From 69ee88f49bf4fea4113e226d88da390e64da56b3 Mon Sep 17 00:00:00 2001 From: SWHL Date: Sat, 11 Feb 2023 15:41:09 +0800 Subject: [PATCH] Add inference module support for onnxruntime --- funasr/runtime/python/onnxruntime/.gitignore | 3 + funasr/runtime/python/onnxruntime/README.md | 61 +++ .../onnxruntime/rapid_paraformer/__init__.py | 4 + .../rapid_paraformer/kaldifeat/LICENSE | 201 ++++++++ .../rapid_paraformer/kaldifeat/__init__.py | 3 + .../rapid_paraformer/kaldifeat/feature.py | 459 ++++++++++++++++++ .../rapid_paraformer/kaldifeat/ivector.py | 43 ++ .../rapid_paraformer/rapid_paraformer.py | 139 ++++++ .../onnxruntime/rapid_paraformer/utils.py | 371 ++++++++++++++ .../python/onnxruntime/requirements.txt | 5 + .../python/onnxruntime/resources/config.yaml | 30 ++ .../onnxruntime/resources/models/am.mvn | 8 + .../resources/models/token_list.pkl | Bin 0 -> 51686 bytes 13 files changed, 1327 insertions(+) create mode 100644 funasr/runtime/python/onnxruntime/.gitignore create mode 100644 funasr/runtime/python/onnxruntime/README.md create mode 100644 funasr/runtime/python/onnxruntime/rapid_paraformer/__init__.py create mode 100644 funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/LICENSE create mode 100644 funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/__init__.py create mode 100644 funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/feature.py create mode 100644 funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/ivector.py create mode 100644 funasr/runtime/python/onnxruntime/rapid_paraformer/rapid_paraformer.py create mode 100644 funasr/runtime/python/onnxruntime/rapid_paraformer/utils.py create mode 100644 funasr/runtime/python/onnxruntime/requirements.txt create mode 100644 funasr/runtime/python/onnxruntime/resources/config.yaml create mode 100644 funasr/runtime/python/onnxruntime/resources/models/am.mvn create mode 100644 funasr/runtime/python/onnxruntime/resources/models/token_list.pkl diff --git a/funasr/runtime/python/onnxruntime/.gitignore b/funasr/runtime/python/onnxruntime/.gitignore new file mode 100644 index 000000000..c97db36d7 --- /dev/null +++ b/funasr/runtime/python/onnxruntime/.gitignore @@ -0,0 +1,3 @@ +**/__pycache__ +*.onnx +*.pyc \ No newline at end of file diff --git a/funasr/runtime/python/onnxruntime/README.md b/funasr/runtime/python/onnxruntime/README.md new file mode 100644 index 000000000..d0fedb9aa --- /dev/null +++ b/funasr/runtime/python/onnxruntime/README.md @@ -0,0 +1,61 @@ +## Using paraformer with ONNXRuntime + +

+ + +

+ +### Introduction +- Model comes from [speech_paraformer](https://www.modelscope.cn/models/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/summary). + + +### Steps: +1. Download the whole directory (`funasr/runtime/python/onnxruntime`) to the local. +2. Install the related packages. + ```bash + pip install requirements.txt + ``` +3. Download the model. + - [Download Link](https://swap.oss-cn-hangzhou.aliyuncs.com/zhifu.gzf/export/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/model.onnx?OSSAccessKeyId=LTAI4FxMqzhBUx5XD4mKs296&Expires=2036094510&Signature=agmtMkxLEviGg3Rt3gOO4PvfrJY%3D) + - Put the model into the `resources/models`. + ```text + . + ├── demo.py + ├── rapid_paraformer + │   ├── __init__.py + │   ├── kaldifeat + │   ├── __pycache__ + │   ├── rapid_paraformer.py + │   └── utils.py + ├── README.md + ├── requirements.txt + ├── resources + │   ├── config.yaml + │   └── models + │   ├── am.mvn + │   ├── model.onnx # Put it here. + │   └── token_list.pkl + ├── test_onnx.py + ├── tests + │   ├── __pycache__ + │   └── test_infer.py + └── test_wavs + ├── 0478_00017.wav + └── asr_example_zh.wav + ``` +4. Run the demo. + - Input: wav formt file, support formats: `str, np.ndarray, List[str]` + - Output: `List[str]`: recognition result. + - Example: + ```python + from rapid_paraformer import RapidParaformer + + + config_path = 'resources/config.yaml' + paraformer = RapidParaformer(config_path) + + wav_path = ['test_wavs/0478_00017.wav'] + + result = paraformer(wav_path) + print(result) + ``` \ No newline at end of file diff --git a/funasr/runtime/python/onnxruntime/rapid_paraformer/__init__.py b/funasr/runtime/python/onnxruntime/rapid_paraformer/__init__.py new file mode 100644 index 000000000..2640f5f5e --- /dev/null +++ b/funasr/runtime/python/onnxruntime/rapid_paraformer/__init__.py @@ -0,0 +1,4 @@ +# -*- encoding: utf-8 -*- +# @Author: SWHL +# @Contact: liekkaskono@163.com +from .rapid_paraformer import RapidParaformer diff --git a/funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/LICENSE b/funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/LICENSE new file mode 100644 index 000000000..261eeb9e9 --- /dev/null +++ b/funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/LICENSE @@ -0,0 +1,201 @@ + 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. diff --git a/funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/__init__.py b/funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/__init__.py new file mode 100644 index 000000000..f9cf27344 --- /dev/null +++ b/funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/__init__.py @@ -0,0 +1,3 @@ +# -*- encoding: utf-8 -*- +from .feature import compute_fbank_feats, compute_mfcc_feats, apply_cmvn_sliding +from .ivector import compute_vad diff --git a/funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/feature.py b/funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/feature.py new file mode 100644 index 000000000..fb5f9a552 --- /dev/null +++ b/funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/feature.py @@ -0,0 +1,459 @@ +import numpy as np +from scipy.fftpack import dct + + +# ---------- feature-window ---------- + +def sliding_window(x, window_size, window_shift): + shape = x.shape[:-1] + (x.shape[-1] - window_size + 1, window_size) + strides = x.strides + (x.strides[-1],) + return np.lib.stride_tricks.as_strided(x, shape=shape, strides=strides)[::window_shift] + + +def func_num_frames(num_samples, window_size, window_shift, snip_edges): + if snip_edges: + if num_samples < window_size: + return 0 + else: + return 1 + ((num_samples - window_size) // window_shift) + else: + return (num_samples + (window_shift // 2)) // window_shift + + +def func_dither(waveform, dither_value): + if dither_value == 0.0: + return waveform + waveform += np.random.normal(size=waveform.shape).astype(waveform.dtype) * dither_value + return waveform + + +def func_remove_dc_offset(waveform): + return waveform - np.mean(waveform) + + +def func_log_energy(waveform): + return np.log(np.dot(waveform, waveform).clip(min=np.finfo(waveform.dtype).eps)) + + +def func_preemphasis(waveform, preemph_coeff): + if preemph_coeff == 0.0: + return waveform + assert 0 < preemph_coeff <= 1 + waveform[1:] -= preemph_coeff * waveform[:-1] + waveform[0] -= preemph_coeff * waveform[0] + return waveform + + +def sine(M): + if M < 1: + return np.array([]) + if M == 1: + return np.ones(1, float) + n = np.arange(0, M) + return np.sin(np.pi*n/(M-1)) + + +def povey(M): + if M < 1: + return np.array([]) + if M == 1: + return np.ones(1, float) + n = np.arange(0, M) + return (0.5 - 0.5*np.cos(2.0*np.pi*n/(M-1)))**0.85 + + +def feature_window_function(window_type, window_size, blackman_coeff): + assert window_size > 0 + if window_type == 'hanning': + return np.hanning(window_size) + elif window_type == 'sine': + return sine(window_size) + elif window_type == 'hamming': + return np.hamming(window_size) + elif window_type == 'povey': + return povey(window_size) + elif window_type == 'rectangular': + return np.ones(window_size) + elif window_type == 'blackman': + window_func = np.blackman(window_size) + if blackman_coeff == 0.42: + return window_func + else: + return window_func - 0.42 + blackman_coeff + else: + raise ValueError('Invalid window type {}'.format(window_type)) + + +def process_window(window, dither, remove_dc_offset, preemphasis_coefficient, window_function, raw_energy): + if dither != 0.0: + window = func_dither(window, dither) + if remove_dc_offset: + window = func_remove_dc_offset(window) + if raw_energy: + log_energy = func_log_energy(window) + if preemphasis_coefficient != 0.0: + window = func_preemphasis(window, preemphasis_coefficient) + window *= window_function + if not raw_energy: + log_energy = func_log_energy(window) + return window, log_energy + + +def extract_window(waveform, blackman_coeff, dither, window_size, window_shift, + preemphasis_coefficient, raw_energy, remove_dc_offset, + snip_edges, window_type, dtype): + num_samples = len(waveform) + num_frames = func_num_frames(num_samples, window_size, window_shift, snip_edges) + num_samples_ = (num_frames - 1) * window_shift + window_size + if snip_edges: + waveform = waveform[:num_samples_] + else: + offset = window_shift // 2 - window_size // 2 + waveform = np.concatenate([ + waveform[-offset - 1::-1], + waveform, + waveform[:-(offset + num_samples_ - num_samples + 1):-1] + ]) + frames = sliding_window(waveform, window_size=window_size, window_shift=window_shift) + frames = frames.astype(dtype) + log_enery = np.empty(frames.shape[0], dtype=dtype) + for i in range(frames.shape[0]): + frames[i], log_enery[i] = process_window( + window=frames[i], + dither=dither, + remove_dc_offset=remove_dc_offset, + preemphasis_coefficient=preemphasis_coefficient, + window_function=feature_window_function( + window_type=window_type, + window_size=window_size, + blackman_coeff=blackman_coeff + ).astype(dtype), + raw_energy=raw_energy + ) + return frames, log_enery + +# ---------- feature-window ---------- + + +# ---------- feature-functions ---------- + +def compute_spectrum(frames, n): + complex_spec = np.fft.rfft(frames, n) + return np.absolute(complex_spec) + + +def compute_power_spectrum(frames, n): + return np.square(compute_spectrum(frames, n)) + + +def apply_cmvn_sliding_internal(feat, center=False, window=600, min_window=100, norm_vars=False): + num_frames, feat_dim = feat.shape + std = 1 + if center: + if num_frames <= window: + mean = feat.mean(axis=0, keepdims=True).repeat(num_frames, axis=0) + if norm_vars: + std = feat.std(axis=0, keepdims=True).repeat(num_frames, axis=0) + else: + feat1 = feat[:window] + feat2 = sliding_window(feat.T, window, 1) + feat3 = feat[-window:] + mean1 = feat1.mean(axis=0, keepdims=True).repeat(window // 2, axis=0) + mean2 = feat2.mean(axis=2).T + mean3 = feat3.mean(axis=0, keepdims=True).repeat((window - 1) // 2, axis=0) + mean = np.concatenate([mean1, mean2, mean3]) + if norm_vars: + std1 = feat1.std(axis=0, keepdims=True).repeat(window // 2, axis=0) + std2 = feat2.std(axis=2).T + std3 = feat3.mean(axis=0, keepdims=True).repeat((window - 1) // 2, axis=0) + std = np.concatenate([std1, std2, std3]) + else: + if num_frames <= min_window: + mean = feat.mean(axis=0, keepdims=True).repeat(num_frames, axis=0) + if norm_vars: + std = feat.std(axis=0, keepdims=True).repeat(num_frames, axis=0) + else: + feat1 = feat[:min_window] + mean1 = feat1.mean(axis=0, keepdims=True).repeat(min_window, axis=0) + feat2_cumsum = np.cumsum(feat[:window], axis=0)[min_window:] + cumcnt = np.arange(min_window + 1, min(window, num_frames) + 1, dtype=feat.dtype)[:, np.newaxis] + mean2 = feat2_cumsum / cumcnt + mean = np.concatenate([mean1, mean2]) + if norm_vars: + std1 = feat1.std(axis=0, keepdims=True).repeat(min_window, axis=0) + feat2_power_cumsum = np.cumsum(np.square(feat[:window]), axis=0)[min_window:] + std2 = np.sqrt(feat2_power_cumsum / cumcnt - np.square(mean2)) + std = np.concatenate([std1, std2]) + if num_frames > window: + feat3 = sliding_window(feat.T, window, 1) + mean3 = feat3.mean(axis=2).T + mean = np.concatenate([mean, mean3[1:]]) + if norm_vars: + std3 = feat3.std(axis=2).T + std = np.concatenate([std, std3[1:]]) + feat = (feat - mean) / std + return feat + +# ---------- feature-functions ---------- + + +# ---------- mel-computations ---------- + +def inverse_mel_scale(mel_freq): + return 700.0 * (np.exp(mel_freq / 1127.0) - 1.0) + + +def mel_scale(freq): + return 1127.0 * np.log(1.0 + freq / 700.0) + + +def compute_mel_banks(num_bins, sample_frequency, low_freq, high_freq, n): + """ Compute Mel banks. + + :param num_bins: Number of triangular mel-frequency bins + :param sample_frequency: Waveform data sample frequency + :param low_freq: Low cutoff frequency for mel bins + :param high_freq: High cutoff frequency for mel bins (if <= 0, offset from Nyquist) + :param n: Window size + :return: Mel banks. + """ + assert num_bins >= 3, 'Must have at least 3 mel bins' + num_fft_bins = n // 2 + + nyquist = 0.5 * sample_frequency + if high_freq <= 0: + high_freq = nyquist + high_freq + assert 0 <= low_freq < high_freq <= nyquist + + fft_bin_width = sample_frequency / n + + mel_low_freq = mel_scale(low_freq) + mel_high_freq = mel_scale(high_freq) + mel_freq_delta = (mel_high_freq - mel_low_freq) / (num_bins + 1) + + mel_banks = np.zeros([num_bins, num_fft_bins + 1]) + for i in range(num_bins): + left_mel = mel_low_freq + mel_freq_delta * i + center_mel = left_mel + mel_freq_delta + right_mel = center_mel + mel_freq_delta + for j in range(num_fft_bins): + mel = mel_scale(fft_bin_width * j) + if left_mel < mel < right_mel: + if mel <= center_mel: + mel_banks[i, j] = (mel - left_mel) / (center_mel - left_mel) + else: + mel_banks[i, j] = (right_mel - mel) / (right_mel - center_mel) + return mel_banks + + +def compute_lifter_coeffs(q, M): + """ Compute liftering coefficients (scaling on cepstral coeffs) + the zeroth index is C0, which is not affected. + + :param q: Number of lifters + :param M: Number of coefficients + :return: Lifters. + """ + if M < 1: + return np.array([]) + if M == 1: + return np.ones(1, float) + n = np.arange(0, M) + return 1 + 0.5*np.sin(np.pi*n/q)*q + +# ---------- mel-computations ---------- + + +# ---------- compute-fbank-feats ---------- + +def compute_fbank_feats( + waveform, + blackman_coeff=0.42, + dither=1.0, + energy_floor=0.0, + frame_length=25, + frame_shift=10, + high_freq=0, + low_freq=20, + num_mel_bins=23, + preemphasis_coefficient=0.97, + raw_energy=True, + remove_dc_offset=True, + round_to_power_of_two=True, + sample_frequency=16000, + snip_edges=True, + use_energy=False, + use_log_fbank=True, + use_power=True, + window_type='povey', + dtype=np.float32): + """ Compute (log) Mel filter bank energies + + :param waveform: Input waveform. + :param blackman_coeff: Constant coefficient for generalized Blackman window. (float, default = 0.42) + :param dither: Dithering constant (0.0 means no dither). If you turn this off, you should set the --energy-floor option, e.g. to 1.0 or 0.1 (float, default = 1) + :param energy_floor: Floor on energy (absolute, not relative) in FBANK computation. Only makes a difference if --use-energy=true; only necessary if --dither=0.0. Suggested values: 0.1 or 1.0 (float, default = 0) + :param frame_length: Frame length in milliseconds (float, default = 25) + :param frame_shift: Frame shift in milliseconds (float, default = 10) + :param high_freq: High cutoff frequency for mel bins (if <= 0, offset from Nyquist) (float, default = 0) + :param low_freq: Low cutoff frequency for mel bins (float, default = 20) + :param num_mel_bins: Number of triangular mel-frequency bins (int, default = 23) + :param preemphasis_coefficient: Coefficient for use in signal preemphasis (float, default = 0.97) + :param raw_energy: If true, compute energy before preemphasis and windowing (bool, default = true) + :param remove_dc_offset: Subtract mean from waveform on each frame (bool, default = true) + :param round_to_power_of_two: If true, round window size to power of two by zero-padding input to FFT. (bool, default = true) + :param sample_frequency: Waveform data sample frequency (must match the waveform file, if specified there) (float, default = 16000) + :param snip_edges: If true, end effects will be handled by outputting only frames that completely fit in the file, and the number of frames depends on the frame-length. If false, the number of frames depends only on the frame-shift, and we reflect the data at the ends. (bool, default = true) + :param use_energy: Add an extra energy output. (bool, default = false) + :param use_log_fbank: If true, produce log-filterbank, else produce linear. (bool, default = true) + :param use_power: If true, use power, else use magnitude. (bool, default = true) + :param window_type: Type of window ("hamming"|"hanning"|"povey"|"rectangular"|"sine"|"blackmann") (string, default = "povey") + :param dtype: Type of array (np.float32|np.float64) (dtype or string, default=np.float32) + :return: (Log) Mel filter bank energies. + """ + window_size = int(frame_length * sample_frequency * 0.001) + window_shift = int(frame_shift * sample_frequency * 0.001) + frames, log_energy = extract_window( + waveform=waveform, + blackman_coeff=blackman_coeff, + dither=dither, + window_size=window_size, + window_shift=window_shift, + preemphasis_coefficient=preemphasis_coefficient, + raw_energy=raw_energy, + remove_dc_offset=remove_dc_offset, + snip_edges=snip_edges, + window_type=window_type, + dtype=dtype + ) + if round_to_power_of_two: + n = 1 + while n < window_size: + n *= 2 + else: + n = window_size + if use_power: + spectrum = compute_power_spectrum(frames, n) + else: + spectrum = compute_spectrum(frames, n) + mel_banks = compute_mel_banks( + num_bins=num_mel_bins, + sample_frequency=sample_frequency, + low_freq=low_freq, + high_freq=high_freq, + n=n + ).astype(dtype) + feat = np.dot(spectrum, mel_banks.T) + if use_log_fbank: + feat = np.log(feat.clip(min=np.finfo(dtype).eps)) + if use_energy: + if energy_floor > 0.0: + log_energy.clip(min=np.math.log(energy_floor)) + return feat, log_energy + return feat + +# ---------- compute-fbank-feats ---------- + + +# ---------- compute-mfcc-feats ---------- + +def compute_mfcc_feats( + waveform, + blackman_coeff=0.42, + cepstral_lifter=22, + dither=1.0, + energy_floor=0.0, + frame_length=25, + frame_shift=10, + high_freq=0, + low_freq=20, + num_ceps=13, + num_mel_bins=23, + preemphasis_coefficient=0.97, + raw_energy=True, + remove_dc_offset=True, + round_to_power_of_two=True, + sample_frequency=16000, + snip_edges=True, + use_energy=True, + window_type='povey', + dtype=np.float32): + """ Compute mel-frequency cepstral coefficients + + :param waveform: Input waveform. + :param blackman_coeff: Constant coefficient for generalized Blackman window. (float, default = 0.42) + :param cepstral_lifter: Constant that controls scaling of MFCCs (float, default = 22) + :param dither: Dithering constant (0.0 means no dither). If you turn this off, you should set the --energy-floor option, e.g. to 1.0 or 0.1 (float, default = 1) + :param energy_floor: Floor on energy (absolute, not relative) in MFCC computation. Only makes a difference if --use-energy=true; only necessary if --dither=0.0. Suggested values: 