saver objects, ma qualcos'altro
\ No newline at end of file
diff --git a/server.py b/server.py
new file mode 100644
index 0000000..ad6014d
--- /dev/null
+++ b/server.py
@@ -0,0 +1,70 @@
+import socket
+import os
+from playsound import playsound
+from pydub import AudioSegment
+
+
+def sendToClient(msg):
+ msg = msg.decode('utf-8')
+ lang = msg[:3] # ITA or ENG
+ msg = msg[3:] # actual message
+ words = msg.split(" ")
+ if len(words) > 18:
+ sentences = []
+ sentence = ""
+ for i in range(len(words)):
+ sentence += words[i] + " "
+ if i%12 == 0 and i != 0:
+ sentences.append(sentence)
+ sentence = ""
+ elif i == len(words)-1:
+ sentences.append(sentence)
+
+ with open('harvard_sentences.txt','w') as f:
+ first = True
+ i = 1
+ for sentence in sentences:
+ if first:
+ f.write("first line\n1. "+str(sentence)+"\n")
+ first = False
+ else:
+ f.write(f"{i}. {str(sentence)}\n")
+ i += 1
+ num_sentences = len(sentences)
+ else:
+ with open('harvard_sentences.txt','w') as f:
+ f.write("first line\n1. "+str(msg)+"\n")
+ num_sentences = 1
+
+ if (lang == 'ITA'):
+ os.system('python synthesize.py ITA')
+ else:
+ os.system('python synthesize.py ENG')
+
+ sounds = 0
+ for i in range(0, num_sentences):
+ sounds += AudioSegment.from_wav(f"samples/{i+1}.wav")
+ # increase volume by 10dB
+ sounds += 10
+ sounds.export("backup/final.wav", format="wav")
+ f.close()
+ with open('backup/final.wav', 'rb') as f:
+ audiob = f.read()
+ clientsocket.send(audiob)
+ clientsocket.close()
+ f.close()
+
+
+if __name__ == '__main__':
+ s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+ s.bind(("0.0.0.0", 1234))
+ s.listen(5)
+
+ while True:
+ print("Waiting for connection...")
+ clientsocket, address = s.accept()
+ print(f"Connection from {address} has been established")
+ msg = clientsocket.recv(2048)
+ print(msg)
+ sendToClient(msg)
+
diff --git a/synthesize.py b/synthesize.py
index 6043ccd..78c7fde 100644
--- a/synthesize.py
+++ b/synthesize.py
@@ -8,7 +8,7 @@
from __future__ import print_function
import os
-
+import sys
from hyperparams import Hyperparams as hp
import numpy as np
import tensorflow as tf
@@ -18,7 +18,7 @@
from scipy.io.wavfile import write
from tqdm import tqdm
-def synthesize():
+def synthesize(lang):
# Load data
L = load_data("synthesize")
@@ -31,13 +31,13 @@ def synthesize():
# Restore parameters
var_list = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES, 'Text2Mel')
saver1 = tf.train.Saver(var_list=var_list)
- saver1.restore(sess, tf.train.latest_checkpoint(hp.logdir + "-1"))
+ saver1.restore(sess, tf.train.latest_checkpoint(hp.logdir + f"-1{lang}"))
print("Text2Mel Restored!")
var_list = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES, 'SSRN') + \
tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, 'gs')
saver2 = tf.train.Saver(var_list=var_list)
- saver2.restore(sess, tf.train.latest_checkpoint(hp.logdir + "-2"))
+ saver2.restore(sess, tf.train.latest_checkpoint(hp.logdir + f"-2{lang}"))
print("SSRN Restored!")
# Feed Forward
@@ -64,7 +64,8 @@ def synthesize():
write(hp.sampledir + "/{}.wav".format(i+1), hp.sr, wav)
if __name__ == '__main__':
- synthesize()
+ lang = sys.argv[1] # ITA or ENG
+ synthesize(lang)
print("Done")
diff --git a/train.py b/train.py
index d69fad2..a5ad5e0 100644
--- a/train.py
+++ b/train.py
@@ -143,12 +143,14 @@ def __init__(self, num=1, mode="train"):
logdir = hp.logdir + "-" + str(num)
sv = tf.train.Supervisor(logdir=logdir, save_model_secs=0, global_step=g.global_step)
with sv.managed_session() as sess:
- while 1:
+ for i in range(0,hp.num_iterations):
+ print(f"Step {i+1}")
for _ in tqdm(range(g.num_batch), total=g.num_batch, ncols=70, leave=False, unit='b'):
gs, _ = sess.run([g.global_step, g.train_op])
# Write checkpoint files at every 1k steps
if gs % 1000 == 0:
+ print("Reached 1k")
sv.saver.save(sess, logdir + '/model_gs_{}'.format(str(gs // 1000).zfill(3) + "k"))
if num==1:
@@ -156,7 +158,4 @@ def __init__(self, num=1, mode="train"):
alignments = sess.run(g.alignments)
plot_alignment(alignments[0], str(gs // 1000).zfill(3) + "k", logdir)
- # break
- if gs > hp.num_iterations: break
-
print("Done")
From 9bec2e35956a6e18aff7c681ef5ae37533882a28 Mon Sep 17 00:00:00 2001
From: Simone Sasso
Date: Mon, 22 Mar 2021 00:21:21 +0100
Subject: [PATCH 05/29] deleting license
---
LICENSE | 201 --------------------------------------------------------
1 file changed, 201 deletions(-)
delete mode 100644 LICENSE
diff --git a/LICENSE b/LICENSE
deleted file mode 100644
index 261eeb9..0000000
--- a/LICENSE
+++ /dev/null
@@ -1,201 +0,0 @@
- Apache License
- Version 2.0, January 2004
- http://www.apache.org/licenses/
-
- TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
-
- 1. Definitions.
-
- "License" shall mean the terms and conditions for use, reproduction,
- and distribution as defined by Sections 1 through 9 of this document.
-
- "Licensor" shall mean the copyright owner or entity authorized by
- the copyright owner that is granting the License.
-
- "Legal Entity" shall mean the union of the acting entity and all
- other entities that control, are controlled by, or are under common
- control with that entity. For the purposes of this definition,
- "control" means (i) the power, direct or indirect, to cause the
- direction or management of such entity, whether by contract or
- otherwise, or (ii) ownership of fifty percent (50%) or more of the
- outstanding shares, or (iii) beneficial ownership of such entity.
-
- "You" (or "Your") shall mean an individual or Legal Entity
- exercising permissions granted by this License.
-
- "Source" form shall mean the preferred form for making modifications,
- including but not limited to software source code, documentation
- source, and configuration files.
-
- "Object" form shall mean any form resulting from mechanical
- transformation or translation of a Source form, including but
- not limited to compiled object code, generated documentation,
- and conversions to other media types.
-
- "Work" shall mean the work of authorship, whether in Source or
- Object form, made available under the License, as indicated by a
- copyright notice that is included in or attached to the work
- (an example is provided in the Appendix below).
-
- "Derivative Works" shall mean any work, whether in Source or Object
- form, that is based on (or derived from) the Work and for which the
- editorial revisions, annotations, elaborations, or other modifications
- represent, as a whole, an original work of authorship. For the purposes
- of this License, Derivative Works shall not include works that remain
- separable from, or merely link (or bind by name) to the interfaces of,
- the Work and Derivative Works thereof.
-
- "Contribution" shall mean any work of authorship, including
- the original version of the Work and any modifications or additions
- to that Work or Derivative Works thereof, that is intentionally
- submitted to Licensor for inclusion in the Work by the copyright owner
- or by an individual or Legal Entity authorized to submit on behalf of
- the copyright owner. For the purposes of this definition, "submitted"
- means any form of electronic, verbal, or written communication sent
- to the Licensor or its representatives, including but not limited to
- communication on electronic mailing lists, source code control systems,
- and issue tracking systems that are managed by, or on behalf of, the
- Licensor for the purpose of discussing and improving the Work, but
- excluding communication that is conspicuously marked or otherwise
- designated in writing by the copyright owner as "Not a Contribution."
-
- "Contributor" shall mean Licensor and any individual or Legal Entity
- on behalf of whom a Contribution has been received by Licensor and
- subsequently incorporated within the Work.
-
- 2. Grant of Copyright License. Subject to the terms and conditions of
- this License, each Contributor hereby grants to You a perpetual,
- worldwide, non-exclusive, no-charge, royalty-free, irrevocable
- copyright license to reproduce, prepare Derivative Works of,
- publicly display, publicly perform, sublicense, and distribute the
- Work and such Derivative Works in Source or Object form.
-
- 3. Grant of Patent License. Subject to the terms and conditions of
- this License, each Contributor hereby grants to You a perpetual,
- worldwide, non-exclusive, no-charge, royalty-free, irrevocable
- (except as stated in this section) patent license to make, have made,
- use, offer to sell, sell, import, and otherwise transfer the Work,
- where such license applies only to those patent claims licensable
- by such Contributor that are necessarily infringed by their
- Contribution(s) alone or by combination of their Contribution(s)
- with the Work to which such Contribution(s) was submitted. If You
- institute patent litigation against any entity (including a
- cross-claim or counterclaim in a lawsuit) alleging that the Work
- or a Contribution incorporated within the Work constitutes direct
- or contributory patent infringement, then any patent licenses
- granted to You under this License for that Work shall terminate
- as of the date such litigation is filed.
-
- 4. Redistribution. You may reproduce and distribute copies of the
- Work or Derivative Works thereof in any medium, with or without
- modifications, and in Source or Object form, provided that You
- meet the following conditions:
-
- (a) You must give any other recipients of the Work or
- Derivative Works a copy of this License; and
-
- (b) You must cause any modified files to carry prominent notices
- stating that You changed the files; and
-
- (c) You must retain, in the Source form of any Derivative Works
- that You distribute, all copyright, patent, trademark, and
- attribution notices from the Source form of the Work,
- excluding those notices that do not pertain to any part of
- the Derivative Works; and
-
- (d) If the Work includes a "NOTICE" text file as part of its
- distribution, then any Derivative Works that You distribute must
- include a readable copy of the attribution notices contained
- within such NOTICE file, excluding those notices that do not
- pertain to any part of the Derivative Works, in at least one
- of the following places: within a NOTICE text file distributed
- as part of the Derivative Works; within the Source form or
- documentation, if provided along with the Derivative Works; or,
- within a display generated by the Derivative Works, if and
- wherever such third-party notices normally appear. The contents
- of the NOTICE file are for informational purposes only and
- do not modify the License. You may add Your own attribution
- notices within Derivative Works that You distribute, alongside
- or as an addendum to the NOTICE text from the Work, provided
- that such additional attribution notices cannot be construed
- as modifying the License.
-
- You may add Your own copyright statement to Your modifications and
- may provide additional or different license terms and conditions
- for use, reproduction, or distribution of Your modifications, or
- for any such Derivative Works as a whole, provided Your use,
- reproduction, and distribution of the Work otherwise complies with
- the conditions stated in this License.
-
- 5. Submission of Contributions. Unless You explicitly state otherwise,
- any Contribution intentionally submitted for inclusion in the Work
- by You to the Licensor shall be under the terms and conditions of
- this License, without any additional terms or conditions.
- Notwithstanding the above, nothing herein shall supersede or modify
- the terms of any separate license agreement you may have executed
- with Licensor regarding such Contributions.
