实现ui界面

.idea/.gitignore

@@ -0,0 +1,3 @@
+# 默认忽略的文件
+/shelf/
+/workspace.xml

.idea/OSUM.iml

@@ -0,0 +1,14 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<module type="PYTHON_MODULE" version="4">
+  <component name="NewModuleRootManager">
+    <content url="file://$MODULE_DIR$">
+      <excludeFolder url="file://$MODULE_DIR$/venv" />
+    </content>
+    <orderEntry type="inheritedJdk" />
+    <orderEntry type="sourceFolder" forTests="false" />
+  </component>
+  <component name="PyDocumentationSettings">
+    <option name="format" value="PLAIN" />
+    <option name="myDocStringFormat" value="Plain" />
+  </component>
+</module>

.idea/inspectionProfiles/Project_Default.xml

@@ -0,0 +1,220 @@
+<component name="InspectionProjectProfileManager">
+  <profile version="1.0">
+    <option name="myName" value="Project Default" />
+    <inspection_tool class="PyPackageRequirementsInspection" enabled="true" level="WARNING" enabled_by_default="true">
+      <option name="ignoredPackages">
+        <value>
+          <list size="207">
+            <item index="0" class="java.lang.String" itemvalue="flake8-executable" />
+            <item index="1" class="java.lang.String" itemvalue="cpplint" />
+            <item index="2" class="java.lang.String" itemvalue="pyflakes" />
+            <item index="3" class="java.lang.String" itemvalue="textgrid" />
+            <item index="4" class="java.lang.String" itemvalue="flake8-pyi" />
+            <item index="5" class="java.lang.String" itemvalue="clang-format" />
+            <item index="6" class="java.lang.String" itemvalue="pre-commit" />
+            <item index="7" class="java.lang.String" itemvalue="pycodestyle" />
+            <item index="8" class="java.lang.String" itemvalue="openai-whisper" />
+            <item index="9" class="java.lang.String" itemvalue="flake8-bugbear" />
+            <item index="10" class="java.lang.String" itemvalue="flake8" />
+            <item index="11" class="java.lang.String" itemvalue="deepspeed" />
+            <item index="12" class="java.lang.String" itemvalue="mccabe" />
+            <item index="13" class="java.lang.String" itemvalue="tensorboardX" />
+            <item index="14" class="java.lang.String" itemvalue="librosa" />
+            <item index="15" class="java.lang.String" itemvalue="langid" />
+            <item index="16" class="java.lang.String" itemvalue="flake8-comprehensions" />
+            <item index="17" class="java.lang.String" itemvalue="editdistance" />
+            <item index="18" class="java.lang.String" itemvalue="scipy" />
+            <item index="19" class="java.lang.String" itemvalue="tensorboard" />
+            <item index="20" class="java.lang.String" itemvalue="soundfile" />
+            <item index="21" class="java.lang.String" itemvalue="matplotlib" />
+            <item index="22" class="java.lang.String" itemvalue="torch" />
+            <item index="23" class="java.lang.String" itemvalue="Levenshtein" />
+            <item index="24" class="java.lang.String" itemvalue="visdom" />
+            <item index="25" class="java.lang.String" itemvalue="sox" />
+            <item index="26" class="java.lang.String" itemvalue="pypinyin" />
+            <item index="27" class="java.lang.String" itemvalue="onnxsim" />
+            <item index="28" class="java.lang.String" itemvalue="kaldi-decoder" />
+            <item index="29" class="java.lang.String" itemvalue="black" />
+            <item index="30" class="java.lang.String" itemvalue="dill" />
+            <item index="31" class="java.lang.String" itemvalue="onnxconverter_common" />
+            <item index="32" class="java.lang.String" itemvalue="typeguard" />
+            <item index="33" class="java.lang.String" itemvalue="kaldialign" />
+            <item index="34" class="java.lang.String" itemvalue="onnxruntime" />
+            <item index="35" class="java.lang.String" itemvalue="kaldifst" />
+            <item index="36" class="java.lang.String" itemvalue="num2words" />
+            <item index="37" class="java.lang.String" itemvalue="pycantonese" />
+            <item index="38" class="java.lang.String" itemvalue="isort" />
+            <item index="39" class="java.lang.String" itemvalue="onnxoptimizer" />
+            <item index="40" class="java.lang.String" itemvalue="kaldilm" />
+            <item index="41" class="java.lang.String" itemvalue="onnx" />
+            <item index="42" class="java.lang.String" itemvalue="modelscope" />
+            <item index="43" class="java.lang.String" itemvalue="numba" />
+            <item index="44" class="java.lang.String" itemvalue="bs4" />
+            <item index="45" class="java.lang.String" itemvalue="jieba" />
+            <item index="46" class="java.lang.String" itemvalue="scikit-learn" />
+            <item index="47" class="java.lang.String" itemvalue="dgl" />
+            <item index="48" class="java.lang.String" itemvalue="setuptools" />
+            <item index="49" class="java.lang.String" itemvalue="lhotse" />
+            <item index="50" class="java.lang.String" itemvalue="requests" />
+            <item index="51" class="java.lang.String" itemvalue="numpy" />
+            <item index="52" class="java.lang.String" itemvalue="datasets" />
+            <item index="53" class="java.lang.String" itemvalue="torchvision" />
+            <item index="54" class="java.lang.String" itemvalue="selenium" />