0.1 or 1.0 (float, default = 0) + :param frame_length: Frame length in milliseconds (float, default = 25) + :param frame_shift: Frame shift in milliseconds (float, default = 10) + :param high_freq: High cutoff frequency for mel bins (if <= 0, offset from Nyquist) (float, default = 0) + :param low_freq: Low cutoff frequency for mel bins (float, default = 20) + :param num_ceps: Number of cepstra in MFCC computation (including C0) (int, default = 13) + :param num_mel_bins: Number of triangular mel-frequency bins (int, default = 23) + :param preemphasis_coefficient: Coefficient for use in signal preemphasis (float, default = 0.97) + :param raw_energy: If true, compute energy before preemphasis and windowing (bool, default = true) + :param remove_dc_offset: Subtract mean from waveform on each frame (bool, default = true) + :param round_to_power_of_two: If true, round window size to power of two by zero-padding input to FFT. (bool, default = true) + :param sample_frequency: Waveform data sample frequency (must match the waveform file, if specified there) (float, default = 16000) + :param snip_edges: If true, end effects will be handled by outputting only frames that completely fit in the file, and the number of frames depends on the frame-length. If false, the number of frames depends only on the frame-shift, and we reflect the data at the ends. (bool, default = true) + :param use_energy: Use energy (not C0) in MFCC computation (bool, default = true) + :param window_type: Type of window ("hamming"|"hanning"|"povey"|"rectangular"|"sine"|"blackmann") (string, default = "povey") + :param dtype: Type of array (np.float32|np.float64) (dtype or string, default=np.float32) + :return: Mel-frequency cespstral coefficients. + """ + feat, log_energy = compute_fbank_feats( + waveform=waveform, + blackman_coeff=blackman_coeff, + dither=dither, + energy_floor=energy_floor, + frame_length=frame_length, + frame_shift=frame_shift, + high_freq=high_freq, + low_freq=low_freq, + num_mel_bins=num_mel_bins, + preemphasis_coefficient=preemphasis_coefficient, + raw_energy=raw_energy, + remove_dc_offset=remove_dc_offset, + round_to_power_of_two=round_to_power_of_two, + sample_frequency=sample_frequency, + snip_edges=snip_edges, + use_energy=use_energy, + use_log_fbank=True, + use_power=True, + window_type=window_type, + dtype=dtype + ) + feat = dct(feat, type=2, axis=1, norm='ortho')[:, :num_ceps] + lifter_coeffs = compute_lifter_coeffs(cepstral_lifter, num_ceps).astype(dtype) + feat = feat * lifter_coeffs + if use_energy: + feat[:, 0] = log_energy + return feat + +# ---------- compute-mfcc-feats ---------- + + +# ---------- apply-cmvn-sliding ---------- + +def apply_cmvn_sliding(feat, center=False, window=600, min_window=100, norm_vars=False): + """ Apply sliding-window cepstral mean (and optionally variance) normalization + + :param feat: Cepstrum. + :param center: If true, use a window centered on the current frame (to the extent possible, modulo end effects). If false, window is to the left. (bool, default = false) + :param window: Window in frames for running average CMN computation (int, default = 600) + :param min_window: Minimum CMN window used at start of decoding (adds latency only at start). Only applicable if center == false, ignored if center==true (int, default = 100) + :param norm_vars: If true, normalize variance to one. (bool, default = false) + :return: Normalized cepstrum. + """ + # double-precision + feat = apply_cmvn_sliding_internal( + feat=feat.astype(np.float64), + center=center, + window=window, + min_window=min_window, + norm_vars=norm_vars + ).astype(feat.dtype) + return feat + +# ---------- apply-cmvn-sliding ---------- diff --git a/funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/ivector.py b/funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/ivector.py new file mode 100644 index 000000000..5577be1f0 --- /dev/null +++ b/funasr/runtime/python/onnxruntime/rapid_paraformer/kaldifeat/ivector.py @@ -0,0 +1,43 @@ +import numpy as np + +from .feature import sliding_window + + +# ---------- compute-vad ---------- + +def compute_vad(log_energy, energy_mean_scale=0.5, energy_threshold=0.5, frames_context=0, proportion_threshold=0.6): + """ Apply voice activity detection + + :param log_energy: Log mel energy. + :param energy_mean_scale: If this is set to s, to get the actual threshold we let m be the mean log-energy of the file, and use s*m + vad-energy-threshold (float, default = 0.5) + :param energy_threshold: Constant term in energy threshold for VAD (also see energy_mean_scale) (float, default = 5) + :param frames_context: Number of frames of context on each side of central frame, in window for which energy is monitored (int, default = 0) + :param proportion_threshold: Parameter controlling the proportion of frames within the window that need to have more energy than the threshold (float, default = 0.6) + :return: A vector of boolean that are True if we judge the frame voiced and False otherwise. + """ + assert len(log_energy.shape) == 1 + assert energy_mean_scale >= 0 + assert frames_context >= 0 + assert 0 < proportion_threshold < 1 + dtype = log_energy.dtype + energy_threshold += energy_mean_scale * log_energy.mean() + if frames_context > 0: + num_frames = len(log_energy) + window_size = frames_context * 2 + 1 + log_energy_pad = np.concatenate([ + np.zeros(frames_context, dtype=dtype), + log_energy, + np.zeros(frames_context, dtype=dtype) + ]) + log_energy_window = sliding_window(log_energy_pad, window_size, 1) + num_count = np.count_nonzero(log_energy_window > energy_threshold, axis=1) + den_count = np.ones(num_frames, dtype=dtype) * window_size + max_den_count = np.arange(frames_context + 1, min(window_size, num_frames) + 1, dtype=dtype) + den_count[:-(frames_context + 2):-1] = max_den_count + den_count[:frames_context + 1] = np.min([den_count[:frames_context + 1], max_den_count], axis=0) + vad = num_count / den_count >= proportion_threshold + else: + vad = log_energy > energy_threshold + return vad + +# ---------- compute-vad ---------- diff --git a/funasr/runtime/python/onnxruntime/rapid_paraformer/rapid_paraformer.py b/funasr/runtime/python/onnxruntime/rapid_paraformer/rapid_paraformer.py new file mode 100644 index 000000000..262c2581c --- /dev/null +++ b/funasr/runtime/python/onnxruntime/rapid_paraformer/rapid_paraformer.py @@ -0,0 +1,139 @@ +# -*- encoding: utf-8 -*- +# @Author: SWHL +# @Contact: liekkaskono@163.com +import traceback +from pathlib import Path +from typing import List, Union, Tuple + +import librosa +import numpy as np + +from .utils import (CharTokenizer, Hypothesis, ONNXRuntimeError, + OrtInferSession, TokenIDConverter, WavFrontend, get_logger, + read_yaml) + +logging = get_logger() + + +class RapidParaformer(): + def __init__(self, config_path: Union[str, Path]) -> None: + if not Path(config_path).exists(): + raise FileNotFoundError(f'{config_path} does not exist.') + + config = read_yaml(config_path) + + self.converter = TokenIDConverter(**config['TokenIDConverter']) + self.tokenizer = CharTokenizer(**config['CharTokenizer']) + self.frontend = WavFrontend( + cmvn_file=config['WavFrontend']['cmvn_file'], + **config['WavFrontend']['frontend_conf'] + ) + self.ort_infer = OrtInferSession(config['Model']) + self.batch_size = config['Model']['batch_size'] + + def __call__(self, wav_content: Union[str, np.ndarray, List[str]]) -> List: + waveform_list = self.load_data(wav_content) + waveform_nums = len(waveform_list) + + asr_res = [] + for beg_idx in range(0, waveform_nums, self.batch_size): + end_idx = min(waveform_nums, beg_idx + self.batch_size) + + feats, feats_len = self.extract_feat(waveform_list[beg_idx:end_idx]) + + try: + am_scores, valid_token_lens = self.infer(feats, feats_len) + except ONNXRuntimeError: + logging.error(traceback.format_exc()) + preds = [] + else: + preds = self.decode(am_scores, valid_token_lens) + + asr_res.extend(preds) + return asr_res + + def load_data(self, + wav_content: Union[str, np.ndarray, List[str]]) -> List: + def load_wav(path: str) -> np.ndarray: + waveform, _ = librosa.load(path) + return waveform[None, ...] + + if isinstance(wav_content, np.ndarray): + return [wav_content] + + if isinstance(wav_content, str): + return [load_wav(wav_content)] + + if isinstance(wav_content, list): + return [load_wav(path) for path in wav_content] + + raise TypeError( + f'The type of {wav_content} is not in [str, np.ndarray, list]') + + def extract_feat(self, + waveform_list: List[np.ndarray] + ) -> Tuple[np.ndarray, np.ndarray]: + feats, feats_len = [], [] + for waveform in waveform_list: + speech, _ = self.frontend.fbank(waveform) + feat, feat_len = self.frontend.lfr_cmvn(speech) + feats.append(feat) + feats_len.append(feat_len) + + feats = self.pad_feats(feats, np.max(feats_len)) + feats_len = np.array(feats_len).astype(np.int32) + return feats, feats_len + + @staticmethod + def pad_feats(feats: List[np.ndarray], max_feat_len: int) -> np.ndarray: + def pad_feat(feat: np.ndarray, cur_len: int) -> np.ndarray: + pad_width = ((0, max_feat_len - cur_len), (0, 0)) + return np.pad(feat, pad_width, 'constant', constant_values=0) + + feat_res = [pad_feat(feat, feat.shape[0]) for feat in feats] + feats = np.array(feat_res).astype(np.float32) + return feats + + def infer(self, feats: np.ndarray, + feats_len: np.ndarray) -> Tuple[np.ndarray, np.ndarray]: + am_scores, token_nums = self.ort_infer([feats, feats_len]) + return