-
- 6. Trademarks. This License does not grant permission to use the trade
- names, trademarks, service marks, or product names of the Licensor,
- except as required for reasonable and customary use in describing the
- origin of the Work and reproducing the content of the NOTICE file.
-
- 7. Disclaimer of Warranty. Unless required by applicable law or
- agreed to in writing, Licensor provides the Work (and each
- Contributor provides its Contributions) on an "AS IS" BASIS,
- WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
- implied, including, without limitation, any warranties or conditions
- of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
- PARTICULAR PURPOSE. You are solely responsible for determining the
- appropriateness of using or redistributing the Work and assume any
- risks associated with Your exercise of permissions under this License.
-
- 8. Limitation of Liability. In no event and under no legal theory,
- whether in tort (including negligence), contract, or otherwise,
- unless required by applicable law (such as deliberate and grossly
- negligent acts) or agreed to in writing, shall any Contributor be
- liable to You for damages, including any direct, indirect, special,
- incidental, or consequential damages of any character arising as a
- result of this License or out of the use or inability to use the
- Work (including but not limited to damages for loss of goodwill,
- work stoppage, computer failure or malfunction, or any and all
- other commercial damages or losses), even if such Contributor
- has been advised of the possibility of such damages.
-
- 9. Accepting Warranty or Additional Liability. While redistributing
- the Work or Derivative Works thereof, You may choose to offer,
- and charge a fee for, acceptance of support, warranty, indemnity,
- or other liability obligations and/or rights consistent with this
- License. However, in accepting such obligations, You may act only
- on Your own behalf and on Your sole responsibility, not on behalf
- of any other Contributor, and only if You agree to indemnify,
- defend, and hold each Contributor harmless for any liability
- incurred by, or claims asserted against, such Contributor by reason
- of your accepting any such warranty or additional liability.
-
- END OF TERMS AND CONDITIONS
-
- APPENDIX: How to apply the Apache License to your work.
-
- To apply the Apache License to your work, attach the following
- boilerplate notice, with the fields enclosed by brackets "[]"
- replaced with your own identifying information. (Don't include
- the brackets!) The text should be enclosed in the appropriate
- comment syntax for the file format. We also recommend that a
- file or class name and description of purpose be included on the
- same "printed page" as the copyright notice for easier
- identification within third-party archives.
-
- Copyright [yyyy] [name of copyright owner]
-
- Licensed under the Apache License, Version 2.0 (the "License");
- you may not use this file except in compliance with the License.
- You may obtain a copy of the License at
-
- http://www.apache.org/licenses/LICENSE-2.0
-
- Unless required by applicable law or agreed to in writing, software
- distributed under the License is distributed on an "AS IS" BASIS,
- WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- See the License for the specific language governing permissions and
- limitations under the License.
From 51d9d606c9fd11d73000d0c7441195be3ae56bf3 Mon Sep 17 00:00:00 2001
From: Simone Sasso
Date: Mon, 22 Mar 2021 00:22:56 +0100
Subject: [PATCH 06/29] Removed fig
---
fig/aaa | 1 -
fig/attention.gif | Bin 3213348 -> 0 bytes
fig/training_curves.png | Bin 264024 -> 0 bytes
3 files changed, 1 deletion(-)
delete mode 100644 fig/aaa
delete mode 100644 fig/attention.gif
delete mode 100644 fig/training_curves.png
diff --git a/fig/aaa b/fig/aaa
deleted file mode 100644
index 8b13789..0000000
--- a/fig/aaa
+++ /dev/null
@@ -1 +0,0 @@
-
diff --git a/fig/attention.gif b/fig/attention.gif
deleted file mode 100644
index 87df49e23d929f44325f2a399ebbaceb79da9457..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001
literal 3213348
zcmd>_`#%%@`~SBYn>o$-%$(;?Nphx*VGcQz3aOAHq&bF4!!|R+oX^a8jyaSR)ria~
zhm>QQLnTQ=2PA!7@9)3yeg5$L;riiuyIt4qem$Sp^|)P6TRR&gV;?+-4g3xG-{Rxr
z6BQMelarH@;8Imp)zQ&8dGe&0nb|3WI~Jz*P9glw0QM+=GluKDHIIi4?2;w#1uNkz
zwt_ykqSvru{upV0TZtQ3$s4v(fmqp_SOuJwa)_;Z6jnLhQajRCCl;$2W33lwhPr?3
z%w6@XaTpB7(b3V(&CScp%g@g*I5;>mGVSU3VY2=Eb^trvHZbkoU7U3e*v^{MY)0v*vj-7WEUzXCKtt
z-XDFksQGA8xoX~^enF2uXIwsZl=W76W?J&otlY?w!sN2-!ix0ARq2h7vfCf!dp@Xq
zpHp5MMX=b2ezwxDPl)}`%85gksqd^4ryP^tVl$^Lb7oN12Ju#=iKncS3{EE+*(IE^
zC7U}WAG<_6;YK*_Ni;c|VsMda8C1mzk|Ld~R83TENI*8@
zi9{lmO3lvBE-o&vuC8usYHDw9@9ysI@9!TO8JU`zT3A?EU0wb9_3PHw*1v!M{vQbc
zKW<8UU?}20$Cdu^Ia#}%wL-gC>mAkQ0{&;x`CmN#mnHz{A;5!I4%Pp*I--qaGvhfp<=uCn$aSo@b!hk
z_S&&BR3c2?{vl(c5}T!UY3O0yn_8Dj`x5&{^=}*8+XLSZJ!+VK;Qf*!@6gdW+aB+$yw%tSs#
zr|!CQ?
z6Gzf0|L2
zaov+tnj1xNt-MAP)~fO&?`ueB#n#tbr{=ELHyoBz)&uJs-pRN&Jor?o-PG}G2;Tg3
z-Gt}CbBR0f7C-7%ecMp3ePio5Y-(L3
zXzuvaCaC_IEiixi18_Hhs3&*?#Hc68wFjl(c~x^lcg09&#zBwpD*7xwZ|9ByCv_
zDvTAkF%2L$!-vrgKH*~)Z}u0fQJeNuysq71f3?~$##g+4
z`f``}Ii|uNdq5+7aias09=n{N4S+Na6BFewS)o4
z3F3%@6%-r?h+oS{2ec@FDy7hgcuq2pF*m?mnBuYvP2%4m0ARWnV193aHeC@QQ)elL
z?-kXeEd@DE5fQIL`I-qIC7#kv%|!YDhq5e9n?aasKqvwU$D?p|v`h8EcU%c7mQW)A
zPg;h!(NsCgWyvzWSQ&6q9|$m0YT|-siEGSUz3Hww99DQX_n}Vd5rH#!FQ@RLD+kr6
zGn~;V5kld>p;r$KKp!DQQ~(q#dv?{0UOBJoir2YVYi$ft_TvZXzbk8j{#LKUJtJ}-
zd1+{q>l^@H6M3T7xM4OAU2nJb=0o@a$7C}pE^u9Lk&z{o8$daS@&>Fbj#c~@O#uSv
z0BvYp5UY+D7Ff=}600R9>s(24VF*+1Dn(vPtd#^HgVI+AEFGkQz=)bjzkUll^t(2h
z@hr02CoN1l=)V)VL&12lH2dkNbLaTZu+~f}q3Wt_mg3u2D@4ns@Xwtx#Gef-f~Ic3
z#T<_pp3DA%c66CJITVhV^J~1}q(9viEAb%qks9>LAmyAGodzJIw!*x
z)mi)S4}=^_R;Gyoje5`+;YBCHav3Lko!3%{;s62wO90$_Fxn^(vmbBpNB0>JUSU`?9S
zu3K9L;|c(;bs8+i6cahGqRq3*2DPxCRC{dVZ{KBeIiq{{aw#eCTx?CLS>{<34xe|h
zvdZIubpAP-T;aW{b#ZU#^sB9`WWCi7j!%Z%pzRr-z5TS@jA2fT+lS6ib*=$io{aby
z?1(L!5sgQHqn_t?@HN4oEq)ikqIe$4t>XRX?<&VY;Ps{DdLPkaXu*3553AHJ1CH^YklEUY;_emXt68P)aYJtysP_vh2!W9R=YD(U^QsyQIV4flLFOiPplAMzM7
zHzx}gBX(M1C>wFN5EZ&iK-*}(8!L;Ok-CwDWkLjTcsSx83Ur8JZwcpj3n86RAW%=3
zS*ID71jh)%!9ns?&>xL10Klqr05=r{5~DMf1eJjFouxGx7C(AzF)VMT`;)mdYD}EH
zWrbLV%aLQoj5$C-xe_5|ES?+b9WT6lu%?K;Gp
znJDg$hYRamhR5%kENa5{D^Dp18%f4P8wK}-X5@skQG}0@V`1psxOJ7zy$M%EvW3{G
zANSaaqgm~Hg--*ku>4Y97`d|K2LrQbkJ$ta`S_sVWV5ud}ZT!?o2
z4CDi-G=cyjEx~_t_$oPdpU>MT`UG0YFh2B*q(p;1Fb6fN=mQUKe`G9ZG6R
zj6i@Q;Dji8VkUyOKo@$ROpIegV!a`$C`d4isLUZq!;{R{fRQec5O?0_T@uL!N`#a0
z0MIzZ9c3KpEEF0wPB;vg0x8j<$t_SLECdb!-e)Ey%p|tmCScv6_fcdMcL)Lw!bc}k
zT_EOIa%euV?J?17mte$zHZv%ZYd~A0q)03&5kZg`CrGxWAP}JN8K44<0C!0?XAvwr
zfoKLWYL`f*K~vpNzB24D<=Hq$2
z+>@v{>H?Nl=nj3>$$O%Ocaue&gn}XfnGtLX1w})|By(6vku5PdY~UKpJfWOJp5xM4
zdIj;Js4N8r9EkGOWZwiG59OKwEn{pJvTbM+o;#~qdQd>L3)wh7(PSpkI-eRzCX&X1
zwo=LOche|H0@#D#&L9RO2_^v$A0)3ID~W_ki(^7#@QG2KM4UTNnhlgeC8c8_(F~|F
zBiDZr5Fc5Nb^o53p(wd;bxS4pF=33XKYGkPVSPF?J~iGBxfXOnHQARnF$-u
zbo7fi+@r%~%5mPfI)u`|N%%1%_;D;8IPP4=IbA|?KZ
z)gftA=}a+Qpjb<0B#pOV9H=x7oHR;}z^5@;N%vczsYXN&IVne%xGW7?n1L@;U`nlCw*t{BZCM8p8ijcW6xlIoyE3W(GJ77!hf7r=l}7|@g!
z=uxdZ;@&`UI5EQxa#ib2C^Q|v262II-Wqp9&J1J=G4G{fSH8((0
zt2A$y1K#-khNINqr1}4DD{7O>_VbVRR9!oA@C7q#EIg^ueQN#`2>KGhp37Q_f
zMoq*YRwLMH7Xyj@42WqosSwZW8%y#V$|pqE0P1U~9!cth=@$D1DeGxbS|Cvupg|#U
zfK~fhs|mYH0`Dh(xlkzQ&5#*~M$v!~dno|TJD=+*FQn>%$y8Q8RR*7C{Hv@4ZmmkAA%uu>rZ*an30qh|qy&(-p#MH$zVmVD
zaIBY65?(eV5*1Ywlj<~F=`@#$hEM@}1VMV!!dxl~N#4*bt^EoZlOcS0+RV)d4h29S
z-^uE{)76#Ig+iev&%ozj!OT72Dnqd20WOHlv#G$F$u)2dP=sw~1!BD7<>6lHjUiGL
z5)`D7;RAi@SD2T`N|nC|U4oLA0(qmc#IQN&W(rC3B0Yjxd}KJ)gF^DRBjtFM#L;;T
zyV8a2>iwl6ZkQBC&eZHBCu>M&=C1)aq0|KTN-=t^EA&y^P-R&RA<-mhko73S1K4fy
zsK%uJ`EE+)P|Iz*;^28;VLdcFp#Jim)O&YIlw#`37>}RM)yHw_`Q;iCO`l2)lcGEe
zyJdQJjCvwG8qgZ07xz-pNe#F^Eqb01aX>F~?TNH*Q@>~bzunZ?r%j^lw8ecAWIuUz
zA0X$ICr@vViTba-2jk!aVP*s2a*!8);%lQSRE@CC*|1v9z@d3RsPmDuM&=_&skD?K
zXcUU}ZZ>5Ho4GU$-JD6~8X|Rc7WZ`)7A`!Jrn9K@j19(Mj0Ac0nx}#aql^K8Z|(c*i@HD
zEChBL6>INJ-VBBxHy^M?4MY^bfh~3WM_xP0opRNCxg;P;E80iBJJ
zEp^G`uO>K;kg(JO7JCjx)&{&oaxp%dXY2)`Nx;0n^UN=>^u1JbQ@XPPk(<
zT38KODy#n)RPK5xX#rB~F;65Ac)ykpEQ#BLs5NIF+k8AByuxA_k1U8%&**
ztNJC&m?qDie?BLkLQF*PW&mdB+VeB1)JHex@huDy^JLZ8N_pnkO_>E*I?Ko;U973W
zQjV(dd;w6uayX|A8W$j=<(_`WLYg$+Ndhvb0}2}ka~J=X7~9(osCbNNG!AfzVk!^==*m`m>1e@sF_JWrBNK{MkJ(^20b
zW3x4J^d@OjiaT~f>2LDU^)jkm6}^GgPyp$GLhpO?rn~SydR~m1t)lRM8xKhg#%8V=
zy`PV!yk6^|xbRXqi4lvbl`~N7x++I9vtEjr3+K&XcZ6tu%WX+~oj;6cU=##E)~DSG+=#SkV;F9u8#Fi$#5Cc@<`h3tjmjC`n1_mQJ=_qO{ydplynw`W64g7
zhc$Vq?+qS>2ybGCR6{V9DCzVRzdK?;e<^=&Czg}Qd7Yn#f+{@ULCdDe43p+=K$@fH
zC$UuqjKPzFWK%|~#nYikdR+=2^VpwhVhcp~@2lFypNx=o8~S?VQMu;Dz(6sgtuLkQJl
zGiSWaO>izWpeiJ2bLuiwe13z<+?>;F%MMQLeF7A|^8K|tB&ua(zZAMwP=6Q#4NfC<
zkjsWW2<1kfZ;s6p|GPI{{8uTYR3CgKA~M8ZKqX%kf86X)H>cpgZl&ELnI1A&nvUDS
zux~%&c)fxO6Y0r*YYR93L98rlPo#4h(Hl;r>_mAJ=Z>Zh8<&|rXtlEV5y>d|m{w-@
zAbGr*<^JMDB&+kx_m0ECRDdgETmjJFE|1CoPkA`s000<^s4i!y;E-n?Vui72lDv5w
zSo+K=HdF0%DW;&U((1ofij4GQz)J+>k9sk(-i$e!L~g7OUQH6N6p&mS=RJMC_fhN;cHhD*?@Nh`cxvMNhvU7hPI@BUcerWZ-FU;2@si$PmL!edM}&D
z{FrKVQA$i=$d4qf0eY>Hn8sZS&*}!!N*~Khm@+rm@rk^Wr51;Pzksi21jOf#mgv4q
z-d=d^{`1IGca%g3{6dg;U;g?Hf?Y;&25g7zauy5E3VeH
z)?}{gKDvBoq#pL@2lMz#sP?*gmmcox5yLN&fW+O7$-|%rgWJ*CEwJbl{Kvl=1i&9u
z8n{osFcUc&%r&YP&Q^blDnjWSs9bZXdv-=P<&n|J8CXXnkGunT9>ctOtXD#Nhf$ue
zBidWsvT~G*p8wK*q_skpU~;PaWeC!=v%&795Nt0o?a#
zBv`;;rd}}mOIHX8Ch=x{O#c`Qf3jvdGw4CF@uB&2;xhz|s#kP$P4LK*z%%?UF!0a%
z_KLg?9sO~ml&Lf0HdaplQ*P~u&&5xc)3gPl?cbWt++Im-%nh$^RPRo+I+}R+W6kYs
zr&H?DjP=8p&Ckp_zaFXQz@DTF+UX4vu5{1My9>>&?j$zsM{hDgu%n#x8*dwB#xD+t
zEZjhDs?*brc!Nyrj|;)OT+;b%e^Xx8oAC=XRBjI$E|er}=XFX(@=M|4HMM5gvI4I%
ze3HwOEMmV{%VcH*3#n|VYZ;e~rJn*EchP)E-L`k`I=ZG4+^o;sTPQ62StL@RKN{_1u-OzxwQWL27ilUvE9
zw|>aQ<9!I4nB*U+pi?iDgksj%WkaDvK~))VkyQ~+>5wn(
zYMuZYa!jW}$ae!&DyYu@g~;&T*5M3V1!<6ueiu4J_!LL&f@qBP12MBaHypR+oKYq&G1A%}elOegyS0cOwazw!7Q@C-=Qxla2
z6Z$DD4*yZy4eHhBM5dU+N36M&cifK)%Kvnfh5y3|V%V9Ly@%7yZ7n`9sshHFIeF+N
z=3deXH$ldtkTLWMh1Za5Gw+FGo^sM|rH@o6>ISp|JfuCR&c1s02zKn_nHA-^H?RKH
zaT~rZs2nzFDJ=(B$7zoaw$nS}%XfuoC#@$+7d{*A2G~k^5U_hKD=am4G2zr!2)xSiNuzEI$d(idK
z5ON%Psh97_^Fe1v-_FBhHe#aj>vrm(V@k;{ex?XV=-A(=*|U|uPP@_PGiP$Kwz-n9
zbJ2DXVVl01=sk4p_P4vAFK&-RxeKsp-=7la$8*N+No)zFzY!dQ^Uvvbhz
zp?MUpVUFjVsFhE@X}CP*f>z6kTo1SIjSMF;zhS(_dXNUA`1CQf3!=H(+Qn8HqwLh5
zV;@=UP^*yd*;JtpP@ZFC-sxzOp0@*#`*24tX
zT5)u1XtUB8fu#3U^okLlo6F+0=PYr)CRI}82&0TjUapnPfZ6Do@WamwbpBOKAd&so_ON;q1rIS5@)ipRo)`Ii*F
zvb&Ms?M-7>rFYR)2TgOv1*sNwT*9uJB^cC+T-qBuo~@rr_EY!Zq;a?8E6CF3&kbI4
zo5LE(O_3?qyJ|l&q-Dj$9RBHC%WJk0%??WuH}*9=s+Tf#dC^k&U%NOBn@>^W{Qj_L}(w(FeL+Ot`AIUDr>CJGy0Kf87v5Z!Yme=uYO9$DVyT
z?Y#fzLs?*j^ZUc!$6`TJJzmraja*Nm2hI8X0?WM9n##V$u$W=q?fjEXblyVGoMLz7
zhu)?p%O}};0^H?-fgpl?3&FnA^u1PV$x4@;h_>xK;N%KSV})XgB^<21=o6*%%c9KP
zVeJ6P_%{rCfS|D_^`)OP)Ujy3m<<$@dn)sf91drde4*&yH8aRg
zj;NzFA=NkW=mv+P$8eyzP6Gs0-M$8Rx(gR(Jwu5}-wT9&-%BiHt9on+ZLPH2oy_8T
zBY`Cxt8C~~iU!$JPbo=(3gr?USNn>m1_?NFyD_Zejp#!urvGY{+mA#`@5kY28-q+C
z6i4v;R^p(ysn!HOc*VwGy&>$0*dnxp$5jY0o0ocAST{y?alDKVMBs?Zo498&Stptq
z^beC-g)^;{0buaHszjgF;@SU!p8&a((Z7w_HPQ;JUYN7Pprw
zx|L>EnI`4{{#9p>k`jwUKWm2abu0;JB*+f4>BWp1iS_JU*VL}A(;q~{PKP@9_T;w$
zn!B{L=n6OM9#^d+@uX
z$q%k^Y+nB!kMkM9K^eYaL-&^OkePgOCaaPNZ
zsnTSQx(DrXjDXt{(?Bh-K!1wbaynmO@-*K>*SbXGONZW>6
zZ=wlWNujDa@EftY)$|e3a{S`I9CXJ(yTR
zKcf)K`)+osq%)_9PVd1r5ps*IWaXTfVd|DylT;Op-O15*W
zeF622X_bp~Olylm?+{h1ZIfkD9cp9o)qEQ++ru>U)k}FxuKX>PLR3pvJ$E90vbHF+
zqYw(*+D+_dF+efhe~weLCrTNB0W;%<$gc*34SIv6ako0Y
zu;b}>XNtY~E}c$4Jb9eDP^1fUuPF9;IN>2|I9=~&Z
zt};h*WWPlaD%=N+-&{WNFVv>n8x@%Gyg^*sQ2MyWpPJdW6p7Xke~!>95MtGc^X>g2
z?O522pJML==Y11Y?)(G@>UwXlbu(2fezc{$`eFP0?OWA1j0bn+4TW?j;sn<11Ry?C^dc
zC0(ML`u^1JVU2up+l&bsH@^V18xtyIpO_nov
zsRRuoU$Q?O=sxfB?gE9jaqH0{3$+b075_!H{bASiA^6~(mBle1p
zm8OtMV;^TBS?!@VFiG+*nU}4BR&-{yxHCT
zC8shXE)z#u|1~bUj)F;6e73bpS#H#S>nYv&{xh+kQ347e;h+l93F>RXIX^cJ)j1ol
zLqLg6Yj3V!21FC=aVN?H&D9=8&pc%Qc#NA
z>X`nGKbPt%t`n?Z@ks=J$)5y3qjH`Pbq-+(xzBYG3sM|uipuPl2UM7VYk3~Dywrie
zOAMwKK7a0ATJ>g%d+xkP2N!R1jlBzafn|N(H&yE|`=92O?=$qrf?vL)>1x|Te`b}g
zD~O%qmF)O*RbUor{8cU}ZkbP+)3vB*WSX)0_PY(cU1X(QkmGgGEsBzW9j(@Ay6AJE
ziuciqq2R7GhheyUB4ywm{=UW9rmDxpuPsH$w)$@Th|M?UW2G&~CsU$_r!Bia+*Mxx
zCB=sUE4*5%E1x0ktWn@>H$Ipt*H0hco)cr&Z2ry?PLG2xP2_NArm@WW1B)7bWhiEG
zZSXBG*9~4Nv2<)F!2#M8DI#>x4E}ja-h{YbLnJ9&nH@%~}ImNGG!Rx-ZW
zJ`Csfxel*8_X0_HGPG6j=_tspX4XK=U=DO-rW`(B{(B*L=Q;f{;YU=fZpS>GpRZhN
z%394@^zIjdll$sc#u?dP1oARdG#7RpqylXWk8Em7f|Vui3oq9(QwQ<>w&w}f-YV}p
z^_!*qdpW8G*M-1m#7(7u$`{C^g4+?2H{Td+N8Oi7A_JgE$&-gSudcnUpJ=_JNr7Pw
zlIM03-zcT&eIQ&83^C+2jYpWpqtpag?c7hh(C}!1K1mcU?R$YncstL5q
z?5qBHU_76%x*cb}y&G{7_5%}L=in`O-7>
z3qo7oNNA!($4>t9anwJW_Y021k(mVRU$%=rrqR}Wr*=_gS5U>Z`xv9)cgd!g6jN9~