am_scores, token_nums + + def decode(self, am_scores: np.ndarray, token_nums: int) -> List[str]: + return [self.decode_one(am_score, token_num) + for am_score, token_num in zip(am_scores, token_nums)] + + def decode_one(self, + am_score: np.ndarray, + valid_token_num: int) -> List[str]: + yseq = am_score.argmax(axis=-1) + score = am_score.max(axis=-1) + score = np.sum(score, axis=-1) + + # pad with mask tokens to ensure compatibility with sos/eos tokens + # asr_model.sos:1 asr_model.eos:2 + yseq = np.array([1] + yseq.tolist() + [2]) + hyp = Hypothesis(yseq=yseq, score=score) + + # remove sos/eos and get results + last_pos = -1 + token_int = hyp.yseq[1:last_pos].tolist() + + # remove blank symbol id, which is assumed to be 0 + token_int = list(filter(lambda x: x not in (0, 2), token_int)) + + # Change integer-ids to tokens + token = self.converter.ids2tokens(token_int) + text = self.tokenizer.tokens2text(token) + return text[:valid_token_num-1] + + +if __name__ == '__main__': + project_dir = Path(__file__).resolve().parent.parent + cfg_path = project_dir / 'resources' / 'config.yaml' + paraformer = RapidParaformer(cfg_path) + + wav_file = '0478_00017.wav' + for i in range(1000): + result = paraformer(wav_file) + print(result) diff --git a/funasr/runtime/python/onnxruntime/rapid_paraformer/utils.py b/funasr/runtime/python/onnxruntime/rapid_paraformer/utils.py new file mode 100644 index 000000000..adab79084 --- /dev/null +++ b/funasr/runtime/python/onnxruntime/rapid_paraformer/utils.py @@ -0,0 +1,371 @@ +# -*- encoding: utf-8 -*- +# @Author: SWHL +# @Contact: liekkaskono@163.com +import functools +import logging +import pickle +from pathlib import Path +from typing import Any, Dict, Iterable, List, NamedTuple, Set, Tuple, Union + +import numpy as np +import yaml +from onnxruntime import (GraphOptimizationLevel, InferenceSession, + SessionOptions, get_available_providers, get_device) +from typeguard import check_argument_types + +from .kaldifeat import compute_fbank_feats + +root_dir = Path(__file__).resolve().parent + +logger_initialized = {} + + +class TokenIDConverter(): + def __init__(self, token_path: Union[Path, str], + unk_symbol: str = "",): + check_argument_types() + + self.token_list = self.load_token(token_path) + self.unk_symbol = unk_symbol + + @staticmethod + def load_token(file_path: Union[Path, str]) -> List: + if not Path(file_path).exists(): + raise TokenIDConverterError(f'The {file_path} does not exist.') + + with open(str(file_path), 'rb') as f: + token_list = pickle.load(f) + + if len(token_list) != len(set(token_list)): + raise TokenIDConverterError('The Token exists duplicated symbol.') + return token_list + + def get_num_vocabulary_size(self) -> int: + return len(self.token_list) + + def ids2tokens(self, + integers: Union[np.ndarray, Iterable[int]]) -> List[str]: + if isinstance(integers, np.ndarray) and integers.ndim != 1: + raise TokenIDConverterError( + f"Must be 1 dim ndarray, but got {integers.ndim}") + return [self.token_list[i] for i in integers] + + def tokens2ids(self, tokens: Iterable[str]) -> List[int]: + token2id = {v: i for i, v in enumerate(self.token_list)} + if self.unk_symbol not in token2id: + raise TokenIDConverterError( + f"Unknown symbol '{self.unk_symbol}' doesn't exist in the token_list" + ) + unk_id = token2id[self.unk_symbol] + return [token2id.get(i, unk_id) for i in tokens] + + +class CharTokenizer(): + def __init__( + self, + symbol_value: Union[Path, str, Iterable[str]] = None, + space_symbol: str = "", + remove_non_linguistic_symbols: bool = False, + ): + check_argument_types() + + self.space_symbol = space_symbol + self.non_linguistic_symbols = self.load_symbols(symbol_value) + self.remove_non_linguistic_symbols = remove_non_linguistic_symbols + + @staticmethod + def load_symbols(value: Union[Path, str, Iterable[str]] = None) -> Set: + if value is None: + return set() + + if isinstance(value, Iterable[str]): + return set(value) + + file_path = Path(value) + if not file_path.exists(): + logging.warning("%s doesn't exist.", file_path) + return set() + + with file_path.open("r", encoding="utf-8") as f: + return set(line.rstrip() for line in f) + + def text2tokens(self, line: Union[str, list]) -> List[str]: + tokens = [] + while len(line) != 0: + for w in self.non_linguistic_symbols: + if line.startswith(w): + if not self.remove_non_linguistic_symbols: + tokens.append(line[: len(w)]) + line = line[len(w):] + break + else: + t = line[0] + if t == " ": + t = "" + tokens.append(t) + line = line[1:] + return tokens + + def tokens2text(self, tokens: Iterable[str]) -> str: + tokens = [t if t != self.space_symbol else " " for t in tokens] + return "".join(tokens) + + def __repr__(self): + return ( + f"{self.__class__.__name__}(" + f'space_symbol="{self.space_symbol}"' + f'non_linguistic_symbols="{self.non_linguistic_symbols}"' + f")" + ) + + +class WavFrontend(): + """Conventional frontend structure for ASR. + """ + + def __init__( + self, + cmvn_file: str = None, + fs: int = 16000, + window: str = 'hamming', + n_mels: int = 80, + frame_length: int = 25, + frame_shift: int = 10, + filter_length_min: int = -1, + filter_length_max: float = -1, + lfr_m: int = 1, + lfr_n: int = 1, + dither: float = 1.0 + ) -> None: + check_argument_types() + + self.fs = fs + self.window = window + self.n_mels = n_mels + self.frame_length = frame_length + self.frame_shift = frame_shift + self.filter_length_min = filter_length_min + self.filter_length_max = filter_length_max + self.lfr_m = lfr_m + self.lfr_n = lfr_n + self.cmvn_file = cmvn_file + self.dither = dither + + if self.cmvn_file: + self.cmvn = self.load_cmvn() + + def fbank(self, + input_content: np.ndarray) -> Tuple[np.ndarray, np.ndarray]: + waveform_len = input_content.shape[1] + waveform = input_content[0][:waveform_len] + waveform = waveform * (1 << 15) + mat = compute_fbank_feats(waveform, + num_mel_bins=self.n_mels, + frame_length=self.frame_length, + frame_shift=self.frame_shift, + dither=self.dither, + energy_floor=0.0, + sample_frequency=self.fs) + feat = mat.astype(np.float32) + feat_len = np.array(mat.shape[0]).astype(np.int32) + return feat, feat_len + + def lfr_cmvn(self, feat: np.ndarray) -> Tuple[np.ndarray, np.ndarray]: + if self.lfr_m != 1 or self.lfr_n != 1: + feat = self.apply_lfr(feat, self.lfr_m, self.lfr_n) + + if self.cmvn_file: + feat = self.apply_cmvn(feat) + + feat_len = np.array(feat.shape[0]).astype(np.int32) + return feat, feat_len + + @staticmethod + def apply_lfr(inputs: np.ndarray, lfr_m: int, lfr_n: int) -> np.ndarray: + LFR_inputs = [] + + T = inputs.shape[0] + T_lfr = int(np.ceil(T / lfr_n)) + left_padding = np.tile(inputs[0], ((lfr_m - 1) // 2, 1)) + inputs = np.vstack((left_padding, inputs)) + T = T + (lfr_m - 1) // 2 + for i in range(T_lfr): + if lfr_m <= T - i * lfr_n: + LFR_inputs.append( + (inputs[i * lfr_n:i * lfr_n + lfr_m]).reshape(1, -1)) + else: + # process last LFR frame + num_padding = lfr_m - (T - i * lfr_n) + frame = inputs[i * lfr_n:].reshape(-1) + for _ in range(num_padding): + frame = np.hstack((frame, inputs[-1])) + + LFR_inputs.append(frame) + LFR_outputs = np.vstack(LFR_inputs).astype(np.float32) + return LFR_outputs + + def apply_cmvn(self, inputs: np.ndarray) -> np.ndarray: + """ + Apply CMVN with mvn data + """ + frame, dim = inputs.shape + means = np.tile(self.cmvn[0:1, :dim], (frame, 1)) + vars = np.tile(self.cmvn[1:2, :dim], (frame, 1)) + inputs = (inputs + means) * vars + return inputs + + def load_cmvn(self,) -> np.ndarray: + with open(self.cmvn_file, 'r', encoding='utf-8') as f: + lines = f.readlines() + + means_list = [] + vars_list = [] + for i in range(len(lines)): + line_item = lines[i].split() + if line_item[0] == '': + line_item = lines[i + 1].split() + if line_item[0] == '': + add_shift_line = line_item[3:(len(line_item) - 1)] + means_list = list(add_shift_line) + continue + elif line_item[0] == '': + line_item = lines[i + 1].split() + if line_item[0] == '': + rescale_line = line_item[3:(len(line_item) - 1)] + vars_list = list(rescale_line) + continue + + means = np.array(means_list).astype(np.float64) + vars = np.array(vars_list).astype(np.float64) + cmvn = np.array([means, vars]) + return cmvn + + +class Hypothesis(NamedTuple): + """Hypothesis data type.""" + + yseq: np.ndarray + score: Union[float, np.ndarray] = 0 + scores: Dict[str, Union[float, np.ndarray]] = dict() + states: Dict[str, Any] = dict() + + def asdict(self) -> dict: + """Convert data to JSON-friendly dict.""" + return self._replace( + yseq=self.yseq.tolist(), + score=float(self.score), + scores={k: float(v) for k, v in self.scores.items()}, + )._asdict() + + +class TokenIDConverterError(Exception): + pass + + +class ONNXRuntimeError(Exception): + pass + + +class OrtInferSession(): + def __init__(self, config): + sess_opt = SessionOptions() + sess_opt.log_severity_level = 4 + sess_opt.enable_cpu_mem_arena = False + sess_opt.graph_optimization_level = GraphOptimizationLevel.ORT_ENABLE_ALL + + cuda_ep = 'CUDAExecutionProvider' + cpu_ep = 'CPUExecutionProvider' + cpu_provider_options = { + "arena_extend_strategy": "kSameAsRequested", + } + + EP_list = [] + if config['use_cuda'] and get_device() == 'GPU' \ + and cuda_ep in get_available_providers(): + EP_list = [(cuda_ep, config[cuda_ep])] + EP_list.append((cpu_ep, cpu_provider_options)) + + config['model_path'] = config['model_path'] + self._verify_model(config['model_path']) + self.session = InferenceSession(config['model_path'], + sess_options=sess_opt, + providers=EP_list) + + if config['use_cuda'] and cuda_ep not in self.session.get_providers(): + warnings.warn(f'{cuda_ep} is not avaiable for current env, the inference part is automatically shifted to be executed under {cpu_ep}.