z(g-V*ioKTAZOZ1*d?gNH4Hffsvyfe+7%9Z8(
z{dBtfy7^CJ_-Ug*L-8qabTnk+rt;*uB#rPB$BxKnrX`o9C!Kmvr-gTO=3Dq+1Gp@{
zFgg}$YNsMmAMyjUrntibLx#YPx_c;1Fz>Vb-Sq376M2nw;DMP
zroQF;zWcQ6Do64`b&7n8$H5P7{q*=-dV~vV9`{4H{;Vyg1d}WQ1g&aJAgqY6cz|#(
zy_70%63rPeFE2}#mr$ESK`^D%xP$R8D1R(qyi?QZpN)-dR_pcvW=nVVZ
z=E;p8_)nDGHt=a9B!76}Sa=ZN@LPAek0N;@WMlXEDl1b%v17g#&JB|>xBzZ_bJgw9
zjkost*^?scmG_T&&3J@<<`$5??myE02EJJ#TfP1SaN;TnCT(EWBxTSSVs4gv%jl_Y
zMvn>CwJW;$kJO`*%vv|6<_?^dlBt}TEkL+08D4rRgx}ef;18c@z211Xg=?E%Mq(>T
zNcn~eRXJgFm_H8#^T;$a)&z-fJ
zKkjRLTQ`27!t*2gXOfcr^+5%w=wF{&|DD+jAJD8w-J*gNH`N6Ew;D;y>(Kq*u8B_P
zC0B2f+yf>EmNxFT_A!1n1&2H($-4t#Cu4NqU65EilJx0OF!!ZmCXk2quZ
zjBWdp`9*H|Nbew?=a&$o4@A9mRa1d2Ebo@}EMbHS
z(gqUzIj`tB@IFK@#RaG5T?>}Rfa9LZ0OHsLafLvl#~mfUp6eablPx5JyZm!g5~@XL
zHOmNmQ836be!8(LLK6SRKJZL7#mpHOdhu=DQ5sM~{k0+22zA8(WjN*-V&qv!gOZcz
zUIn(=C0WY06%`Q0RvWI={6U-Zw}Cv*9;he00R&yP?8@lT)GR@JSu>K#!q2g9i()DA
zlP6@kEDx{gcx*FCHN6i7%*M
z<-PP;-892v%?#h^FK}9j%Tda?{!zxY>qv)_vDcL<>LmV*jua}7)2(Ar>O0S^np@?HavqsKM;|6kQ5467(xnP93G(bI~uVT3fQo
zj1EID7-d|}f%Yl=70>fLiZm23lT?16PHO^IL&wpblKBM61A3I
z)Z3P31R$qXM9?62umVH~JZa&1HlZU(F2g3^Wyzp;nCcnwnxkKzDk=R_K4qD=3aniv
zUYg|&UaUUnT-Ck5>Dr9?4ck@
z&1FBX!9`ick#L^ByAIOv@35(jw7KkGQ;ivRH&ums3DfAIOBO{rIf%O&(0@z)$1L(R
zN2VEpnvp?PA-}G1k_qs03!xS19;ACPBm=8hw1g=7)^)f#7X=rh?<^L2&W~
zI+R(!EW(@Qmr(F8M+?mu=6sd63FhCQuH;P9}r_l@3D*FdpkADU5GqDgQo@M)!>vDehyR&W^E$
z0tJ=rR`&4B3D%V}jbOajf1a12f|Gw9k&?$Q9WBpXgePBM7M&LWb9)0GZt5
zt_Gk$9mq{?uLsqAc1&1xXwtEtlP|gz^aCrN{yiTf(Cf~y(cDKf7`Bn-j!H8C!5Mry
zlF;W4x*H5q_de|9j$CmcnX$oP?Zv9q!Jiy(iq3v>m*1gP@+%bNiZiZxMnv1-TxM+C
zE!EtH(=~lut1zGAr*|4Lp?2K*a|4o2M0vu}(lxuRNQDQSGoi)<*%1mMMi{
zHy(G`^D!6+zIHyrWU*RZPQ6vX5Q`(bd0Tr2f|^)1KAF{~zLkm$qQXY0Y6PYRopi4l
zyp}(lBGz}!#?p0MA^r>6U$8Hgg~1mK|HayMhRK)^1v3&
zu~BMcVK5$CHwm{ai0wb#-IA|WMkItCV4@;z8^Ji()glJTdPEP_A;|0
z?tXG}BNR!M_H|Pl_rObvUDD^qjt>${NVM+b-ecJzM6z~)eJ#+;AB%<0
z=WROAp^bDxGcnF?pc1VsIX$RSQX}h=CjYJ`y;LWEmiQ6s97~4^l{kKhu(8-5Z+IrO
ztTSoDO77PKjiH#q?tpmrzAiDzFg(6)fvOm6=^X|RpCVj*SRk!pZyZXLG1nT=5qq@(Ujaz&*-;3pl@LK4Ztl&vK^MiZH@K*
z+tcAltbv|O@O!0MGIpj9a&oip{kCFai_J~A?T15|UaA{D5HqNEGLNT|DsQBdhzMB@
zn|IX@X#pi1+HGQtFgI|YZ5zRNW}2&ioUQ3Ac?f>DMw)Xd36T!TB6N!d(Y=X-uUX?E
z*8J>mv3>Y0Wz8^^!n-_i2hFjK1kPnYA+hfQMzuoZP?2n%ZLMQ6Nzwou;G5ez&(=5(
z_E^08fVjq+KO@w!A)eobn47OhxCk!eg$_{o)PM2aTLihYmT1^`WLb;G)fJ|dNST-?dcqiUbP>dAq91&LcN8ZJE!xyNgv&D-0UNW
zhuIT7ydlseZgvecFO7KYc-(p|V4Bylon9-
z^0QpPYtSBe*4tp8ZapKM(An
z81SW15iafP#Ea9kwH~JzGLjNw#x2Q;>c0I^o-Yma+jhXUCE_8r7kLrvJ~ad%WPI47J}{(U2{TBq%kk}R6A$C`xmVT6
zQ}Kh!ujKaI!$lsV#-gN(S}i{qTq_13i+eSTnE;(3m>57a8w5h2Y`9Q!W{uDD$ucqQ
zSEv-MN=g{I&+aylb8C*r{1tg<%HzI>Gkzp)HVNVfP<<=vq;g#zAs8PFcC~cixh9H$
z^0sbWe;Nsao!^>xB9Y~;5{?5D<6gb#?xRWdS~AXgOGADwf31t~R2xx=^Wt@T$mdKx
zd8&;(qMrHlF!IJPD)IAsPiQdRZpu#j)7wur^{bXtrQnGd1lvB}+quwTX6vh^I+25l
zS1Q3^L758*Q-0hxFv%Od&Y*?&KEu;J1S_*rHFroTt0pllujcKWAa+LMOr^f>HsQ^i
z1SmL!1r93P3IhT7rGB9Pwdc$JI5lq&QoH9@1p;J)v>BEFa__}r%c3p=fIG8Q7qnl)
zuwoGi66ID00gry9=UBMlWUA${)R49us9U3S$aD_0bixT7+(HVz_LE1fA(#RFW#9NB
zu1}GFoaHERJpD)F5@%+KUOoukuexFg>yWjFM2Q89b4Ozp;Sq>9-9
z{gB7eA+w~^QO>&6K2ez1O<;y2cf{u06B2+!fa<*UBt=h0K!r)WyZM@3m)Ld&^2h
zWrtKqg%(1ne1*1dU;o1S;eFnZ*X#9so|LNN@}oI&!vR&b(xZ1zhC@7#yM33)diXtE
zYI>1ox~L$`HUC$UJa2yH;nLZvak5Jg^S{S&AkgvdOa8Pgg!i=3ykGdU1A^?n(!c><
zHV9A-QelDs4ImFbn!?1Ga`~ZhyWvfRWO>CeHcP-pcF%FVByBpg>sG+WznK{a-+m!@
z{4R>-87j&Av;ZkEJ~M7XhRMOuooLU?=hv;?WviWkdiGbuhsL~9O_$x6>p_ysk?)Q|
zv!AXJvdtahutBiv`1g3~AqV&!Xk2#kaX>pX@5Sp5U*2$o->awTR?q?pUNVke8hRR%
z~t;KU-Bf7IqiM$$MAhKmjFzt}q^pY&@CQe9(zlo2gk2)Brt`e30PU=$m
zMmvGWWOj{r{6}+$Z$HCqGnKGZZfK*#05O#L=J$}Kfr;nBm;mM4O-}S!*{`wv2jk7Z
zU}E|zbRuAM4A@Td=mr36{vAXER<^G^-cL~CT-gSJo|sv$dX&XBCqF$)Jl|D7ekt@{
zudhsu`C~Z-GR$l01-&Qio)3Rv{YNXFYm?X9if1yp@C%I}QnApa*iA?KpO2m$H@+lq
zL&Jg~lB5*nL`pYDaWUUuWLHVd
z+}`|J{jsuwCCdNyvz^80@mm_kwB6;(5q2JKOLhUkXr{E?U~G6RM{nAh3U9e|9&FZ4
zm|?eST3n+?xD&4{+Vt)o_(LC_XjJzf4YHfcPuhpxcY%hFlu3y{ZEp0
zJEUJ(k^`w0MJFrbV?Vv>2z>RRqCY|yZ2wDu3-A*tGJLl74AUE@YvGN2KJgBpae<;DDv{sd_x5*ylm1oP)8SWbbS2>KFQ=v8Z8T*X;LWif`DG5tmV`an05n-WYQ`tiy`wjAU=gAG(`O4MRkB+|=
z(hv+Y1W&SMinR7>&CU{IMy#HV&B&OFcQ{VxiHuO_jYbXlz;x_p-?K8j-8Xss786D>
zcLAq8m3kr`(r$FH=FntDiieZCw|_pfGMb+21F`fWbV0oGwjc!g7Z96dz8wTPZv>I-
zw67PdK055`r8LLKHg~q_68Um%@g*TvSV(btE?k?&^Emd^PP*|2{;EKnO;`54O0C3ibFhMpPBM-mDh3af1sXM
ztC#NF=W9|Yg#5eDZ?1KA?Vo+d9;jH?PV^yjTp^`_uS(eq2!9g@LMaN(ByQAKCX5D&
z&H!vxeHDN5n=4)h!DU1$+xcD7LL#E!gb8P4X;9Wx(pPb!l#^+r4Hp8pD&!P0SGp$b
zL!s?izTGAk{HAPa#x&TxM<3{$Ed;6+1ic6_$SWyY0O|#3b0|5zr;wgV-l*}(@^y`R
zi79@S7=N7)n8&?|047d2OxC%J;++o0dKr!V!t>xExwjq7mE`8>@_L#$;xB11nvHi2
z+N1FpEMLOdr#o;W5^ir%6;Jl%K;1eS0;Q$=DZ
zXgkA9Vykij`OGd=(iP)e1v-1O2y$c3C9WHxF+`wr4(
zF2@4YWBLbcX`q?kpJf?Rc!m>bsP7B9p4;%J@WwF3u)-%hGw5LJOhS_+fUlsTa`%a{
zGnNKmV@+EB4A#t<53%0VpK%l#I)0)9m9ja+L<7#@gwrqamfv%~~(Aum8F5{O>lN=bzkev$y1hZ85Q-AFl@!|9LU_-l1gp-ZJT2zZi=@
zAtMch*7NGCO*WsTfqrVE3>a}X287e_m>j)x)d!kGpYntd+$tN12$5jKb$TxXS^S5gpJigOxZt`sWxUcVcA&EZ$
z@`tIjU;E!qR1Wu)%CWx#DM-giB|cj{jTAuQc~jM|?!Yke!{vU93)^$=_9PXu!nB<;
zXWVM;YH@H*2{eY4$Etmyt)JilCQKvmSOQT^ZY97=HaVZx*j0UAVj0cj@F+nb$}ZNq
z-|eO6kKfH1T|AI~lM@Brp(cbV6!hN`BW5sMHsE=l%?311dIp}V4@Ot%^n=Zg0&$8-
zO?c(^i?pB)+4=HWjmzV~UtyWOku3yv(6?&}!Sv$`@`*?d;>7P&8Y_Vh&FhCv`sxui
zUDo3JpLO!-zblSBdOrRL*#Li)t-PUsGpTMQL5TOIe26LD*X|-)(NCBDg6DR1
z4S5xJsYdgF_I^Wlv;DC`Pckw@@$F~&?C&qtM;?SX&9wsIvZ29pXe;@u)K}aI?m_;)
zC1h``ZG?}i;7r;c_xcn`ai1e&<;_>m%&jHu40G3SryJzps!o({vdMew${mP2q|qXD
zlJR+J^Ii>cH1rRxQK3q
z+XK7cQ2)hiAVhkNx+)ELJE;!3E8Myx?4Bijyj>)$BSV%3lD9jl%(OV7G02IR6Xlh!
zH%#cTQ^<&53PLAnTz>NdHr*T^791w~^aRAm|8&}xGl*vv1cB6cz2R!zed0-PPq~&D
z_L-e>=%tau=OsIaEk)G+=BI{?Vkv9hEeIi&kH*iWyXRKp59&uLD&PtFvQ0Mn;Vj(K~U9b8F$iU
zQ2Nm4S?S$KIq_*+)TlONE0?3Gy45^790B6^PPbdXLgy+2)a1>{{**b@WoD=zJ#2=k
zgh&jkYpC2A7S!7pyvZFplkR8Ws|@eZVpg%$2bmtZzb7Cw+#hlcWc;?5ZE>pFs3Zl$
zuL@VJ<=@!~e$gNtj1joG)RmKz8d1jSGTcp1Io$7rctUAH@A6KrcPZW0GW*tLpiHt-
zCz+9JEDSo-_i!h0KC)u#xke;Uj==8Wa&{-t2|a{De2f#rExjAdk5Me`krUkUA&20q
zYPNofuW*@-_Z;BW=J{mR$%3D&leVk>Gs=vN+FZMtf)jg!QXWFR+kkGZ@xPcA>>gK`
zt|AD)nR?z}`85AGQAy&mGU~&JPft)eV{zYW3)5v)PWvDlppPO(5%1(78diZJG53^-
zKD1Ri(y;PpcyQ1uBykgSe>T?$OS5XGUfREOQfZChb6$(4gOHQa+tF19U%v;J%^rV>
zQ_XANs(}uN0mgl1y0mt?vLOiNb&mE)LFT7}E(gqxSKO%D
zQ1xH_mv>P?V6z}m48cMr!Qhk*`J!*LUCZM^&{Y
z1jV>a*JagPR=LjC$j=L?J91UmtT|1`kjgcl2MLC6W@{F#l9#NwzI{jyU%f8ouRm5Z
zw08TW&;3Tw$v<1=4M%@8vYg5asdN|5K(4e#50D>{G
zij*nS2X%Iun*!%|`2#_U?{dyC>FHjbcrW6q|9no*?W!IQyVtx>K)0IzgE;91Xp}Tt)gzO_UXlC6W*2>QGHL5=!nl3hvzzBo^t!7aWI5Ijdbf62%9MYFGWHExv-
zLN{o?lL4RO64Gug(Smr2DM+kC0z3EMRjJiQDG
z)&TdPPMTCFLA2
z$5};b5m51KT;9a3&n@=o)$3n&gAIu5`NLY9@-*NI8C55DE|nr4au4SZJ{RZ-c+_S`
z36`qUe)eOf1{P>C_y;bVl*ir1w?#O)V21ghzVX=i7THwDa~IzLwjGcVw6gv=C1wyO+dGJNO1_BZ>F6TRtm*#ugDOe}EcZPmp>ykJXeY=*(X37$Aiv1n@
zTeonDozjEAcjGad{+C)JD(aOHbEv)skq8c`gAaSXyHx0eWo~v%voDE(&$xwVq=~z
zFOO$&WO;_BfDOUwvyDF}bNCP_1tsHUMS0+t)bck>St>>4ZFQ0XOC_j9TI&`bfvdTW(
zr%4ivTh?>f^>3K`b1P`LE&S(te%0v{*jaZ{`^WxwXwDXpA(|VrnQ>tlKd(f4XK>PL
z;t|gPqDASBqMe{>gQ^Ih0q@bwsv09^dJ~npn2HP9&<_jMHAcB~i~M+N7+?SwoAQNNQ7w;pRT<
zva2^t&PrZD3H<;!_VI39uEH%OPjruiC)PGO*Bi|wK5M5IsJMabKJVB_9eTQ1ZG|@4
zwM8f`n5TO2_P@)r4j)$HiwOn}$z3?$`u4?>8=QKkDScfTqIy+9J?PuGYSR73T8QM#H=&PyDiWr@jeF*SbfW_SNU$F
zJn(Pv^tb=^o+vs3Q*^S=n(>kBWQi9(t1F(E7et)gIkv}jV^f`|m?!%HvEySgonEsW
zmi*Mprnuf-$1B%}VbiX+7CbCx$X??IjowYmUdOF^&mD=#~#=b@P&3dW>
z*ZaEkyXfHV5F|I@nenOe`*aRG{Y?IKj=$G?TJY*t9|#{6z5UN?pylS15J%meu0mDgn(|E{{)L>?W8xgbV$i@#brOhDvF$*&cYBF3%eD2r2
zLe7L)GF-Q&fyZCcSoMqF3^PH1A)tf&&B99yR+}qgUL+_22bG=48RDkt`7e|KItGUS
z!F9L!DlMDc?x*WEzijpBkN*qT3w==w+t%>WQEW}QgT>3nTX)XTVcc7E0AWrLi$|+_
zXy{OCq?k9vwv=2aHmXjgiQMvs*x!~y=KoC(`q_J0(&(5;JYQr{Hup2g$J3O>IiEnI
z(z8pG4wJc0daufO-dpSVV#P66Gz|&oEf0WZNL@?Z1^Bdbrzjy&I*Jj)Nw8ziSHnLc
z+bg@`fokEey#rEH_)^b`7E6=3_RywzJS$z)Pw%yVfVP(j@`0tYCv4bWOXspC9=Su>#?U;W&O|jfJ~#%@2F-zy!T=?qi4MC4^@%Is$ua?J|ap#N_5miNxAj
z$XxeaMpEM2(XRp&?*#G4sNL_DkkpCw5;`5YNT>|JQ{`UTYmvF+9JMr*oC}=KrTG`S
zh4eXMUu)uU1S|va1|o4^lqiiAZoqg$VV0?-$+Fm*Q4%=DAG7GjpI
z1yCtl^tqU=fEsFXs+33&_9Jun$=dR{l>~N)_jShs(HJSaOK%?e*~1F+$_!%n9F%Td
z+sh|~!J?LcaTk~biOn0xh=wNxL9j%wVA&jQbShN)l&A!*9)}_ZMm0@k=cRICrHb`
zO;?e@(&Pi<1)UHfJtOI99RVA9er_4UPCr`CkfhFjsk#y0PURMj#n+Wer^UQkDtm#H
z>VVO_c9M?KAkVoZ^vDX5U(fkiNLF8B&iVc}H=p4`lqA~^-r&o$8yYyHzB7#820DzQ
z1fRyoirjrsEs2igV;qf!2Z2%2>4(6{^MV=<)9$iFSS*60K~R>jKV&Cgbg}Uo9oA);PUp
zvo%J+u<514!BXFn$MVxtU{dI+xFAAFE_{!~gF7fB12$sxf}y?SQ}Qr@Ri9EIhE)E2
zU~IkJ5PgMjYiY08wO~}AnZ}1boRNWhb_68>Ei*No2)dP@2vW=W#U>8GhnXE<-qE6O
z2UEUc>csK3idnaVvXSh0~V?dmgO)g54x2`G{Qe&w+w<516@
zlLS6$A5RnWAqLWO682JW{mB2Hzqcl4>U4=#e%R`NXMc?XKs1`Y>?t%Clm;}X=8MeR
z1LT=K{|6RLeTq^O#BMmn0xW$H!>Ig)R4n~U=9%&&v8lLCG6-p7
zfSF@ru8^@8UC9^bxE^&g8+Cs_NAe#xIBvA5i9~x(28cz?skGGy>~veop{Z$I^J-1w
zAcD2TuR%-uTvtfnTDaCY)HbSPQasYn~Hu5>Yr8EeqKe(E9+XF)yR7+7_F<#uJ(b>vudG@M{t&6
z)PM6Wk@dmn__6n(e%d-;@iN6Pm@&Q+ZrUMFXmoxDqnIZPs&h%~b1@=QfKGhpo@4y>
z-pyLK{w*%6XUjx()x|$!70sC3k^%n2nEF4aW2h|d<+mDLGp_}Imqu;F=YYJ3a+T1x
z^g-Y|XIlVPnA7YsHky+Q2S7`>1of=|L~8ySC|j`_3IK3p{W7d?{I)Xe4eUMT17ONVxz3uCEm&ql*#LRur0U~Ku{
zM0r{a!G%u*Await~>#v}u(XMqrl*Y{wVta!CHh6K
z6(Fb6!peNZr?37b=5gV(Y_VD{PhdItI~
zI*zSyT0X=a>@_wdd7Wmd@|EEQItiV{>>MOZdx>QEiS^}*OW=Jn@z;+nyWJZy`=k#<
zXo#YB?>>#1OLY*X@6JSmng3@guv0HgBy5QzNAW7p_L7z?mC42
z9Uxls{5h!5Ak~!@Brc^W3j#_MB}lTRvrj~6Y>H(b6*0=IA-XIS(O0o!j&nV0UL#Q2
z@EKCdn1Ya-7caXUa8fH40h9m`H2a7YgI%D>#zq$p18!+mJALRz+LPZ%Y4@*@+^pFb|JIA
zD5fZqq_qb0`w<11$a=ewcETsQAd0yeX(szhq!8y{kO(Z~0dfQkd(qJbRbMK0#Y8d8)G0?h~d`YfoV`ntf
z%a=k0Tye?&AI!#@0L>Vbc#|(;=bwGMCx%1dfwg-JAMjYCS&~jDnHq~Hj`(LEc=G&U
ziw<^0-bE9&7zPzpvXVfjW<`R;uFt4)@7p7sp{1iZsJBvqW)PR761_64^n((ABTm3R
zTdK2Y%UkJ!w?b2bt^C*LYw8*Ax>1vdAW__RG%M}uC
z$fm4AEMw5gSQ9(brT9i#p6#t6_}k()PUV`jB&Xc~h+9Fn7icbx=p{DNXAiLU0k*GL
zG1KyK7)10=3=r$9Ryty4xgJ8ezB)K92z4iQY2p~)CISh%W)U)H#%
zt<{v#Nm7i2mHi?p2D5BBeC&T|xiivte4&VHFb@)kD>1l+%9H0*}+OZ$@68
z+1DtY6?2I{0p>N;SgCEPC#tP*qz1jqemh9sHUM(HKuoP))|`XB-8cQNg;j<~v%iYK
zw53*7V)s`>pR8a{eAT|nDVufLIm%3M18ydi~Y%R|OokGIz#8@XW
zK(}Q93W)z2w!5{en_5ayatDss>RVV%8DA7EH^zdwCt`|L;EET$zisGy3ON)ndx;$<
ziJ`V?wqEaY1f=N{T*1hZk@qc;VbZNA@6Z?VDgu?}Y!Ze}0FKVL3QQKSPZWJ<%29}F
zbC5??gVHODPqdJPTBY%^Y(Sg6h%Xd-77A*PR!|TM0oi5pjQ_
z17CROyh4?w#-B83?t1&FVf9z{%BODYNFF`Ei{W2I+>9;RcRVaD5&2YTk{$F+wW%Bq
zqn;~OHFDf1235FuZZ@x+b=kC5=O!d(&mJf5@jlAq*I=YEMKc?m_hd*w%h
z4ZZ9bm32QXQ0I9aO9?`6Yb;k1P^AYVQB_VQ5y$c%L0{d8a!65h6*L1$Eu@VYsFs&T
z9y?PDopK@S^0a=}lxB9vZWu(Xu3#Z6JegD()(X!y+4dJ%_Ddv2Y*AyYNJOYs{ggvN
zEWn$VAjUyQx%XNMW!S+^`%}KmBHZ0jOj)BFUP|LpnDI`GS_jFOOZ<76V$i|T2u-&zDsZ*S^0vSfKX$J2P?@R)0VbjBJzJRq_Zk`O~$
z`t>(L=5y49pPtbW{RxY5ameYrG$pW7s65~APu5_y<^*`FKI_C
zjS)pB4nJ~(s;FA#(Iqv?MIyarAd$7ftPz7Z;Q_NFXSmc(B1Z*vMy~<~_Ai^rq+<0!