\n' + 'Please ensure the installed onnxruntime-gpu version matches your cuda and cudnn version, ' + 'you can check their relations from the offical web site: ' + 'https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html', + RuntimeWarning) + + def __call__(self, + input_content: List[Union[np.ndarray, np.ndarray]]) -> np.ndarray: + input_dict = dict(zip(self.get_input_names(), input_content)) + try: + return self.session.run(None, input_dict) + except Exception as e: + raise ONNXRuntimeError('ONNXRuntime inferece failed.') from e + + def get_input_names(self, ): + return [v.name for v in self.session.get_inputs()] + + def get_output_names(self,): + return [v.name for v in self.session.get_outputs()] + + def get_character_list(self, key: str = 'character'): + return self.meta_dict[key].splitlines() + + def have_key(self, key: str = 'character') -> bool: + self.meta_dict = self.session.get_modelmeta().custom_metadata_map + if key in self.meta_dict.keys(): + return True + return False + + @staticmethod + def _verify_model(model_path): + model_path = Path(model_path) + if not model_path.exists(): + raise FileNotFoundError(f'{model_path} does not exists.') + if not model_path.is_file(): + raise FileExistsError(f'{model_path} is not a file.') + + +def read_yaml(yaml_path: Union[str, Path]) -> Dict: + if not Path(yaml_path).exists(): + raise FileExistsError(f'The {yaml_path} does not exist.') + + with open(str(yaml_path), 'rb') as f: + data = yaml.load(f, Loader=yaml.Loader) + return data + + +@functools.lru_cache() +def get_logger(name='rapdi_paraformer'): + """Initialize and get a logger by name. + If the logger has not been initialized, this method will initialize the + logger by adding one or two handlers, otherwise the initialized logger will + be directly returned. During initialization, a StreamHandler will always be + added. + Args: + name (str): Logger name. + Returns: + logging.Logger: The expected logger. + """ + logger = logging.getLogger(name) + if name in logger_initialized: + return logger + + for logger_name in logger_initialized: + if name.startswith(logger_name): + return logger + + formatter = logging.Formatter( + '[%(asctime)s] %(name)s %(levelname)s: %(message)s', + datefmt="%Y/%m/%d %H:%M:%S") + + sh = logging.StreamHandler() + sh.setFormatter(formatter) + logger.addHandler(sh) + logger_initialized[name] = True + logger.propagate = False + return logger diff --git a/funasr/runtime/python/onnxruntime/requirements.txt b/funasr/runtime/python/onnxruntime/requirements.txt new file mode 100644 index 000000000..ed40eb497 --- /dev/null +++ b/funasr/runtime/python/onnxruntime/requirements.txt @@ -0,0 +1,5 @@ +librosa +numpy +onnxruntime +scipy +typeguard \ No newline at end of file diff --git a/funasr/runtime/python/onnxruntime/resources/config.yaml b/funasr/runtime/python/onnxruntime/resources/config.yaml new file mode 100644 index 000000000..fd243c304 --- /dev/null +++ b/funasr/runtime/python/onnxruntime/resources/config.yaml @@ -0,0 +1,30 @@ +TokenIDConverter: + token_path: resources/models/token_list.pkl + unk_symbol: + +CharTokenizer: + symbol_value: + space_symbol: + remove_non_linguistic_symbols: false + +WavFrontend: + cmvn_file: resources/models/am.mvn + frontend_conf: + fs: 16000 + window: hamming + n_mels: 80 + frame_length: 25 + frame_shift: 10 + lfr_m: 7 + lfr_n: 6 + filter_length_max: -.inf + +Model: + model_path: resources/models/model.onnx + use_cuda: false + CUDAExecutionProvider: + device_id: 0 + arena_extend_strategy: kNextPowerOfTwo + cudnn_conv_algo_search: EXHAUSTIVE + do_copy_in_default_stream: true + batch_size: 3 \ No newline at end of file diff --git a/funasr/runtime/python/onnxruntime/resources/models/am.mvn b/funasr/runtime/python/onnxruntime/resources/models/am.mvn new file mode 100644 index 000000000..681910cd1 --- /dev/null +++ b/funasr/runtime/python/onnxruntime/resources/models/am.mvn @@ -0,0 +1,8 @@ + + 560 560 +[ 0 ] + 560 560 + 0 [ -8.311879 -8.600912 -9.615928 -10.43595 -11.21292 -11.88333 -12.36243 -12.63706 -12.8818 -12.83066 -12.89103 -12.95666 -13.19763 -13.40598 -13.49113 -13.5546 -13.55639 -13.51915 -13.68284 -13.53289 -13.42107 -13.65519 -13.50713 -13.75251 -13.76715 -13.87408 -13.73109 -13.70412 -13.56073 -13.53488 -13.54895 -13.56228 -13.59408 -13.62047 -13.64198 -13.66109 -13.62669 -13.58297 -13.57387 -13.4739 -13.53063 -13.48348 -13.61047 -13.64716 -13.71546 -13.79184 -13.90614 -14.03098 -14.18205 -14.35881 -14.48419 -14.60172 -14.70591 -14.83362 -14.92122 -15.00622 -15.05122 -15.03119 -14.99028 -14.92302 -14.86927 -14.82691 -14.7972 -14.76909 -14.71356 -14.61277 -14.51696 -14.42252 -14.36405 -14.30451 -14.23161 -14.19851 -14.16633 -14.15649 -14.10504 -13.99518 -13.79562 -13.3996 -12.7767 -11.71208 -8.311879 -8.600912 -9.615928 -10.43595 -11.21292 -11.88333 -12.36243 -12.63706 -12.8818 -12.83066 -12.89103 -12.95666 -13.19763 -13.40598 -13.49113 -13.5546 -13.55639 -13.51915 -13.68284 -13.53289 -13.42107 -13.65519 -13.50713 -13.75251 -13.76715 -13.87408 -13.73109 -13.70412 -13.56073 -13.53488 -13.54895 -13.56228 -13.59408 -13.62047 -13.64198 -13.66109 -13.62669 -13.58297 -13.57387 -13.4739 -13.53063 -13.48348 -13.61047 -13.64716 -13.71546 -13.79184 -13.90614 -14.03098 -14.18205 -14.35881 -14.48419 -14.60172 -14.70591 -14.83362 -14.92122 -15.00622 -15.05122 -15.03119 -14.99028 -14.92302 -14.86927 -14.82691 -14.7972 -14.76909 -14.71356 -14.61277 -14.51696 -14.42252 -14.36405 -14.30451 -14.23161 -14.19851 -14.16633 -14.15649 -14.10504 -13.99518 -13.79562 -13.3996 -12.7767 -11.71208 -8.311879 -8.600912 -9.615928 -10.43595 -11.21292 -11.88333 -12.36243 -12.63706 -12.8818 -12.83066 -12.89103 -12.95666 -13.19763 -13.40598 -13.49113 -13.5546 -13.55639 -13.51915 -13.68284 -13.53289 -13.42107 -13.65519 -13.50713 -13.75251 -13.76715 -13.87408 -13.73109 -13.70412 -13.56073 -13.53488 -13.54895 -13.56228 -13.59408 -13.62047 -13.64198 -13.66109 -13.62669 -13.58297 -13.57387 -13.4739 -13.53063 -13.48348 -13.61047 -13.64716 -13.71546 -13.79184 -13.90614 -14.03098 -14.18205 -14.35881 -14.48419 -14.60172 -14.70591 -14.83362 -14.92122 -15.00622 -15.05122 -15.03119 -14.99028 -14.92302 -14.86927 -14.82691 -14.7972 -14.76909 -14.71356 -14.61277 -14.51696 -14.42252 -14.36405 -14.30451 -14.23161 -14.19851 -14.16633 -14.15649 -14.10504 -13.99518 -13.79562 -13.3996 -12.7767 -11.71208 -8.311879 -8.600912 -9.615928 -10.43595 -11.21292 -11.88333 -12.36243 -12.63706 -12.8818 -12.83066 -12.89103 -12.95666 -13.19763 -13.40598 -13.49113 -13.5546 -13.55639 -13.51915 -13.68284 -13.53289 -13.42107 -13.65519 -13.50713 -13.75251 -13.76715 -13.87408 -13.73109 -13.70412 -13.56073 -13.53488 -13.54895 -13.56228 -13.59408 -13.62047 -13.64198 -13.66109 -13.62669 -13.58297 -13.57387 -13.4739 -13.53063 -13.48348 -13.61047 -13.64716 -13.71546 -13.79184 -13.90614 -14.03098 -14.18205 -14.35881 -14.48419 -14.60172 -14.70591 -14.83362 -14.92122 -15.00622 -15.05122 -15.03119 -14.99028 -14.92302 -14.86927 -14.82691 -14.7972 -14.76909 -14.71356 -14.61277 -14.51696 -14.42252 -14.36405 -14.30451 -14.23161 -14.19851 -14.16633 -14.15649 -14.10504 -13.99518 -13.79562 -13.3996 -12.7767 -11.71208 -8.311879 -8.600912 -9.615928 -10.43595 -11.21292 -11.88333 -12.36243 -12.63706 -12.8818 -12.83066 -12.89103 -12.95666 -13.19763 -13.40598 -13.49113 -13.5546 -13.55639 -13.51915 -13.68284 -13.53289 -13.42107 -13.65519 -13.50713 -13.75251 -13.76715 -13.87408 -13.73109 -13.70412 -13.56073 -13.53488 -13.54895 -13.56228 -13.59408 -13.62047 -13.64198 -13.66109 -13.62669 -13.58297 -13.57387 -13.4739 -13.53063 -13.48348 -13.61047 -13.64716 -13.71546 -13.79184 -13.90614 -14.03098 -14.18205 -14.35881 -14.48419 -14.60172 -14.70591 -14.83362 -14.92122 -15.00622 -15.05122 -15.03119 -14.99028 -14.92302 -14.86927 -14.82691 -14.7972 -14.76909 -14.71356 -14.61277 -14.51696 -14.42252 -14.36405 -14.30451 -14.23161 -14.19851 -14.16633 -14.15649 -14.10504 -13.99518 -13.79562 -13.3996 -12.7767 -11.71208 -8.311879 -8.600912 -9.615928 -10.43595 -11.21292 -11.88333 -12.36243 -12.63706 -12.8818 -12.83066 -12.89103 -12.95666 -13.19763 -13.40598 -13.49113 -13.5546 -13.55639 -13.51915 -13.68284 -13.53289 -13.42107 -13.65519 -13.50713 -13.75251 -13.76715 -13.87408 -13.73109 -13.70412 -13.56073 -13.53488 -13.54895 -13.56228 -13.59408 -13.62047 -13.64198 -13.66109 -13.62669 -13.58297 -13.57387 -13.4739 -13.53063 -13.48348 -13.61047 -13.64716 -13.71546 -13.79184 -13.90614 -14.03098 -14.18205 -14.35881 -14.48419 -14.60172 -14.70591 -14.83362 -14.92122 -15.00622 -15.05122 -15.03119 -14.99028 -14.92302 -14.86927 -14.82691 -14.7972 -14.76909 -14.71356 -14.61277 -14.51696 -14.42252 -14.36405 -14.30451 -14.23161 -14.19851 -14.16633 -14.15649 -14.10504 -13.99518 -13.79562 -13.3996 -12.7767 -11.71208 -8.311879 -8.600912 -9.615928 -10.43595 -11.21292 -11.88333 -12.36243 -12.63706 -12.8818 -12.83066 -12.89103 -12.95666 -13.19763 -13.40598 -13.49113 -13.5546 -13.55639 -13.51915 -13.68284 -13.53289 -13.42107 -13.65519 -13.50713 -13.75251 -13.76715 -13.87408 -13.73109 -13.70412 -13.56073 -13.53488 -13.54895 -13.56228 -13.59408 -13.62047 -13.64198 -13.66109 -13.62669 -13.58297 -13.57387 -13.4739 -13.53063 -13.48348 -13.61047 -13.64716 -13.71546 -13.79184 -13.90614 -14.03098 -14.18205 -14.35881 -14.48419 -14.60172 -14.70591 -14.83362 -14.92122 -15.00622 -15.05122 -15.03119 -14.99028 -14.92302 -14.86927 -14.82691 -14.7972 -14.76909 -14.71356 -14.61277 -14.51696 -14.42252 -14.36405 -14.30451 -14.23161 -14.19851 -14.16633 -14.15649 -14.10504 -13.99518 -13.79562 -13.3996 -12.7767 -11.71208 ] + 560 560 + 0 [ 0.155775 0.154484 0.1527379 0.1518718 0.1506028 0.1489256 0.147067 0.1447061 0.1436307 0.1443568 0.1451849 0.1455157 0.1452821 0.1445717 0.1439195 0.1435867 0.1436018 0.1438781 0.1442086 0.1448844 0.1454756 0.145663 0.146268 0.1467386 0.1472724 0.147664 0.1480913 0.1483739 0.1488841 0.1493636 0.1497088 0.1500379 0.1502916 0.1505389 0.1506787 0.1507102 0.1505992 0.1505445 0.1505938 0.1508133 0.1509569 0.1512396 0.1514625 0.1516195 0.1516156 0.1515561 0.1514966 0.1513976 0.1512612 0.151076 0.1510596 0.1510431 0.151077 0.1511168 0.1511917 0.151023 0.1508045 0.1505885 0.1503493 0.1502373 0.1501726 0.1500762 0.1500065 0.1499782 0.150057 0.1502658 0.150469 0.1505335 0.1505505 0.1505328 0.1504275 0.1502438 0.1499674 0.1497118 0.1494661 0.1493102 0.1493681 0.1495501 0.1499738 0.1509654 0.155775 0.154484 0.1527379 0.1518718 0.1506028 0.1489256 0.147067 0.1447061 0.1436307 0.1443568 0.1451849 0.1455157 0.1452821 0.1445717 0.1439195 0.1435867 0.1436018 0.1438781 0.1442086 0.1448844 0.1454756 0.145663 0.146268 0.1467386 0.1472724 0.147664 0.1480913 0.1483739 0.1488841 0.1493636 0.1497088 0.1500379 0.1502916 0.1505389 0.1506787 0.1507102 0.1505992 0.1505445 0.1505938 0.1508133 0.1509569 0.1512396 0.1514625 0.1516195 0.1516156 0.1515561 0.1514966 0.1513976 0.1512612 0.151076 0.1510596 0.1510431 0.151077 0.1511168 0.1511917 0.151023 0.1508045 0.1505885 0.1503493 0.1502373 0.1501726 0.1500762 0.1500065 0.1499782 0.150057 0.1502658 0.150469 0.1505335 0.1505505 0.1505328 0.1504275 0.1502438 0.1499674 0.1497118 0.1494661 0.1493102 0.1493681 0.1495501 0.1499738 0.1509654 0.155775 0.154484 0.1527379 0.1518718 0.1506028 0.1489256 0.147067 0.1447061 0.1436307 0.1443568 0.1451849 0.1455157 0.1452821 0.1445717 0.1439195 0.1435867 0.1436018 0.1438781 0.1442086 0.1448844 0.1454756 0.145663 0.146268 0.1467386 0.1472724 0.147664 0.1480913 0.1483739 0.1488841 0.1493636 0.1497088 0.1500379 0.1502916 0.1505389 0.1506787 0.1507102 0.1505992 0.1505445 0.1505938 0.1508133 0.1509569 0.1512396 0.1514625 0.1516195 0.1516156 0.1515561 0.1514966 0.1513976 0.1512612 0.151076 0.1510596 0.1510431 0.151077 0.1511168 0.1511917 0.151023 0.1508045 0.1505885 0.1503493 0.1502373 0.1501726 0.1500762 0.1500065 0.1499782 0.150057 0.1502658 0.150469 0.1505335 0.1505505 0.1505328 0.1504275 0.1502438 0.1499674 0.1497118 0.1494661 0.1493102 0.1493681 0.1495501 0.1499738 0.1509654 0.155775 0.154484 0.1527379 0.1518718 0.1506028 0.1489256 0.147067 0.1447061 0.1436307 0.1443568 0.1451849 0.1455157 0.1452821 0.1445717 0.1439195 0.1435867 0.1436018 0.1438781 0.1442086 0.1448844 0.1454756 0.145663 0.146268 0.1467386 0.1472724 0.147664 0.1480913 0.1483739 0.1488841 0.1493636 0.1497088 0.1500379 0.1502916 0.1505389 0.1506787 0.1507102 0.1505992 0.1505445 0.1505938 0.1508133 0.1509569 0.1512396 0.1514625 0.1516195 0.1516156 0.1515561 0.1514966 0.1513976 0.1512612 0.151076 0.1510596 0.1510431 0.151077 0.1511168 0.1511917 0.151023 0.1508045 0.1505885 0.1503493 0.1502373 0.1501726 0.1500762 0.1500065 0.1499782 0.150057 0.1502658 0.150469 0.1505335 0.1505505 0.1505328 0.1504275 0.1502438 0.1499674 0.1497118 0.1494661 0.1493102 0.1493681 0.1495501 0.1499738 0.1509654 0.155775 0.154484 0.1527379 0.1518718 0.1506028 0.1489256 0.147067 0.1447061 0.1436307 0.1443568 0.1451849 0.1455157 0.1452821 0.1445717 0.1439195 0.1435867 0.1436018 0.1438781 0.1442086 0.1448844 0.1454756 0.145663 0.146268 0.1467386 0.1472724 0.147664 0.1480913 0.1483739 0.1488841 0.1493636 0.1497088 0.1500379 0.1502916 0.1505389 0.1506787 0.1507102 0.1505992 0.1505445 0.1505938 0.1508133 0.1509569 0.1512396 0.1514625 0.1516195 0.1516156 0.1515561 0.1514966 0.1513976 0.1512612 0.151076 0.1510596 0.1510431 0.151077 0.1511168 0.1511917 0.151023 0.1508045 0.1505885 0.1503493 0.1502373 0.1501726 0.1500762 0.1500065 0.1499782 0.150057 0.1502658 0.150469 0.1505335 0.1505505 0.1505328 0.1504275 0.1502438 0.1499674 0.1497118 0.1494661 0.1493102 0.1493681 0.1495501 0.1499738 0.1509654 0.155775 0.154484 0.1527379 0.1518718 0.1506028 0.1489256 0.147067 0.1447061 0.1436307 0.1443568 0.1451849 0.1455157 0.1452821 0.1445717 0.1439195 0.1435867 0.1436018 0.1438781 0.1442086 0.1448844 0.1454756 0.145663 0.146268 0.1467386 0.1472724 0.147664 0.1480913 0.1483739 0.1488841 0.1493636 0.1497088 0.1500379 0.1502916 0.1505389 0.1506787 0.1507102 0.1505992 0.1505445 0.1505938 0.1508133 0.1509569 0.1512396 0.1514625 0.1516195 0.1516156 0.1515561 0.1514966 0.1513976 0.1512612 0.151076 0.1510596 0.1510431 0.151077 0.1511168 0.1511917 0.151023 0.1508045 0.1505885 0.1503493 0.1502373 0.1501726 0.1500762 0.1500065 0.1499782 0.150057 0.1502658 0.150469 0.1505335 0.1505505 0.1505328 0.1504275 0.1502438 0.1499674 0.1497118 0.1494661 0.1493102 0.1493681 0.1495501 0.1499738 0.1509654 0.155775 0.154484 0.1527379 0.1518718 0.1506028 0.1489256 0.147067 0.1447061 0.1436307 0.1443568 0.1451849 0.1455157 0.1452821 0.1445717 0.1439195 0.1435867 0.1436018 0.1438781 0.1442086 0.1448844 0.1454756 0.145663 0.146268 0.1467386 0.1472724 0.147664 0.1480913 0.1483739 0.1488841 0.1493636 0.1497088 0.1500379 0.1502916 0.1505389 0.1506787 0.1507102 0.1505992 0.1505445 0.1505938 0.1508133 0.1509569 0.1512396 0.1514625 0.1516195 0.1516156 0.1515561 0.1514966 0.1513976 0.1512612 0.151076 0.1510596 0.1510431 0.151077 0.1511168 0.1511917 0.151023 0.1508045 0.1505885 0.1503493 0.1502373 0.1501726 0.1500762 0.1500065 0.1499782 0.150057 0.1502658 0.150469 0.1505335 0.1505505 0.1505328 0.1504275 0.1502438 0.1499674 0.1497118 0.1494661 0.1493102 0.1493681 0.1495501 0.1499738 0.1509654 ] + diff --git a/funasr/runtime/python/onnxruntime/resources/models/token_list.pkl b/funasr/runtime/python/onnxruntime/resources/models/token_list.pkl new file mode 100644 index 0000000000000000000000000000000000000000..f1a2ce77858f12695ff1dab468c3d1d5109817cb GIT binary patch literal 51686 zcmYJc$97%IvLz-0dxPhd{zJDq8Y-hIsm6FiqkcpUkOWCwf&d=?9=`Kh@ZNjxz4vaz zwDk?@t}xlFPf*O6n^$Zk7uH@YS7yYF$jr!y$d1zN|NH;^PyF|P&;0w5_y7E%slNH! z|1~pG`_FArmj3yF!++)a=1>3qzsF5Gos2Rq9^iWADN5N~h*FOZqx8*tTwm-9JB>0;;}Lzvwc`uX ztG+1h=o)wW(NEt;TV&xaO1%e- zYo#kn8Abit!@hQ`MlZg!vKR7*j@j-q%CqQRt?hJ1=_9x=BlA(}>t>Y28ymfV`;EvQ(! z8a=UNeAG7Zt8K8tl|?*L8ICeNvN2ugNm>B3el)a3DQ|#F@dyyg6WEx#P=i0;;rfu> z*&zte+3dF?Jb42&*y<~u=>Rq58u0czu-b-k$JeKDeS+?%4zNpa&_QF(Ev@X)D?59& z6ZicQa7dLNJb4>^sr%US{YM7n5IXCW6kP#3Wu8)20aa56x@i?OAPu4;=?NW_Iegl; z*x-SsxGr9!Yo*4ItTY_~S~iWo={j1Vn`Fin+TlB>G_3=QvNs*2&4gr>+OwGW3&n0xJS{8!XrRJY}Sx2JrH6Q+p z``$?%vxU#g9ZKBN%oXp6rR(Q8DtRX`qFLLG8JfmvW3z-(w2FkKcp zalJ+4@AJuB@IuDHFWV!#NBHg&8B!szpV6Z98Vo75SLw2MjEc@frt7{pR*B^k;IwbY z%~~6SZuJYasd&1MQU;?hp~?TsTsa3gWd=M;8>md@ppE<%Elfw?jh}%+dLFC}ap3c> zXSlONc)zio>wt;Bg0FfB4#_wK$6q0IX^Ft;j8>F>RN?zJ8$T2;WYh*&=MR8`E#bZo zqavFpv35bAv_PVC0Z`WUh;j>1@(~c%R?+k{kMeWFFWk9A2jmn@)M0?)a|>i!mm!}v z7Z8ZX0g`@;uggB5S3ICLKj&K=P_23uO|Sda*#4FDvQ_ZFFNnKEJSOweq?*Q0AnKQT zJ|t(<8K>w&y27FUF9GEUn^%IdGQJo+9^hs`es~iM*T1ilGq|@vSV<^W}p*n z3V8Doa8styOCM)9UfIwqBK8=J(qTNQE5yb?NJFhnkZBdvqpx6;?&2*U3!0TbHgFg! zeQltYlL7Eb9?%$>M}^Xfp6f6gnT}Dd9g@Jq;JDvH-_yfzjU zl|Qsr>2L>`w*#V{nn;Jq-U-qHjEz|enY>9gdlC#Ne{9ON`x0t&0?y9>xO7Af`i6d| zW!(7>853txwHib^4Ubbcio$z%4E_H;K4T&Kzwu1FF?QLDjmd^3a8<;LVq(Ub;V;vAhouw>n1Vw6PmkOl=kg#q@ja9&YdnWrqoD!5(-0~T; zw!Nh#N?Bpsjvxlw3q;5)LHK~a=x#`M=|f`%)zS`b-J&gx&8W!k(5059Kz@)3cvTG3 zrKYN0UBIX1Ai!StqqXG&Ytbd3K3%bQJydyvV1|w1x_BfBp8`Oo57ntxA26loXmh#- z$I=?wX}2pVuZgd5ye1Vi%`bpH=>t(x=LGI-;scEzzecI+Ac|ZQvky0MT|5!vqpW8Y zr4FLqsfViJ9C(onB4v_j8pO5F1u;sWDDCfl;gLx+CoPil+W@4D1VGBazBV?I=x^i? zRBvtS#>*eut7L_`(_8c@U4g~k8`2UKYiP?9yTAi!ixRREkCofAAjLck^eJofCM>U1;^Tq0q%TG4=JZdWc+0gv60l%zJDrZR<&JE;%bUVZwvQ8XcQog2fEwA7}TEZ7>1lW^N z&{i>4m)hzv$j7`2pZM@IgGt6u@GurVK}5jYo~hiP$m(rWYh9p_aK)#~(V1H5?TAuu zAY!^oi0uVY>b|zLH#D`yO)@%Z#FM3}V(=x1tDU2E znF9GGB9l3(w$r9Oqfp*Icfw~@@FaN4Te|26#80J7Ne9IK-D%uT3LVsNPSQV!ZA`iah;|)(u@ZH${(vEws5tq%M{}^ zbl!(IqqI2)e0o4@b(I1z0a~XHls-bW%b~?L*`ls^rlvV4S@*rZ^;<(bo;m|&(?Z}h zW-XXa`7qmo*-xpe8|0C?qZ{b+31agBB(QGiRNp0fZcyqIn(u3+p-pI2ZLSQalu4!n(lZQm-BgkSL zZ+s4IlU?eC1Ar{8;?*<@A{5U+hGHgB`t`M*G?;*h*%BVLZZt?|q1tpK%F=IbvkEvA z&ulNJakfix-GB*t$!48B;Cc*%*RAMBsXD+r$K_N3>GTCkGI7>vgP^s+X7RqU`5!}A zA4s4BaW}@)XnIZK=eU#&FkL4AJADS2{T5KNdG>T1Urii2q!*ymTXe|h&@$ZuF#Up1 z9|AWWBXNue@sW3}aTOC#{C!3F-b3x$O^)sn<`-zFZpHg`?Y|nD+y9ErHCOfU2{S5P|_0RMu+g()IA-g z4H9e|EK(Oq+z)1?F6a~=0qOk}IJSMTI8~y9rLXm^P)|xrg;nAkwSVHv)&+LSgSyH z8@Py4Cby$ZU64Mx;M*{nmD;#IE5#x&T3-HMpQl%P+Z)g68mMmvfVvMtJ(-0}+#Z$eKLPtovldv&EzJfi`=h@RTr8JhVuFZH~aqC5k>`4}3d11C|Y zVF*#;oZ6-Uth9g*`3;_NS~Js8yj%YCd$n>)ndyd*`!*rAh`Q4>g`tP_Ek(1#gvXk$ zW^g?STGuzlr4CU(*Fhy61M+1KAo2BRa0nx-AA~G;H>ywFz<^J`MQH~pr3=)SH9*m3 z7Nbn7!0N}ZRr8R}qC(vOjMEEVFDiINhl9-JDt6)?P+%K@+;@#}<(Y94fSLB9A3g*O z|N8n@l!&T>BHK>tkq~oS~CC%!%G4?h9AG zau~cT)i-|(B!n3lx=XJ>M7bdimZ>$*Q0g2=tTd9Oz93~7I5^Z)!0_S`9aXMQ(=^d{ z4)i1%UG*xSDSgRy0h!>-Cy)*q!$Wo!