z>N)EYv2G!P=&?`q6W>v%W8YAwJ5QmK5!j(qtw5)AridIhE*@|O?Om2KDsvxn5nF
ze9tqXfOX!Kq%OQ)mD3-ac(O9c*Q7{qh^TI_M%SEjSaToAy*pp18C$MtwBjDWJM}J~
z=X0~DMU+SciHD{M6$lNr_Ea)r5BcF9CedUESE#?QJ$)MBaqYrAw~AreuhZ6bPEu4N
zP;C|+^(2k?q;N3QL;{BR{v?PcVm@u;h8`FSH3?wl^H-1NCn0uBJZQ^GRI=>l%NUsN
zLiH&}S*Is(Hb%%iZkarTmaq+Mf$DvIhICa*utAyIh@rn6OISi}$Ddi@d<{A-NkpLH
z;vDdoBj>v7b5&bm#q6$8+czbHUCYOwKlY5YmL@5%P@_wpNRY0;6V+@=CP^mtht_kP{VxwjCqcpQ>G^UUPBPBe7HI_?%4MIWoIONl(@D|oG!EjB+
z%~s6FyEsMmSg9%0u>A4AISq|S2XKeXTUl6%S8>xPnM^yBPY>p()_$gl%0e!MREfe=
z#pb`raNmKZ<44On94;N$CMQsL6hHtX5GIzENw{@h$>>x6Gb{lIjjR=1LzXaMxq!6{
z7Ez9ZQ3`h`o5e`GK?5Ue^^gN=U*E*(kvPXL#zo84`}%N-=P5cJ7pM5&q}seGjex0M
z*%;4)O+I;ZQT;z_VEfD$o*QI@m)l!GTM0Eytln&Gz>2bPP~6S$j%usetnO!_mqawI
zFajatR>e6Q1PX1}4wvwF{L+h;uU|pGlH?BXhpjw$2k((DWNvcs%ZbOOnUow;f??)X
zVI9wjNC(8ElefcKe=}2!wGx`E;?-i|_C@h*Hys=PD(x!|_f70S7jG&u!ylaEHsAAXZ_J
z*aew0pL9PJQO0~9I1g38iUGUD_%-11-HE?*5!syGui7?avmHa9g!t1eyy}t1G;eHr
z>tbX`eAhM;)w^&uLV_B44|CPDS47cRBh{gt3)Iqi04{G)*+?0UYWft
zd+omjG`FcbwJoDi^|+ysJvvo~4vZuc9OgjpW6GXO-bEcg>Yo;AiF*O!*X{yg
zGoomOsN(H}`wg~VxP@+XCd^T>H;CBbNK$6i@mrTg?8~J@y%hf&wTRY~B_3CMPf|Vq
z4^KG{yTAg%_FP`99K^2*y{`t^3yL4`IMk$klb#`69{Q%ZY1jipNOBH=2ag0LjU?ilXnT1mXgiudvvh^QRqE{NHt
z*8mi&8Tm&R4nTsI+-X5c7b8cjnBZwXjKQTrC}kui;W}{!=KbijWILIl5i-g|E*J^P
z&j2;4c3HS9*cYa
z(W6J5*C;#5Ry^bHT4>s
z)JI%WttT6OH&b(CbY$+rrG|
z`9uR#3p>@&e41m#x_i*#t5gTak3P|9*4y>Pwrz?b)T%bI8Pa=$_SVGq(PjP>?-LDCsr~(Dk7OO>bcnpV2X3;b{2G72BC+anCtgZu3-wC
z9cgG|Y4y^G?Vl%PSOns=jO^x_&X~sjkSc&27Zo=(zKxGlC*RyiSL!z+eg??Vm6H$V
z`0H?1l}37Mvc?Fe^%NMU%$QWHn$b#UpG$Ruz3wq5!Ym4~G`zg&OlkANw%6OHIwGGE
zyF((I#Ed>TG88potbcuXV|8)xHt&ZI>jN
z#TiXc2NloB7PKOI`jtm>#8?+=NN{ys@eCsA|v_2;}}78kxKzc#LI6JFD>sIN0l#0om}S>eEYdm
z-J>sN*-}DiD5+$9pSyy{66^GhW54Rm2mLNq6O3b)1{S{5F$-SR%vHp`l=z8}{TRnMP_hA8a;;LpKQXNf_oFp=(GL8WKuIO;_2rV+Ro&+q(5Hz=Q@$EORf6@^
zL=kGY9R*A#o<5R}RtWK*xfMTi-exe7dMRP>=9><5vijgG!$uaLXPw5aMbC@F3@eTj
zDL4@p!?t8Q99x{#+)gP+X!zg-Y51k5B?uo`E^SdFLAw)UXuV{$v+Fz
zvfA~%Y1_ndCDw_*mGa`g$D&)dco;#cNDuKex}7d#v3OlwWY!g@@mtziq^gQ+XlJwB
z#1n1GzF8NU!{6tZzw`#pHYyPyny7tpWp37sYl9sSn?0l;EEJ-ErGxeS+X0nDm>
zyyDs{Pj_JNDGUKPkOq4lZ5UHks_=6-AM!c@B&9K9D(k;0riqife6EVr(1
z;Yqgtus3ZMq+&X{1sKKd(gf~&B-_8`G?5uKoqv&ID&7^=>FQUC_ED!G;{DmOn7NlF
z1KMq$=02X&mqV^&d1Qh#gCq~{xezh(sQx2c>4^^O^B{-f%;Zp3*0qMy057f^u48Zr
z0)Vbr)k^80xoPVdk~Bt9|B$bpnASI0v}D|G#}2B{%B!}DdbQcKr-gsx)2zv>D3QpQ
zRF)g#PsLsALr})k(3ieh?LF;x@2FvSQmDs&V&&a?M(cL!8Dz!b+bz1*LZx0J2Iq%S
zh1lH?fo{erS;D+wZ$c<%=1zhl0d{(Z+{{yzYCh;Ed;QKZ=IQP&q}Gzk^{i38fOc4>
zDrnB!6XH%;A2nCp#n*qi$-!~Dwru|`XkwN|<;B!KTANhDl8Qa#(3K45dVaOE@u*zD
zSE#k$#Qgf`#9d+J35kV4ZujrsaZ}CDWo~oN+bTU%;Kll?>176AZ&_bUJ10)1c2L@
zPpF6ilxz|J5D#-)S+X
z6t+XUUH*F>(DU&jCrf!-mRZ`yl3)l)JQK5-;P%m6>}Y+q9QyQRne@O0o!
zWS94M;I}&YxE5sF_U}-|^oT3o-r>AfPze^GF!N{=m54FQ
zL|P~wfCsJbpXi%%zkcITFhH0an+z))h^GF`Z~)Nk+xe6)Nc|dustVovhrP#7sjmKn
zxSIhp?xI^jaYt^x|Gh@*f4C!A-DR7mQ{yA&{2Okp!CwULM@#qPNy3&3ns@dMOoNOo
zCYZ-|6Q0a+lGLt3EJ$ht&vii2xQEe+=qm)nr<>!SfO52f1Tlm|
ziLT`?$o-FCW};r>s>w6!+?g(hk+izP62w-c3%qsnpOnR(|hS)7T&4G6zZEuH~X#K@+*4mu=&)46|jUvlat{Bh~+;N
z1Dw&o1K3rwn&|=J5VME}lh47aM~D#ydUJx=VTm%sY6uu+^bw${e~@Ngwa$r*fW`bC
z4gy%T*xWtKpR$}KBQDi*6`4jkmNs~|i;|c46WnD0GmgOB?nkfqLP5OBDbO?+R?T#<
z(+YpLPDr0dEb8sW)HH)n9F<(qrZ|FV`vf?JCel5qaUaLCaAQS1Hv9tN*Z$^8$QHYyL(vsz`4M?
zybxxe9h5d`04RBG=H!y>2%tfWYhebgR0Ce;ToOYXks3E{w)+@qZ^Xf4dwMKc`&G=a
z0|V2~h;|LdDlf(h8YAmhdWr#SauOFe$AGXAN1*e|j_D^lNZ{sU|MpDxT4~dDwGA(l&l*c_yCS*;Muy3w5YO7a@z=
zCbP>_A=TZ`<^%24&ZLXt&^8H?+k8`Bed!(79zJy`?IIhWpQ2maS@%rawODci4U!)+
zteuM=-y3*5NYMfqPyu3++^+Dl6+LtY%sy8zs9bKPnM1gx(bcoX49zc}Tp3>yQZR}T
z`QNYf7cmo^UUVg-g~MuZ(oA;$q(y=~a_<cY3s%ukc5M9jhX;$MS*0LVS?ylL-tnWb?
z-y;fzSaNr;a079HwMyjTp=0EtOrF7QlOTGUhLDmlFhSN55s?PRr=hReQ==;WtO&-$
zVD`X}QID*zjfzMY{y=0AVwctrtaWVy)E&ap>{=n{lcr)1(-dW3WM~!zc~gy~+OLY-
zU1+De`jVj2B^ilMX87LY1;yorAd)~+8sDzosW>CVPXM1VKOh3q41PK6bN0Blv)@~H
z9=BwNpRXMZXlMg2Cnycz$Hj5P;=Mh@kb^Io%TP-nX1FjewHmHXVNrb@!vvX3^j{2
z0PIuV{>kKdlb>!|mGPp$HALvxD$MK-ILB5pY8Jrv_)#vunoNmYfxm0fi>%}z{@YAf
z=WPc8l)kwJ)ZCMwB-FL&mK7tO<~=!h7zG2McvD=2LvxQm6^G4MGquFqjQAHQ;@Zyl
zRkbp1Y5DP5h6w!zyzWG}tE#v=6WzZV)?aL<^Ch}N*ZE-In)^=~BTx7BuDc7FqdgCK
zg&X?HjA@q6UDDwQSQhoQ@Fc4GPIhiD
zZ`LwQ3&TsbO83CPw<9pz*}gXoK3u4I+TgVy{tepR?wiYicVGpBjB{Q@XjUG8lYfb(
zP}TG&la!?SZ};+MLZmq}-kDUm0*dz-b+I-Av)=etuN(9C>!XX$vu;PcmX}}aPZMyh
z7As_uTrwW;RWkPm6!>beM+q>cxW~a>d6h6Ma#jEy%Yy}Yb)3|#pm{nm~B6Og$Xl1
zR>&V$>!bzwAE6l02h>P`brfgC{MLY7faTSs^R={#k}+dr^g5b0yo7^lUJ_b{IPc2utUt)+CnZqr8aMtM5hira!pb
zIAn3Z{`M7`*SQT=*=zkg`thU-%}JK1fKLGrK8U1ymt!`&1^m-rG#ja=o=8?m3z3Kl{OS=L7TS2Oq-Lk@(bUU+o)_Jx)P
zHXuXqbBX=e^o#D0#Tcu*wP-Eaob4o`zE>$nZS}{TeKhGpliR0rweEiMFAeqwt?dE^
zZbryx^O+3xTDfE`Rfu%sz5XENE{?+ZgM{FkaBp|yP{7t>SNK{l$7NGTI1kGDWbkyZ
zDO;ed!TLdCnVW{s^THcxvTuTNwAICPl}1zv8Fh}lqFYrD4*-S2yoGoH-Ut|=En8eS
zS2;Jbp#6MwUs<+6TCM@Ar`IWWT5Wx(qJe=twpUSOm&RdS+(PP4FyoRATsF*|Beur^MUfX#Ckt1?fY-nAgRmu`{2WnSop)(Qyse|5qb__oA%=O
z0`&%ThGLQQP(i!H@H+q{4*?^H*>4H#0;WaI^HyV+Vw}5G4NU&O?b503CD)Y?0=lK`
ztm)`qMcdGWo1j`5jF@!#v5D>^crlHSTGex9uiZSm*$vvkG+#Q2T20HN$w$p5Jk-lb8YUIsG0k%+!E^J;~qi?NjLY9N~Pi1s
zJ2+QGMb$L}hW}?jHuQ4A_*cJ)>OUQKubhd_?x}!ENn8_=OT2PL){&}bvT?&P!{~-~
z&i&j%$`2jMwpTM6k292S%FA#0@J0K2WId;`SODeruVj;ay>&i;<9P|#1Y$ZZ@^w@#
z`|ZnCo?gepywWeIw^5ffM6V{)e^=(a{fX1bRJQSDz1PlUjOQ7b#RU%Cs(#l{kejJ_
zl~#8l_TiXdS6Z}l?bSk$}MR&DGWW64uAD}5LP29FUwMI&I$
z()fkHZTglbBSw!F(|u~RlR*}6`46;L>@_s%wBm`FgJ4z-ksj-9k;#Zku|AFGig@=>
z*07b}7&1U;jws?(bc0Kvp_3crlU24x5h>SRJT-`!lPjv>v?ttQW$JDGDR
z$B-Ecz!k+1UZple@m5Xp&R1iozzt6`&u`ib?0K}?iHwLVFdY~5)u)B$%QdG;I2WI%
zayOGd93=BV75xju#t*}xCVFZfO)<|M-0D&stKH|;aKfWEl`7~4<
zV5y^J-iO~s{_pRfUne?lC-NYv>|`jmluZ}VU1q0<2wU9&%Vpq^+)L~Xh%-dm_+Ws?
z%b);7%G6E+tU4uxylT64%Ve>ZA{ge}m%Pwfm<-N7Ddkv;LNP^w5
z#9p1ig<~?rfdwpf#_+kT>$&?x_pm5)1oZ~f+{r;Qe$>kGB5|aU>NO!ggiaGANuv^W
zA!4ZO2>{Y8h?5?wuKE38!EklAC;LuQf?-M9(U3g9hUr=hof6_Z?3(lWvlu
z;V$joo5{Q~;NSU{-DAC#s8J8ly{;>36m6T-1oRS7em4!_^jMsPOo&x!8pmcb1FhpE
zvK~u!225K!+vSytt<-x|>5GxQUW77hO=ZC2PUPu;kcki!xW>}Jcj1DApmkAn9Vdc9
zsmX3}6ZI%D@XMc5O(y$4>2WWI5@}JCXZA}#85v5uaFLviy^ZaW3Wx3Bho=-|HTdw*~EHahA1#1ru$$=A)vL+CC!8LRse
zUzlU^sdYTUCURa7n?bf{(thvCRI7k69YQ)R(ZHfPsb;2)Vzv$cD*ai
z14ok1sgsusrxc-#eza;4O|xR|kyVBTL9>HLGm6$PC#mRsXuVqoHc{tj4sT7opB)ex
zXChd+2M}ZQh9{wd%jKoUB>Sn_Jo1#6g3H@YP1AxYgAMyqOstCO
zNWuMPQ}@kDGadqXh{u{&%VM$Z=uKrlDEz)H_n{)`0TB@{^XYT{OILrhwR3$OA6{v76v4J4l)Uql(?I2x~YWpjjs3^
z=~hO0ouw>0NrUKf{oo9ztt5HvTa=Nnm3;VAWfilhp@xyFuGa(9<9DrWbR+Mz0>^F$
zNl%>({skrI^D8GD*rf$MQ^*nqyp7Vun_$Vi;}`G!o4i2<
zTb&~0F|hMj*C}Bo4~nB#tAs$+q~RTkmVR=LVw@nO2OfQ*?<2s!}lOAcVKOIA0iGYcPM3AFZ#a#%dVB
z=q0G~I#`KkZ}JBHBuTjz9@Rig&tfee4uFT6nC;vxx2Jq6VYFIzbLCG)pBHJh#>KLo
zDsW};yxRJR3+f+jzgx8Fb3Qs0=V6emR*GvW#A&{HqRCl-5EAe|ZE*Emj5hdeQ_#>y
z5{F19sHP$&ZS%3o@%wWZG4kJ9}re9*2n$=gS%=j*BZxu>q9TdDv8hk73HAy+W7!K#`NHsot*@Pi^cq
zeTol=Oir`E)1c(pZlzd}&_)Frt`?-!NP)>KgvIO1=Ynip_}ln%`F-D?6_u6h!W{iWeM*N`c=6>BiCkc2(J;*;LgFF+Y_`0{=rYGPab3
zETS^jFCs+Zp*Rv^vK+xrv(#TWE4WA<)8oR{R*r9CPveq40+`lON|!*jR{a@QB#r4?
zMd*QC^O=e#g_|D-znRRi=~n5E-^l0`nAEW!(=o5KEHz6x~Q
z2`=zzEunWBZO8(yo%GWHNXpvSMHgimi=4T-M_6^946GF}UyY&pN9^gsBy%8P=xV{M
zimQ23OJFN=+rDALD=~mGm7n;uX*`T_{imzF`uAI*ys8=r
zp8)S(1^;ZFJL0UWJbmHjz-{YB>LhQucLEVYjQ4mGLwj9?^ZA9pC=sfMpSH4O4@3q6+Tjbj$?xWTO(sJX2pL-8YlR=m)PxA1W9yF0-1
zf)2MIjgGI{oU52@%N}!fn+)6me)Dm(3+6pQpLbL~f7$nye2AA^f0;sy!?zb}D9=|=
z8ZFn#7H>tz(?&`@oC7$K+vLErFC_fsz+v~vL|2B@;klL?$45@cKq~<&^|WIHB4!Iq
zNxNk0bF@n%Hqo2J`hn*|@wA9wfJ*@U?m|}9_!L(#37S*GuSy5>4~$#b4LIt%Ovcl)
z*ie^4l)cW#2x`Q~l`u(uVsq5zCT^iLTAs7PY$$uLH|_YtwzwXHW@x6Zh%a?QGN)O08yXvI^;JSkpFke
z^VFMEGnQ4&h(!CiwcTMR75}(C*j>BoNKC6@;-
z#M5V?Hcd4)hK071oALQdiL*8Kp>xM9h8z?H#}6REfWRaa=MGBYMC0!Un{^k+`SQ=|Sy_
zw%~V;42${RV}Zd#3ajNcn?&Gb`Q^RAG_+5&pbge8BdZieNv6exze|23AlHGf`slg+V31|o+vG`G!5K?z`SUj8~{zQ*KNN$+{7GTcM{uU4rX>#R%M1&c-ZUWE*w-J1w|
zVY)BaZknMpRvYdPfY8XNj?Y1u&qA1x3~ew39hw|dUnQX`=T4eb4V}`JtjzSQ5$K1nN7J`&`hH3EL5UW)-9F2^SQ^G?W-Xm-feaag&jWvvEdA~
zpQaPE#>zf06|?XBuY`D4?X%#8OycYwCYAQ$XCZG7lkZKgO=3n9*F4-(uz6*V**+xG
z@x@BuIKqXKKK!VhEI4L2!b=&ku>HfZVUt&LAvvjPPV|gRt}{-NmW%HBuij)$e;NHp
zpYEsqqF!60zB24g$H=9y2;WU`j$liURM@|N&u!URv@k3~#CiJ%sNsJt+|K2ZgK4zI
z2q@d~T&722)V`?>=?;8)+_0>f3hZIAks%=)kS?Ui#-Xw3-p^AqO!H
z0fw0SJAp6skQEGWb?mo0!|2M-oW}OnNORfdMRvF(MG!${jhr8PbKZg0ni{GOnluMThdW)$
z59ka=W3vqAu_7cK&-SIAT`Zp9PUO9kwE9BsLlpb2{dHYs0ei~H4@uVMrw!#EKzKR0_0gF7{(aHh+qYg>I2uoQ9*lNok6h@P*H!iIx0e?9pebii
zM^tv`tsHZH0gO2A)nE8+r}!iDOU=2HFTz~4AaDVtUqtNkAfN64g7MX
zZw!+Mo&5lqadum)woq&rDb&&3(8{eE{o_*t0Z-2!zLrSY$&9EjG|>LF{bcr0Yp5I3
zv8)3*kyqDXF@8IZcl)CaJKMF39L;8>S
zB|oGXIh53m#K-@53H@(B=1;uw&nhw|p5N8VL3fh>{7*hqu>37}dTZD1A1rV{wl{Pq