{Hdw`8kc^$hPSTpqfe1|LxjOTDD9lN$qqX3 z>uXDtdJduYTYyRy&`E#9^JyPiss8u$MkcAGLbd;`4bL2)%JmkI?x>*2E_e^Ab#fe_wmhd4b4&73ZAh;e8KPWt`DJB8KkH@vIi4j zxX|J*9l&q#jNS;fvn=9UG6>Qb2UnpPNzZ3-hZYEfl`QOP)Rr~P;G;bUWy_zd!Xw;R zupLm|v$cuFvEmU;HJTpvoCG)nNo=ndPd=hMJ_7zLMc#YS5naBCQaaGEbTk=d@wSKZ zfx>go8m@!b*w~=O#ZIUnt~r)3Li4Ak@hIgC#PGbxy$__88UkVXVS(^~O(g(9aoIp) z5^bA_Gt0CK`S)#BJrOSlUS6Vg=1=&e6HQK=prxFE(h3D!s&8$mk3VdQbv#C2O7))!J6%TSzvP9{7CxsB0EBS= zM7!hd7-^{Fu=Wn9fh&4R#c+%+$PMH@ZQ)MARdd1FrA%YUNmn3c@dDV{9B5L!ve_@F zv#3x5Jb+@t4KsTlj%z*QcOMAWi$D@~2)?tywh(l$6>eUo8FsOo1jo)<*nud^uPkA6 zhy^}+E*H2Xr=Xj2_?Cx#+@U=9TMzCZpbOeb^xtga*9kIle*)K|!L`@|$}^vFGG+uH z)=Tz*yGXW;23Zf7Tj`FZf590sPW2p*_RMgR=k zKqJ4_QxmS@JGM^94$sAPK_S=MX!iU1Sf1uNpua-H>I${|8N0E~H#@<6<6J#W;l52% zVLU-z{2a8_bJUww7g0U{OR_~wPp#v63gpmkGGGKxdM}>w;Zc@Qe+5_a&QV*qEYy1t zH(@DQt9ul?(cow(NAxZAf!~(r+=__9Fq+)d*d%T>9w&(dMmxM9| z3RxvyD7XGrMK|bp#+}}yu*`3g41G+!1Py#645%1Oa7zDau6ie(^-auY+9v5~BRB9#4#dE548CWD-5M-H?W$ zPj%;zBa!x@zdjXeGkT5d;t{o_2Lfpnh?FO^yx`X4S5CQ8-e}+kS)v^+Q3Cxk+E#9A z`2}gOsRw9eh?A9Rb{LkYKVVzV8y_`fp`QcNGTk4g4v~<%fVy;6aAzDOv?VmbmZ3Uj z0KcXqio+(lOgFx8<$Aav10GFmy$-KJ#1q z_?B^_NuDP8bT+}It>}4JbGQt8_rK$*3dM~plHw`Y7;;**v5z046?tT>1WZcpU$N;_ z`qI+!DSoNiCbl}`THZJ=a&hKQ;DVFEGD)Fb0ceD~O$n=)nrw%=0)}midbR_ENid?l z|Jd@gxxMu-{B{+Muhou6{)Vcon?gQ;I|98O_78l{oYDpkmsKK_L_-s<0zK zOtV>g6Q6P@?V3EEbObYF2^L)}|7yvDq>fGDwKFu_DrETp)P?ncr8)veBy5V*wSR4F zg*6~}3)(AT58(=t%pq50dTd3PS zpwX|!NAPT#M|bre^~fqkGMg#jvl5u9qG1-wL#EMK!8)t-DQ~}|PLLx_p?9(!BWq1l zOI#Y4H@1OqTPO5-Qv+0yjRiw$e*TC{-C}3xOi;XlXLby_7=0#e1C&a=sMm)<9qS># zxPk4%1j&4itL6H}Dsax??b2^@Z*#DIgQAozls-Yq&}YWBfYFcK8q){1aUFMbEC5jY z&_F3nGu?RY0*x0s+@x$UC<58R*&wYMtz*+MEEA0&t4xqg z+%oY2ptN|VlHLz|)>?mPidgv$;d*m5N}WPeZ6U4$df|=Lh6nP54D+QwKDIQ6O`NzX zUZF?y7Io-0*>@VvEH{3v8m`7kj-8qKT?QZo`iyHkgXG&Y-mvp%dJF-m!!|$&nV??i z2KM9)*Sf*BE`TU$H`-Tfr*4h~75Ire{0o3@f_BGt!4L^*|i- z9;(b=Ax<&_>go~z;7j1A-w#Kby3zCm4wt@FcWH0gr+sQ7_P{pzLI-rX-hnlr>ZxK* zg59P1U-^#%5F~MJW#glGl(V|jOA=n93Z)a49HYI`52gK)1B~3W>P583yC8TzMwLE~ zwU|9;Ee8gtp2aA2Vk21bxoTZ4DnOl|V`3s_z?jjVDJv9VZXRe?u(R&N&#C}f1>Ke% z!fXcvzF}uJOgCt_tUyH;oXYw=p7aZ#$j8vkbPV0-Z$#H4!9PlkL1(RkK4jVx4Ga%6 zkXz4CXL?~{_E58)p`$|mm8JmiP}`gx`U6qb14jEe+eOb4p`EgzJBKfVl|tWw)B%xA zbZko8y!I|Ac6nZiQnO#+B?zOPP%=IRNC=&dY>aSP4P2JLRK+XS+W36|@R3{oJ_$vY zdO;nz9LMzlhYmV%={#Vl^AKj63yQ>kRUQ}iSp^*f{q81y_6>-poRhC>XpCHrMX3uE zl2d31C&E*8!ynHithZy%inn|SgqP8Hvh*cmH7$aJslsvUIUX&2%*G>~L?e}E4ev(t zbPCL~IjVq-5Vul8R;DsToiW6kV5g{QX~L5gd^2&NLM7lIwtAX1L^W^ zAk?nN+%=Mz(-0ZQ*V?|uaPYqUYirBTFJJN8DL!a#kO&zKlGbqQnoc3`N;^*Wfzm~A z&N(&KG00Tg9bFl*!o(9~D(yVr-QIRb{cM3LSN$6{PMLOuj~v4L)&1}$@9l~>WHBrZ>f6Ay7%$gIaxXbVN8p>vybZ1Gytm? z6Hyi~i&1rRuDblSY93FIK!^8%Jk}LY#Kap!vT?M(Ew8Meqrc(RL4niFXj9k-iIOqU zuqkg&41x0AM~tsja6OLB%K?N!mI)8u(c9s;lMORI1WpP1UTVvbNS#2pO;fS-0)|Be z3M%!0>b4%0#-RG!kBAVI83?sxuczTz zo;`mwbG-_ie5o-5-eANnRR`k&&U5eL*Pt@Vz;%+0K^lY}1=_WQ%IG?6LqLVS06$h? zf8a=2^A~iVW((>W9+yqfPDi4ea(h+yePRT^j8fxp+9VVBq_0F1!j*vcX|ryo)0{4V zBYsYR?vkC^(PN%sGaSQJtPS#1bs8ORHY>Lr#^f--{O|60G=r?BIe0)_QEPbpi)crDwHXr z4|ejHU4S43eSFgcp|eii&h1Z~hQ>)VxFgj6O7pd4JJ4=?p!9p)wuaspuas;mA-STE z&Y=}LiSEc~G`{?+x<<3t5TEdPW25Rmx3*-_lRZK#Z{2Zl6?#s~4&a@7qP?ZxJK+={ ze+!C-a)Rg?S6E6jnB+~^C1|Cz2xus`GYUF=`6@KAaa2vSAc9Qe!}ZkMtz2^3b8xq= zt-ZQSQanXfe|(H(F6Zgk2+$$)F|9#0`7(ef>ljh>8AFNmgZ;h;85M40*gRO4;7naZ z=WiV#TH*XqZma5Id&O%q0r(|4qUj@^vg7k8mD3kF4PklrpYa6usbn-75?1sq?D~c$ z<)Q~~LD5A$1UYxywI7g9^jA{)Vc7*xr5&Lh-hIIpUAhcDJLS)xvrs377^FF%QzoF; z6E+LVVGGV!U+>2+Z$Onk6EO$4E?&_Vp8B6Ue8_9D5)Sbw65~asWh{rDi+9 z#K4CkLunPXkbbmFb|_~z)XLB74gF#CDkvDk0JX0VK~()5?Uf6vrg0!iZldZM`uY89 z9zx)>5lcyGCc?pA_|ns2+}8(uD?Okk1?5;VzSR9_s0yjlO-pF-(gU#HA^I^z&h`kI z2Ex3T1YH56WtGCZg6WIqg&a35(=0G5+|x0d)qNgh&>M~&qtIir9B>M0Ky%U*rx4qy zO?oz?bh;!todI+m-|}pXUm%TPzX`1la@sK7a{AH>yfR|b6dlzCXotqXeaEjC_@b}8 zp?m?HibqNd_lYFic_6PLdB45*)}WbW9*lRcB>%*C314Em5v1%hKo=aDXG3zBM=!fR ziUukfS5)#i2bSpuWp5bJ@d-5QYa<)T?GU|>ci;UVaW@SUoUW7I*Cc1khzL(_yC!*L7}3 zr4vv)K`_Gz3`%&+1W$}bk;n7<9(UND&7Z_{(q9aOKr2aX=wb8#PEv(=LIuGj@K5g9Uv^mI= zPJmXL0~lsdb(#VVy@SH@KuU5?#ty*QGy@$KXwJs zbgd9t>#Qx??(<@=A0i-&=)6thU75%8c82!*3M5;(lJv`Xw^Uyp7H2t*a060%g0dlu z)%B?7$?G$iVdSGPK`>;Ms%8`|uqTrFG5TK)2a7!I8AacH6u8y?0lcwGYFwj_His7K zL|2r>11KQe<&>&35HRAa=!@Ng-Nq5vpj&I`sbnu8w#>7ejI*t2`qF*uaC<*a9Dlx*Q6LQSZPK$uK{H$H^wZIvL!%rjPq=)YV7+>TGh zItR)^?mYYtS#qlRNN7u7OkABI=iLC(rOwUeYgs?7jjH>n%#3KwF*r2EK$}^)mQI{Z40R z*<+U;sI3=(JRf4^BjgLCgbcIl(9;i5S~tL(I1%t}JSn}<(~|uRXpNT#a(oZsSWgY* z{b;ss9^%RDl#F#?sW}_l(;5h%yXa5qAmh9H@aQO%bDBpJy@Ma-`p}M{H`5d`nLTH_ z35W%#`hty1Q}~!Jg4=fTjOz__z=sGU&PL>IIPS}yG^G6Fw;KWcFl*BhP^{Z*?kj1t27$GqU`5z+@Qn1L z142i&;*pYZ1=b0pHG~^iJ`|*hgy_nI4J-RKP&3g&D4#3C^-u71rgB^ebzKizx8Hj3hhn7_; z$SqVPqnL1bKk%ccF`0-_7xxw0Wi_-@c;2(kL&z+&k_|$>l>WzQ2HZ6+H4Az|r5=u) z&y?KpjrdjW(Mg$%CWPO9_LESLg|k#m_85d4T={c-bz@=|)JdGs8cnKgwAAYyKv=xsGeT2| zz*#ZoZ9$Q``mo#3zro<&9uf=8+d70QO^5WpkL|KQtD0aQs;+p3ib#3TrQ?;pkO^0> zabLQDe(9%qK4RWBhdi?6rB!O^mK_a+3Q^Iu4Qx zbS<2c>M8%(m}mCIbNn7_i&pL|=mWl&?&hPk=M&WU01U|_IZFq$v^9qN@ZpTbHhNw> zq0h1fI!Xtr!|+OD&~e)2lh>rh1JLBRXpMIcN0~^Av;G}o9kOgP=lBYU*un-Lv;=R`rB-`$0w+55(55W&fHoJ{JNsT%Rb1I%|zo(@k*sI zyoIV?LGc2S6EzWO_7?Z2 zL9euoCMX3a-9V6>QK&wIUzOAHgxI*SmS=e`-ayad8A_&wo1Xp>zlUxE+nwe$ISsCb z+UH1}W`HYWxRcPC->xCYvauLt@d)`;8hFw;CHW;9AD_#%2Ae>W>uIQ$SIi3qlZ8;R zz?!XLND1GD>?8REiV0_QsRK`?1$L5SzZ^m>eXnjt6c3PF*+)zCmY9Ev>$;PnsL8wptAlq*rZ|_G zvVjg1*`huTrllp&+LxhMY#7wnLEJG!Z`QT^SzXF07ZHh0VR{6X$O|OQ8EE3Y;Irn!M z5r3xtn5_T_`UI+m$1`50Yq=Z&+Qkc4WW%8cqt8)YSfDt+(kn39&q3A1I6*li{g>nW zbujOdX118|M7_tzb*)B7&46=xGFHGI7CZ5Pp<|~x<}Vyr&hS}-a|j&6^JdIMG&HOq zRT%G6oxA{!G7tzVeXj0n`d@x-Nprl1Q(V5XWzm-wwSfy`4 zS|vnhqP4N>La5L)O4jg6??Vkvb)kD1er)~>HO&B6>RC>&Kqz|)S&1NYAV|A7jk?2q zVTfe`ui6+dIsR>NETG%715M>GKdbxOo8;Xol%>$uQlx%d+9C8VjbxRbJ%FgtD6>v2cufA5_8>LWD=9mst=rVzni$Z9L+WYQ0U2|JW9 zWs1mwyz4A9#mNa#yhd0I%*L~p-Lc*Pj^YgjbvFs~p|AMJ3t;KI_9N9(+K@ z?FIGgB{=C+XDLhh99WR3;FYTt&)$*rN7ZFY(jnW%y`Z#0q^}3EN^MoGauHS3G_*%O z!aYt|0fS^J&=i|Dh$t+#wO9QIVe5wvQ+Bx>KyxHVWz)owI-tN5tD{(m0pG&uy3fUM zR;q8}pyB7hpmYA2xE?R){pMp}sByQ=j-uBQ-$#X}3`_QPbi?)!qfE=-uG55?9)T6T z27sH{t~|>4IPtNFiK5coK|0ZSSz-;Hfw>UZ`~kR>h0LIe#8eN$Rf~5A{bEecCL6sA z$i`xk-;h((qIam#hEWzYvdTSh?LFYJ(Vy4Yj>J{8GpmEmk#6w#itYlr&dfLoVTE;V z4gXSgq@dreEzdCuK)6*cw4Domh;)alv1683kJ}J$>kEF=wKgtSzp>gvl%~1JF zLXRoE_>E3GHV4T!CUme08jz-;V0Dbm&Jp^fP+2w>=qR`SHoww;N*Oh7eVz3Ar zM4|+&1xp{RetyRgZqFfe3;2ZQY$UT(1LJqNLt{|tjsdAwvZHBslUfHo5`(aH^$kxh z?%)}EGsrI9O?`p#00+9}XXuVSNr+XC?D69M{I zSK?Y2r7mw|_X2pM%c#d`R7qIP|5KRO8p7|7*t8*#$p$cYu8G9L{?^~wWYN!uAO$*1Y<}D2R`KYkT<;rd*n3kmwqQc8U_@cR8;0`Nl0@Q-62EfG#ycgBemVpUrVc#gbhlFEXoABNJL_Wb|S|1@zvR_sQ`*jaR=U6($)C*0P1p?xNeP2aG zWP`eHAM@?6_0_yNiSc?0m*&!sh;7DmLxjT_RISaT<89f=tpa`+13nc=HNx0T^?DAj z_zJ$Kd8DB8oy>xO$~Gk~#e91=?g#;{*%Om)xryd&&5n;}gLvlkIZDg>m-7%NXJ!@2 zM&umb+5<4!6?wc8y$ag{b;useLEKNvtc~eJQsQ^uf9fOk;h7zCOY~n~A>g(iRDSor z;4a3dSos0q%6)w!$g&57&T&jyM!#hbC=zZl8P~#w3=E+JLp)>d)7BDyINA^%IlzDF zAYs?z=a@MXf)$WGO+!X}ANtbA0(_-v79knIcl-g=OueWmaRaEK`R^@%;jj6NDD|A; zwi~J?(0q*h-6R@!_lieUZcu))|M0CL{*bB_l|NuX7k1jvw82hvTDl=eX$r%c-jaoH zAyW}`1!0mta6%980jF=CKcZKO(*c{s^U7UNr3GIuVi=1FWFA6vlf<3|iS-;!^*+>; zjxi}pjLTG7>ud{CEN3Pj;U1qINue97l9V8~BkbVfB(r#egKuJ*nbhQ)x z5d`YTXn=2}ZirLbJwiFnwATMb0DY^t>mRy|rMTtW8f)Duy(YL3FU&nLbeYq|wL5k@YQvH~<|62ZHw& zY_NT(ADJR$=BU^2AqLJp9v>rn=r*8vKQx=yxvnv0;7B{vDr;b&(mP!_GJLK76nD}U zU?hmheUG@g>NNcrebY|r`AxF=+TDc-sdRbWJ6T^*t32-xJBh*o} zQVzi(;UZGENcS~roeoeZM#r?oEp??kkTEBe zF+i%?i8hkP+w(x_Ct1DuskDB8rxBijbe(#%uTg{67?QIO4YNX<7fA9w6Hey;sa zyIt{^o6U|h-Ieq~I%N`#l2ekDDO6;g!*O=u;=;+N!DbyIxprdOQEvaOM^*u$f>EZ9 z(Deu*$#c7N^u%u{MNG4&o5A-QjxEX^E*ZynWE$;Nu56T3LFtWpLPtaNY#CopQ+Otw z0OZ93X3XgdV(VGv>Eb3iEF62MF=s_zZ4f|nZc!w9sR(nTDH$k?k`?U1#~dx&1q)7k z^bX8(J}dOJwKKppEn&nd5JME;gGk8E9Y!TjHK& zCmrS@lSd8+bC5eZAs=aJu~SHB@em9twPnM9I)+#ZdEpb|aqZ{eLYjpd5>j6n?4jJo z(LK_X8zgcQZHygMNKyIY*E|iiYsiIBB8;|gi1ZEkJ)k4O!=t5&!J7M3({^= z0H)G}X*-xO$rRo)Zmpzsz%VfhuH_+|Sxc23v_f}seF5c^s}>I6!LK~~lRao8r@7hs z&{7>lR|_WmDxN?H<7(dN7-CE8_BqQO>7I9|#sFJI=v?#_I#X#_kWm21sDW)Rv}L)z z>g=I(FUrzvH-XjC_o@x*j@_MrqzU6VKUTFt+QTnKd#?^bX6+R8w=sYytx$c@2%=nL z$uSxwJLJw3d7d*9Fd)S+u1AvQ>W)?iuwvI#uXHAG4k1E6abr$3<78d8!5}+E%%EZw zL&bpd$`m)n3k8Rk1kE1Q^Jt>dC#HDAwOq1UL(uDTOoZLe#C2kwN?03rrmPF~D~X%v z`)f$f5D9&O4Y`hA!y23y-TU~kjFMWNQC$s5ZOka(y+lQxB5h?o0R^nr(Dp(*r4I1b zY~*LUN4>-|KEesdI$G#i-@tP{$cMf-MKp2Ekd8qUc}1I@X()`lH%gPB&xHikd~L)d zlUH~oJB{SP{+DRTyT8R<&D$DG`Jro6b$Pb<+=(|=!8u_FV`7kp3>?HAgJC?_fnh#z zkPs$C&|#u;5DHhh%+m_GY0n|D<@(=&MF3);0u2}VhlFv9*Vs0$3%|~+)t&oZ#w;Sn zT(mmCQITs8Wtga}11ruPBdIGSDy+0rc=Y!x9-*uZDyt4+!O4s*KEzF30D)}~pwOJk zG_z}%nkuJqD1FI8k&xN83lu1Z;Miuti^kv{ISvGeJ3%$5ViqJbIu0pkMw%V@7yy|1 zfDFF?3Vjj$wlQ#5nSfdtf2VmL^ldU~jTsimg@XZhFe@pWCDHlfFO{oeN`G3T0Q5C*oCa zAK$q(7EQpi}CF1WDMx8V~>gPZaFi#M~nY2_y+!h76g_VqSO?rB0wBvI?Sj zPEB_YawNux>M1^y80aP3rxS*M`#GQ^o$-ONil7-b&!J%ljFNtgwl<3|DpHUI?C=#G z3tbjI8i1M{#aCj>Ca1Jr3f&#f7!VT8(G~PpltH{8bX7}__?R5@q3j1RWen=m=^w8- z{J0CFg3*AXV-P&O!Kf$Qft3dE%FU&2b4CB`OaeRs?8Z7UfT!J%7o15@V z1IH??9Gi7n2sqP}9m0rdag3=lY$4R*`$!MbkoCBX5N_k~Tm(qa-Yk8bxN!TyFWAR1 zNLqq>V2s$=yDlB#p3*!jgwmh+!;|39-?)5B3xvr7G=}GVMtQ!E+@Udp2^hwfPflx8 zRQMF__{!`wvPb=pqi{HG72MHGdyA;=0?!-mO{pV@83dn1$Qxgl!DAmK`X1!4Aae}U*IgKDf2q845-^k+#=lsjft z(h;Z_VU$KXWG7C+2U&%#Pn(oJZpv8~l^>RO;k<*PR#}fcQN5nwN$aPQ8pgZAq)dvC zq)=f5;cD(Q;3G6$IWr=E_z6vBt1-@ngkl;dGpI8yVmy*Ls6t%?36+Lar(33T{PJF4 z!a1~9e>a}zxdp{D1k`iJxJ$_Xf6Xk<+3^-IJlF!-R%A=hSj@ZPn)=}hs?6p?BZW^Y zU-MNqi$19aa@56au!`PszLPh$;X72D#vlR8fHgk{K!vUtHktN7PPmCyU%0xyQp# z@q=_gI3mqSl(r10$Y5}(RNcFk94a{qA~?+}#^Y5|FPdow^LXKLCd%RoIP_(z(`(!j z1P{LZm$=g1OZ?5*Qjz`IU(g+21Ey0RfsxuVzySux8i6uP1`j|kiWgioSvdW-gaeV&C|J<&?$|*uHC4CP-@s_gKDP9}lpG2$i*M>W6lmg3 zs&O;JdI5F2f@;)Dv{AEjF@0Nf1owTP1J)=si;uGYIfzxFg|n#e4g>p?N#C>+6mx(o z_LhjINq&W1Fa-_Ynj?9bW>#;9q5~i5KYoiBl*5o=$`QW$zZYO2Ra^FNVcuZRxfmt@ z0b_nc&8apR(Udqs6m-?y#u7P%Ds*}-*c_x!89XW^i>^}-%mM~_2UMmN zU|hH_D{uJK^QJv88TEieHUS3O3Lwz(t2xTl$2*9ki1R`?e90Zujd$NgnL8%eq|q4w zTQCH`7Iixf?+P_;qR~&d zM!~3b~!_MK*Km+7y`t&oWcClx|jAp4c58fKh;kH3?tVCJ*kl~c*l&TDPO59%a zO|-&SA+E&(rwhlVWsc5&-KU1CN1|T@J7WqANzzSF#)k;U zYiRs}YcgZ7l8!^%6Ow^M63sl9N!dq;TljMEj5_FZqO=^@BEWAmQFkP-`iMaA2%D-- z5E!OVJYam&yEW829Z#1&Ri%foN(Y8ZKSZ1M_VA~IERsPhHH)b4E2^&_`cZA6g zg`sH5tlOH?UDJVI7|Eei>{UJtYGfd4woNcpxkHdnWAux4NjwvdY8>1OX7)>u9F4BP zM5R5@x^TxfFn22+AyUquEn(7c!+JZGHmF*5sF&{{aJ~RWB&Ii4%$Q>T6HwJkbR_Jr z0ETb_rJ#8E2*$#63@WB&d_AAlpeKBilUP}XNJ%%KA!q24XCL$J;6SX6*bN=U=XDUY zOgVaV8nc9gev8(}M%<9n2f#$^B=9}U&_BDxdxsdzYzNIq-9+z2;4W733@A~_KD5=b#7UeV!ba!q|}+(@15?j(qm z4$*F>Mbvp7nRC%>Op+v1(V|k-2rYDY(o;}R=+|02L=(fcrM{YUl?#B3x z)mrvmln1md4D}gky8uZ-cA)UJqk7KUR#@VdepVd^nNGr);>xyw8XbYUa7K#wB_$GG zTaki1Z}i0eLVF9IpfIIgXr-!&<%GazK#<-=cOrquC$P7%nVjLNWpa%<-lsfAu;=Oc zSq4E?+v6-_9F6-`b>-V3*q2D3?3;wVCP-&i-yW~%HDp^3Mo^w&(3E^a6o*C6f$vN; zs@IkH)ycn4UmD`tFr=3Ltj?RKz+Phx3#Fr>_5s_9jh^^i^(ijr5X`%pY=%g!92{8)BNpaIHRYH5L?>C@?PV_RNO3{ zFz_XIslF{VZ>;0cLSY6wUm_uP0c%+U*mQ>jCC75_gibL!@C(uty{7{KLE<-5i8v}I zjFq8n!^TVrp8_v!1*Ot-mY#t)rR_QKlpLGD+{89W{%(M+pxMNj5?rSMaeIZ{F`gyn zx4?>VzNm-PaJT3~S_J=f7Xn~eFFvAX;duR>=4j~zjEkJmmSeXum!8sB+mQ4j#D>FK zFDb>OuA2RZm{Hmoc;(w6V5N^7UhBSw!y~CsF#CmZvTw{#C0C%CzXE{81KO?YP*kI+H%!+1J|Q*%0V9;9o? zoM8zTS(K~O1#YPbGjGe@6I#m66sJ;@o%6Vp=r*ZvL5>|Ja8Is7-PL>#W9iUI>yX@;|Pw^Z)t7-chMHaJtJYnixaXdg(zTsqt_^H1f?@0q{F*8CP&n2<+krxc?dl( z3%ZC)J`*CJxyeKXvP;gshE^+wi+kSAVQ8M;_VOzzjT!K?i$X?s6(dIMh5|H-;l90| zu2kt7^6T6@lYL71EAGfbsK!`y(_AFXR4D!6oiVsTrdDx%7PKk-YM|Cj*c6MC*|Gx3 z0deq=EdOlI(+@v~C@Vi$JVYB}_nIZnHL9TrRL;?YAC&EMgAvD8sb3gsZuIWc8xsB^ zDvP;v*1?*ma6^H%&>PSm-T_SLdXPA@2Y}-U+ljqPOj={?WXE_~!#^SHtUl6bHpElK z0|pGep%n789pM(ccT<-fp(bOZUZs9di=20GK$jut80{saPf&wigIx(M;JG}cjpZGY zE1<|Hp_Y9EWx)wjy29c?`UBsg3^;`FxFn~SlrW)v!GrX45z0fDjzm`}fL-%($LO-} zJefh5RZtjzE_2b08cv52V;_nN+Mcc<8ENnaujcjI6DTizpkU&wiJN#re{8=*V+HvD z;(30s13KLqgQ(1p;3IsOlt>W-%dLQXTnOLRJ$$)(Cdee{l@2)+Bcom@25>m-fr5$l zID>^TCRdTcGVVE3O)JA{6Z}oX-c5FDcHg?748XzZeWd!RZet5W(RC;keQf3fD85aI zaE6fb?lE&VJFNrS!$EE?^l7O+Tg`G1jnTH5aCT>SP=)}+3>35z?pD|-*~g@No{uLS zz_|XIv#<9!c{NXNQ#T2?25sg$7f~wRB;_9WeHiNs=|fMHk%PhnQaUh#JG5@dAPD4> zn8(@(aRfJ@*bt&r(fG)3;ecZzPnZc@*6~!@V+ZTf4+}YlvNkzG#yj7*yF}Wh3Q@W2;h}@;sj0rrN+Nyt+2)+4sq6{pVj@Q zY04}^)Y5(YT=T0jUN)@XV@?ej(d+|1H4xWniqKdlWEY_t5{<^5{e$SYO>( z_kqk927m)SLD|CLPR9cWv0V%cuU900;3jMfFsC=473L z2EStz$+K!>XsxV6;|LlbRxohid!P&y9?A%)q)VKf0iDM6u67^Bbz)XBqu*@m1g3Qm zL~@3`$PLtH;*{5E^VBoeijD19n?-9Pe1N@7jFm3nC&`#v7(T-2g<(OtuWu(o!*|WN z3$BK%9&8)VR4h)TrscDm9wAdjjya^MyxK!bI1a)X9%s z65N%5eH>2FoR^k~fdp~TM?75QNjBHA#%X$*g;XieTafHa4fmxOK48zV5zyYVb3->r zf+2#gx)Y9=$uD@We+=ab-b@c5zcazITt|;&Ve3&OA2jm$F(xjQ~E#;;fD)&jlKWS z@UgynBmmEtN_1qhHE>fmptqc9ReY47qw8hCRC;;=#c$a^nxQX64rO@(v=uqX;LMQW zz7D0JtE{6JgjD8L(GJ{E_|v}o935)TdP=ybqZ1e+(gT`YykW#~Zs8lxGPN~OIOSY8 z-5BUURF#3!w5Ycj7JdLPsXuyJ%DG3%fB!xG=ed-^lb1B_p<{Ir)A|^C0!l=|8padb zCoxv9P!-oB!h9LJuoidxE>I9VyzFRQq)=(h!E;w=wlZswFN2O&ImaFDWv0hyTntO3 z1ae!}e&Sm`Ou)c?8^ahOgd^_aNlcs-#-f;ck(Rf3OPL-}Hz9$_fG*!g!xK&+jAawm zS02RF3r&!T!jeaTe~(;%&_0Ed3Z%v@e99hSPt`2_uhCjZrhCqI>YjUO`t4z$b|mFB?!$xdz;n0(p> z>d_9eYLHDvLXemXfrtgooU#UeNO>vsSo@GtTK)u!o!u78(G5eDq$p_jN^;O zJw3Vajq8*n^LNPF%@`eHj0dqLo8@^TiRS4cjS;fPuPmL#Gk%8J5~C}e>ee^GT4&}_ zpADA7-y`Qd@?PLW$brN|=#+6gDZ8hv0ZA=C+oH59c4`d@+2=4`IMY=nEEl8kfdZr_ zhF1HYJ$X1%I?p3>26PoKs7^+~59j&eo}+JR1vhT<$pt=;R-vR)&LB8-7uUKJnk={i zfIBs?Y_aJh@%c!+PcjkcSxi!5f0wRBLYO*TV{|e29BH|;Mx6xAidV>mWN)CAcrQ#4Y}zjHE_6TigRQuB zCTTb3;}tAYOEJ8($rgw(d^+;c1`Z;ZVOET~h7ZXlQOz;Kb0XsN?YN(&Fjp1q zwv>OiW*#^bI{GG*Z=#tqF~^EyRgFY7c}x&a)YBmV?s->@@fwQFg>Z2E1d=GCl)~DH zb{P*SSB85F!;!TIDD}7gc;N>(iOQ|NQ|sPBJkvrTC$fLgzTr46y3}5^#b#GA#^WYz z?qfHKlx_P_A8+a*20a;sNca>ODf#@Sv8W^<4z(-07|kUowWpw_N1)RwYvj z76X8mbCIAlfW|q=xWfPwe;Wwp`~%Nn6SDM$Q==Mq3YMz-WZDURt+Zq4HN?!bj}F7G ztP;G6%-nejr1?y}eC+)VRG~79p*{s%OMhjgDT82?37As{2m5ghM)pJka88mBN0PhH zJ+;PtG2aH+bRtCO-G7ZMnIgt$k+aDSJhzIc6(<(OzIc{{;u(Wl;v87|v&`(^?W~Hj ze?!iu&3uu{Q-g#P4!Oo~nUJ+3oI67JLla~vh66<%_CfQXd4hyF`MduFw*4c|ZO-wH zkm{I3tp zY@p#?bI?ydyA9Tk(x?2*G>u7`WH;S>22mwi8r!#=cVyoKM(lbwAS*l*$zK64Uwn)* z@kynhn(;hQ=c$kS5*&}YaqzJ=L7Bxh#t?tQ#6!7I5Cq59eEu9ajC*gs%&FQwv9}7g z$O7ga(uoTE7SJ%Rb>$Q+PE%-pdS%I6D#j2Ir9JLL;}<;7UEqQ#Q_~FREZM*0Bt9N1 z$m*n@YjwSa@XA(vt^D=(K@H5IB8(u`d>AZR8T;?#WTMNpZ1Y719t+*KY<@qg(pQWa zDQ8SzkgU+r-75Tb8?=y%;B~AyL-)l(2PpKuy(kmY0{AxAC`0i~_+b7v6i9_rgfnXV z%tLOL8Z&dIoFai{1f?Os&sZ>=flN7=;rtbeV7yK$u-9O^4U9)ciioPRt9 z+-g3e+s>TUX(`lQoNCVQ$r+k%y?~*vQ+;KZVw^h~*(cg?|BaoT4-%AqwX}YVTQY$W zPUz#P8(@twt&}qlQJ zyp}E~SzkeS|6+F?R@QLQrd6qi`B&`}O%NP6A1=+oh4P1=86HkY<&=+L>xM@1uApcQ z<;&MNCs>(|+qQy1CUy7c96@Ixsj+ebRD~OOmX8HLhmvvxYuB{2;srP_#SLn3F2KAW zU(sFYE4hSn^;?LUy@Gpxwfw~WDRd_?8APf1Ae?Lz;?X3+U{mkf#gk_wHnSl3P7pU7 zFEj4B^Jk3mHzu+xxa2n`tEnPjNKruHiYZ@(jE=Pc2H*9Q69IywLv8uAwBsGrfaU z>L4l;T8eZVK=)fdvlb|fAraE8JfT6Z(LJT@$T*9w=cLS$VYAfEFr;!btf6>pA(#^E zWhw&a>|37EMO_KK7Yor)OSQhsmBTI|ZM*CgT{tYC4#6D|%QoX1$cpilCYhe$^JtCd zR?d*$r%`83-nwNzC(Zuo42~4;*Cgh0wmArW>IBoB=lN+q^oD6|v=98yE_7dbWUTkZ zow~M`-$%kB)oW;OF42UZ8rqMp8J%r~d#O$dO&m)irVFB&Q_SfiWI;J0u11!We;+X9 zkC2IKl%t*o1taD$?-%M2T5CwbTdO?wRQF>)4bq0 zd4dIg=|`TwIs?;`NB7wVQHIFp8agKEJU%I$W!pMtT5`zwS&qh`))Vl;#?cjl@i9htG{P7-Z*-oHSm}7LJuko!>`pHnvIs z7E3RYgrJ!SDWy(Hw97js!e9`I24Too`h95yluIz)mLpe7)0vZlQD%cgUn8uvU{5 zGNP5kK$y3&Rzg+B+z8Xv;)%NRknCgnIl-|kG4F+j^V@w~(@>r5Y6>|8b94HJCA2zC zqSczigW@%F6&;$ui6PI)UgvhN#m;K{EIyJL{Cw_ zn`|jBCV9R`!PwxmejU#g74+4Z_1E(L*EGK6Jlxvg`da?d zlBttAqrOti8CyfhFm7ia=G1xjF{p^oVuyktrK6+ zOSDzE(2@z>$;S!5g*u9zRJ2krajiR`K%%pzGX0vjQIqAF8W(1!LtE z7_`AFl+49#>*#{<@EYeHeLdZYWdN+UXCPg;ADejQgbai{m#T*}_%H-YPcbk=L=+}S z4p+jOoGd5{1?Q=WhphS;gf&r%3XX@2JR*d^Lbgz1wcBYip{68=5Uf6l+N4`$|ZrpM%v+S1SDYJ>b_p@CWx&Fj8bp_XPT5;S$-GyJAhbBRqiSY8 z3PRTK%w#P{~zqf$`120DX*Q{Jy#LDvOqu8^sm2j^f6xANSRf!HZeT!&+j zDt)dhl)S*32{hL<)#C#QqX`8JIy;gp#yT>G?Fg4x&OpAzz%p6K;N@dCQ5IZHI&Js~ zqiSSv&j4sVJ1^-`7=e1qcwWQ&JEOFc~bH?bt?$Ph!fWv7A&cQxz z-+@3NA&`U+7!f6c2nHb#IhdE(73N!X{a>;ha*IF}Fnic*^<=#{@p~f|XS7f}c z0hL2Si=Ln&gjnwIKMP16Zj5dYzoIckce+vA*E(Qu%bKp+@v@c5Q_i`SzBx&fe!gOq zOZ*IVjhIFnF93#pErF zyXO+K&8qQEJZU>Js|M(hu_E=8>sn6(=GLBQj(&1qpD#3EZ~2HS9+yzUfNb~;ZE>}8 zt}EZ< z+Dli~10?FS7!HE3*>$8;%-6A^?uo`!<4o(M)d%pm)}%`&vcY%;HdkI{Qz#U+C(@YK z_-I=U^89r@le$I2YW*5#>3>d{1tz@1q*RMlxk{a*zfi-UAu}j|of~BI1WimQxIB$y zr&Ry3+UOG8b3aXwQ-i9T@lGN!R3=^-a0gijKuj8l*S03Pn>X}w$4BY1cMc0|mC;$3 zR4hsC=Tm00CjsaeP+m7-om1MVxK_Hku#%zbToZLgcR^C|X8Fu}Wl!pk3E(+ zxjiwwZ9tlhboPgYj{}etf7RT|E|>OCLcOMrvvbv^u6J`;BFbOD;AE@~go&Nz4k5#U zc0(6N+uPQUNnr}*YFBQ}ZjG=kqHpxYa`9(5*AY)$Eqzm?ipNE2pF=^NKQieg%gv(C z3az&cauUBLpR}rh`z7gJ7h8JIKmhnw@nxk%O!@Yskc5Z9k#CpO&@VD*l-H8eInS5SBfA3%^B~RZik%U>pYI|Sub@{3CUVRzoy^h zo+v81FiEfsbN@RHj|7Qf@e+akSW9MPTcy04GpZn|n@{Ug-Rvv&SWH=gYe`b;gc=>U z%fOdxsE*Fl*xjmbvH3C&%VcUBb6gcY8g5(vtG=m9xEe{tWJP4Q#5h+a{KrOrN=Q0X zY!BeSuIGPPD^!_-QYDVyO<<(*ysi(7hNM=_&p*>|&aq-t)H#5bo$nBUP>=3(L?UsL zUEoU_bywg;_2(k+@kXqdl<-X67YY1LO}3}vuA|7{#pFxzv%M_7EeB8gQ)~g#b6YF% z#h2}a&2u941FPeMbW_sbbo?P99Lo~RgHW}R)-#?LMOSrEJ1O$1mdyaAO_W70hy#83 zsjfdPE-4p`mUDuhiT8S4WnKSXm-01ZBw0HDf{7nN8#ijyb#kkU+dgq=?N0J$U|+zY}-o}Ov9@nQ0ZyzjwcV0ll`_q1+8ByX z|NbEHq^gDDgyJ=*6s8$FwazodM)PSg+FbSY)$N6^FbywQwYK@y9xDBbBDufS zGUZM?u7`mV>5@-#gfiRO^_Wtx(t%C`=3cs$z{0ygm~7jX`2SHbEzmt9gK?+YCH5Bdm`C6 z7MKD)>!i=St%xXM4X*F34>G`GQ)HXKDh^ti`l6mcd8elha)G`&&_6gvt#c50=Xk`3 zs1|mTxm^LSxr0UnhMKco$+*N{l0@qLb?s@L(UElLiy!JJqj3kVP)7&lR9GYctV<~9$|7rZ8@EJ}v5GiSpR;oQi7qD} zrGK38ofE9j)tX_M5#Oo6 zjQylrZfPlPp9R_!O$MNP-e;gqvB9_~U1f5U@uAkK zY&20q(w8NdYpgOhSUKNDp5gBs+Y(*5i;fj`Rpq!%LqS`j_{8?KU3m7D%#YCpZ6q8I zFu9Dh+Gt3UTi5D0#*U{70Lt#sw1Q*OAR9Qb2<>KFuV`x65;`QvX^%b_1)T{XGs;+}VZ099= zZ4(K8hbhZZU%}EJr$<^`1%{}d;2CrL)Z?uv-_)n7D&H^7g5^luvo|QVSBzZET2;qstq* zy6Z~4_&=ujv)nsyhaR2#`8 zs1x$5`ZVew{oD!f%Pz@3(K#x6rJptg!Z)1_P6iG@i zlO|4sVgkFMopo{@>^+N(!WsJubv=np#G2?n8WNGnrsGi@cSN+=R^uahgSIwKA%*X} z&|Md_;MP~(6i>~3{}m4tMLUTbPUK=dX7L0PU8a%t-Wy=LldY95F#B~o5}krtU)SnI z7bF7&n*jCLX{~L7))GL#`YWX7y6D{2M5@lL2kuq7DGK*ht!^%&O=GPx};SRfX7Pg{LF=PoBF}2v zgXq~H(J?+gQ^yQYrN_ zxDDd6ST_W1S=_~IJ4I7@qKCgJtd^-%(Uyn&#C15+i-Lp)$B&xy%d$01Hidav#4YkQ zfg{Z*)Lm|DX{op;M)G~$scME=U2%Qvp*y7D=BimGuQZB|j3BmSFN}M&vt=`C^#d2M z5mn52jotvF^H=7jW|H;hZ#sPROKBp(-P&Q7g>V_2i9 zkkz5@>dPQmHbJV%mtrA*t})n}#vfP+30OwrP_k8Rv7eR9AD6`5c2yK+tD>Hv{*ALo z`f!uvrsp2J`>NP>=C@<|Equ#Pc{@}dl?bUP_bf>89Tz1QO;w#m9c&5bySSyDPhegK zQHde`spSzjB!!nQETzN1V0@*ECyRuu_E|ruLvVc#l>=0yW$L*3AOYs*Yv2XXNys2J zB_wQ1Vq?OIz&Od7^jX#4G0;JY)u1h%DG$10i47$-x^R}`Ov~6YK++$|)2vk!IYc{1sV_173AtONszCeHmIvc1k@ogMY~bW48#Tc$ zJgV7l)leEyLd@D!TrMFeaKzqRfPN>!xUJ?7Q(rBteNWPnKE=RIMCqO@na( z-L;_4T-2uluwrzN9W=tVKgH5U@^HS?Z~Tr0LUp;pm;(0RvkOT4WhB*4b#Ux;N+LYZPXvw)fd?? znh!LE3HdPN-mCii%U=;wpIDa%;!IbH-<(2lM|9r{z1x(^TfsbN2qYn+_Qw`Q3#L5# zI`{A*qZ5qNd^?tIG(Zq#8snl?G4D%p8?N^}tLfRoLVdcdUOEdtST)(qlbt`~LL(4AKODeepI;JVA@9XbD zST=y#spF@vywzqK&8h4KoIv-oq?6hAU41~$6=y1Woz+@G)NYkXC`G63w?6FwOYD?g z$0NzJ*esbh^MBeP-$CV*k{#3;d6)fNmC=w?b=nKfcQh=z!2f2yntHk}h>z@zI6dZh zJ8oyg?J=I$PFEPG*A%EN`{R18?Dq!ycQ(hODau(OdoVtj5At< zK~Cia(CHvy*k7=#S48(AGMb@%pVMH&6-zH)nq-0RiwIomM5i@YIr^M?`nbNKg=TjD zEC|E#KTY2A?aLZ=5lr!KrS@xY^0lOW++*A25-(?tbgC=Md-mwmao4uR%d(U{i|!(J zTMrsLGR6F+_%tZIvIn}JRBMmh(q`VE8o+(!3B=FkTKV#{l#;J}1PK}=`Vgex21;Y# zAsOqU>0HpxIQ=Iwz6mkOg-x|Fg& z%P`kEI=J^^vsj}@Mo;!WBzF1aMX{%m5S*u;*WZD~YS*wVU$FN$KSm#APKm_xauHoo zkLV*-HHdLdzZ@NBp-Rwi>S<$M>U;80WIN;?swLLu`B_b1v=e1A^wLM4r^QNw2&Cnu2vP<9(QRQF6?mX?{U6 zfm537ul0naCw2l(YxLXwjT)b^F7nHhV(DDciu$^CAUDOiL9_C4=Z_-a!sng3{RibA z0qjYv!P^8SM-#`xqWNlU|F6!l53(EVjHHWkuJ?7RqoJIT!lfm)dIqPw zU#=II1v+1!u~M?9yeHorsefF7t>RSibkevYIcz#-c3Y_tl>1Ao6?cj`4*k1?VO8|8 zq0vF^&ZNK2=6-^lyJVLPStoBJ)qJu}q7til7BTo}TTforIO831<(^nD2|F2B;I^Y@ z`XwYgE~vq2E3jE6@5CEMosyr1{^s?x1am7%ZLtfQoi#e0K&oaBP)9*8A96R8IaEWL z{c&n&(m}-qNL8uWEuY?K#RvXOdvim7UzK1od@crvm*K|p_IP(?3^=q2EKqh4<%xSd zM>69dtqdIm@PIP2VGP&AZFWX{V z88?yA1S;&rYyE}}1g65#S`mcLr}`{#3uftE=c5L|hmRedo(`gOrGQDd2NHr#`EH`V zeN$I#t#0`&`{;35OXV*_H9?7fyc0h-Nx!itHP``}3A