zq@qTCBz|&{X4Ui!GCR(z)#&`pFS0D>S(upGg$zb5pi@^m_Sq?^v{4b%#LBDi-AAiF
zo3CHnN?`*4kbeEl)!t>qh69($g6co!nPQl^y1+KAp_|X-7LyTcMzcE)w5|f=<00(<
zgi{JOrh$aF)+3^Cvl=<+Ty;;5<)+q6F>)$w+tYdEpGbcr_vP#a|I@p~?04hG_VvAK
z;7hPoUS|9*We8W+6QK6wc^_i7Imb7
zYo?%O_Fn_AXR=3SnnM!r!$+KdMLe^DF#_GC?7hT7O3Z#{(={5$WSVy
zv-jZG(*VN!;@CYOg)7UWmgg{TS4UZBhoLcc5(K*Y4|wx>boR!$h=10WRjJ8zALj9t-raLIyYUrA`6D|
zWC2@CmtH~srA*N>F}MY=QmzCg9u{KFcqWybmSa7Bklx3Uhn=<93i#Xc{%gRMzt4|d
z^S}A`FIVKui}!;s&W_wuJc3vms1up6-#{>oOy?7LufO1*E{uYw35sV@fz{9)apeWy
zBW09B#_qAibgG6J%osA$7`F5D*hxLv`8js-46ljK{|XmazC5rh_&(HR9BKtNmC4YXf$x^x6BFC_6akx8WhM7}%cj35;e|lPM
zpRlgJ)y)OwX|2R0H8JvtaMp?V9duh1bcEIycR!~0Vmed)eUDH@fy0|oQDwUwIH}sJ
z7|nm5JFe{K2bYZKP67Zy8_aMcZ6vk*%QDgIidL~|mkvzibhS<8M28x?Ce~BnCRdrX
z`hL>As5|V}D$y1$%YP-XH-E;iNfqba2bR_UT~z)}`Srp7djtRd`tjp`cZ%EqgqnPi
zK-b6sSZGRQfESerAA%`Vu!i_G9>IqNWyK=FC*cd@JpANW{Ov`1ZPhB~-X@_{d|!fL
zY8gB1aaNY@iWLYUg&@}Cd8+bN^Uq%1t?k^q*w9+UpOJwVS}{@8xfMarug#b80NXBC
zRA)Bv*L3(doA&kj@CboB{MBpsO07kXN(dVeCgT+$?RcYR+q(t{FPpXgEzeR;xg%(1
zt6VGPzmFRu!EHoZ+WMxDDX+k5_jSMekVPsDXO$VxJX>6u
zD)&sizI&!dxu?Cb^PImR
z52?FH|FD{86`C({^7r9#4p9h#rwqmjk!!q!UMESuiQlTc;tf1&GX5}ydvLMV;;+>n
z5GqGi2e-(5WseG|QaZc%kuJ{`s!72r4D;ups6LCBN-g)%o-nnW_q{JO>(V?oV%Jr#
zJSwIO=y&6QC${`9V=RrBYMh3ZFcrlr619PT>w_JCt?8WqWi5oue<3`R4)19k(#jWh
z>n+ufUC+9|lhl*p@y#s;h6xX8|J3XtCntnPG@q$T%la4n!|n6T_N@L#g5#qTd3IYv
zfgg)(2c>XxVtL+eZW!DZ0vYV-H%FupQnuOzCAx%`hQdKL@5?{mtN8BuMWl%!i2?3w
zS7t19d;VwEsoWePIeb~1OnGgyym<4?tJt@dKi(*aX(v~5S)WBpS@ta*Y2fRC`@(&o
z_ECP#hQ_XrAE7WthDvu20xlggo*Pv#uM(AEGQhOE4+#x^Y+!qL<0k3u_q#P0&Dm_a
z&7l3Iy42F}DW2n?b(b-32oHRQF6?6WyY0LLv;(`KC$gD
z?8|(vU|uY0o_p~1wcIY%@`NWWK889!s47}C8Y1kh+roD#kw^1R;H6mf{J5vg=ju<#
z0j3Jo!^AOh!f3Ff;@201T}5#P6UM#*FY@XAH@X^4v%piBmJ};I58Y&T6n9;k+|I94
zu+xI4zTg2CBQeL`G$6pEJmAZq;kvRI8?9xvxue9G^XveuqZ8%od7l!voX?1X|b;8%?Gt#g}_Zf{u3})i-YZ^BbEi4|+
z0i;D2lefL@31?;I|997==tt44NNO+bUtwe(xC?QUhwd+@bJg&gTyNc_pWg
z5O*2CW*}XG{;10NHVF89QFTld;M|Hh+dgjz-ln}%4m59dLErPino;j}ocH5oC$l#d
zKM7L6s;*Cd^uNpV*9p?Xh&o`#{GQk?nKMX(^DcXmZYRw76luy+_p&qG0yQsdxxOUS
zvsJ`OT`i1)z5$BtMqwjidsjwX1*)
zkIHiEt~ANVLa4KIW~UX}OYoe>Tn9d5CsekmIAgnpr|_DC-?;uCDlVhyU}GYU-556Sp&o>j#au&rGfJmvR(*W>ehAHjG-xI9ujb;
z=8;X`#nj|=1Onp~RPBU6<+t^%#0~@uf$XCStrBu?&DR&dPtw9u{|1;b^qzR`5zArX
z{4Phsx8+?7J7cHn{-r)~+%8e~YJ19gL?;D*f2-_dr&^%4)Y&18(p%Zo!sivWqqg~~
zcSDEyqsPjiMlbL@8N?tpCy84lUHKhT>f16WS-_^mn%hFWA3|i!J)dQYoJ`f3%aE(q
zW9*Ud|HivIG$5r67VFfwnJ;P8nJf9Z&O2V?ZY5E9YL@T%rDvkv+@yV6g}%^kx}o{y
zi=}GEFiUV9JvWzGs#{bbe60gPV?M(RITsgG2q5H+6U_H(5z{IPQPafl2Cq)uJ^5Z)
z>^}mmydar>Ko9$AIBW}4kUljDSvvWM#>5VZ-hf+
zD7bgyM9M9bmhH6ArssQ|G+Pr#f4`J}T{?G{VcZl_0=Bf)EB!LBrW;E6OS;f|yX4)n
zb~S55O*niLet~-%9wlD0y%Fj)gONP`3Cw_=xKbLV=`3+aXzsyjsLUOcuROo-YPYNv
zO|V){&#KJ}Vf5^5L~&ip$&E8GUEmg|L+6Y4muB0>O(4&r4F~4h3N^(mT2FsokxAH4
zukD(;l$ckSAoYk@(o##izB8AWjFRjo1R~U7vX)7u^AUYN8&Y{zX9E
zcKys@_>&zivjltBk&~)CCTU7k#7v+nqKQeXSb6LwQvw!s%onxSQfiB@7E-L@%Q=TY^6>JZ>zD;%DZBz4be
z%>GT7&}%`Ub6VX&ms&rz`YV1{-+1E!c}B~6k&GmB1bFi*Z74e(*m0Ph5}Tv
zCwRvb#%`|f4IF2XeilB-(a%KZHWS3w`uXu2vy|W4Ea7Ku3fde)VUd@fI1@iVmocs&
zD2f12z-i9V8ws>HJLPLfzq5({gHmQu7D0NeJl%!_wDLt=fUmk>vb{2vx(5(71d9G4
zosZ|l4CJ^{hxQI>XbB9@WSyT{>cZryz(Uo|7Hj6QB`qR>AIn81cDUMrH#AZCOUp3sR>rqQ>1}R;)K^8G
zVEe^*43KU2meWnvAA@dct10W~P}V{Db2^jc;LQ!PFzaJ~s?nyZ3jE!X!ikbxKUP53
ziBrC|k066aFBm$KzvnIWOd*Z3-(u;wc0c;59IZb9x;yB|OawNPc+j8qKdiQCgG}Yn
zI8=Z*n2m(djUQ*?Du;U3+wdg?$NyIU^wN3!ip;%RTw|xscgq{5c4|nqEJR_(@2eb7
z%nBW-#xo%iC}BpRfJgP%`@V_l;K&p#FR;uMM3&N~E^qn-`Kj##E(d|27+yn8s3wJf
zM1p66;ir-NK}u=B^om*X44K@-d(Snc6U%Apdo~jnrr^mq4(7O_0e1Ay`mVcI_36*$
z9MGy`U44Gav`i0VhtwJR&e#mVAn1O+zt%vF0oPCcHXb`5q!@XbM@IPqXZ^7Rj$aRr
zWsYBv$9j&xf%AFTQ>F8=Vi7|oMi8;_{mPVjC$=@6IK=Dp2#^(
z&(4}DN8m!K4SM>1OMkZil$aKR}B>uuOkdpauIT)!w#8-N!iBVYSm@}
zK1&U@j%$l#rdmRMgA-E_>y}pLWP2`N-yua+868jK)w#W~7@?R!RXbxW_^D=?Y?A`2
zVJiQtWTMPNR8LMC(zi8k;kka<46q7cZn4jE{H2Qdg-ix)YC9;!ZQg#SN(wizRgK<3
zcSbgA%naeUT`cEtFjqv;G~Ja9H|V6QE0cVpf^p%ijp>HN7nF;z*R56DQ^TwbT^e{-
z_kA$<)lz*k`+TfAd;_Nsi5fe!-(P)oNYl65XoJ7gIYrkO6Yb)?q#dy?Y6~)RG&|l>
zU}mst;Ay_aS~szZd1y*Fzd$l72(hOw6yCIOUzfvZQ
zIJQ_L+ESt4c>fcSgSD(A=Ummkbd_tOS15UH7^ug0lG8m1x3L35;R_!Spzk)k?#h!f
zzmah*?$3RXbrWm`d{o@yPTLSHS~>7se*|l8;cm(Ich_R{*b#(yHM*#;3mO(BXa;+=
z6LN26o2!wLqw80@&ShhF2={NdVp=+|>VBsCBP~ntP6?l4W1X|(WE{cr6d_&74^&AK
z0?%-ER5^}c(a60s|rpuzpw|0!S|Ue&Ns^3r&;E2$4z{JUpf~iul2QEzV6hd&gm_f%=hw>@-rG=URay
zNY+q)o-`MP`vR_o29ZsEoi<3U)hL{xrix9z!*5ltukieP$=TOf
zi^Fk)7MnG@R7FoeOY!&s_Yk-g4?S!b-L1&(bJ80hrYB3L@bq9Z>vv44P*S|}(#1&l
zec?8SYsZd(N~DG-^|g0T0Z!XCH~g&Z$Mc3E*J782gYl~FL+OtLU1;CxjT+-f1I_z6pkjqd&Skh>xFlz`Pf5V
zZM!fe%$=);^JTF0hd$SOFC+aqBEdBkX=B#TBPWb#7Rw{51aQ(xm6u%kwov?fqfDTs
zir;mWmG|7k}q;D)~rz@C`I6yvg%1>Qt7;u3TekbR>X`7up
z$jy2b#h2=}8nm<^ep{i7`H9#;q>t1jQ_pTFz1R;6)SfWd
zoDmMTR%Kyt;ZknhBzIm{0Y2uajlHm&{#B9C7%SU?d~;YS@E^$;0|BD}f^4y$)ey0l
zUdp$XOd>s_=eO