+            <item index="55" class="java.lang.String" itemvalue="tensorboardx" />
+            <item index="56" class="java.lang.String" itemvalue="lxml" />
+            <item index="57" class="java.lang.String" itemvalue="torchaudio" />
+            <item index="58" class="java.lang.String" itemvalue="transformers" />
+            <item index="59" class="java.lang.String" itemvalue="dataclasses" />
+            <item index="60" class="java.lang.String" itemvalue="loralib" />
+            <item index="61" class="java.lang.String" itemvalue="netron" />
+            <item index="62" class="java.lang.String" itemvalue="tqdm" />
+            <item index="63" class="java.lang.String" itemvalue="regex" />
+            <item index="64" class="java.lang.String" itemvalue="zhconv" />
+            <item index="65" class="java.lang.String" itemvalue="pillow" />
+            <item index="66" class="java.lang.String" itemvalue="espnet" />
+            <item index="67" class="java.lang.String" itemvalue="flask" />
+            <item index="68" class="java.lang.String" itemvalue="einops" />
+            <item index="69" class="java.lang.String" itemvalue="joblib" />
+            <item index="70" class="java.lang.String" itemvalue="fairseq" />
+            <item index="71" class="java.lang.String" itemvalue="huggingface-hub" />
+            <item index="72" class="java.lang.String" itemvalue="xformers" />
+            <item index="73" class="java.lang.String" itemvalue="nvidia-cuda-cupti-cu11" />
+            <item index="74" class="java.lang.String" itemvalue="fsspec" />
+            <item index="75" class="java.lang.String" itemvalue="nvidia-cusolver-cu11" />
+            <item index="76" class="java.lang.String" itemvalue="nvidia-curand-cu11" />
+            <item index="77" class="java.lang.String" itemvalue="filelock" />
+            <item index="78" class="java.lang.String" itemvalue="lit" />
+            <item index="79" class="java.lang.String" itemvalue="pip" />
+            <item index="80" class="java.lang.String" itemvalue="safetensors" />
+            <item index="81" class="java.lang.String" itemvalue="sentencepiece" />
+            <item index="82" class="java.lang.String" itemvalue="urllib" />
+            <item index="83" class="java.lang.String" itemvalue="certifi" />
+            <item index="84" class="java.lang.String" itemvalue="nvidia-cufft-cu11" />
+            <item index="85" class="java.lang.String" itemvalue="accelerate" />
+            <item index="86" class="java.lang.String" itemvalue="nvidia-cuda-runtime-cu11" />
+            <item index="87" class="java.lang.String" itemvalue="sacrebleu" />
+            <item index="88" class="java.lang.String" itemvalue="sympy" />
+            <item index="89" class="java.lang.String" itemvalue="tokenizers" />
+            <item index="90" class="java.lang.String" itemvalue="Jinja" />
+            <item index="91" class="java.lang.String" itemvalue="portalocker" />
+            <item index="92" class="java.lang.String" itemvalue="pydantic" />
+            <item index="93" class="java.lang.String" itemvalue="triton" />
+            <item index="94" class="java.lang.String" itemvalue="nvidia-cuda-nvrtc-cu11" />
+            <item index="95" class="java.lang.String" itemvalue="bitsandbytes" />
+            <item index="96" class="java.lang.String" itemvalue="omegaconf" />
+            <item index="97" class="java.lang.String" itemvalue="psutil" />
+            <item index="98" class="java.lang.String" itemvalue="platformdirs" />
+            <item index="99" class="java.lang.String" itemvalue="ninja" />
+            <item index="100" class="java.lang.String" itemvalue="nvidia-nvtx-cu11" />
+            <item index="101" class="java.lang.String" itemvalue="peft" />
+            <item index="102" class="java.lang.String" itemvalue="charset-normalizer" />
+            <item index="103" class="java.lang.String" itemvalue="hjson" />
+            <item index="104" class="java.lang.String" itemvalue="msgpack" />
+            <item index="105" class="java.lang.String" itemvalue="idna" />
+            <item index="106" class="java.lang.String" itemvalue="networkx" />
+            <item index="107" class="java.lang.String" itemvalue="more-itertools" />
+            <item index="108" class="java.lang.String" itemvalue="nvidia-ml-py" />
+            <item index="109" class="java.lang.String" itemvalue="antlr4-python3-runtime" />
+            <item index="110" class="java.lang.String" itemvalue="wcwidth" />
+            <item index="111" class="java.lang.String" itemvalue="py-cpuinfo" />
+            <item index="112" class="java.lang.String" itemvalue="llvmlite" />
+            <item index="113" class="java.lang.String" itemvalue="nvidia-cublas-cu11" />
+            <item index="114" class="java.lang.String" itemvalue="nvidia-cusparse-cu11" />
+            <item index="115" class="java.lang.String" itemvalue="nvidia-nccl-cu11" />
+            <item index="116" class="java.lang.String" itemvalue="nvidia-cudnn-cu11" />
+            <item index="117" class="java.lang.String" itemvalue="gxl-ai-utils" />
+            <item index="118" class="java.lang.String" itemvalue="threadpoolct" />
+            <item index="119" class="java.lang.String" itemvalue="Cython" />
+            <item index="120" class="java.lang.String" itemvalue="lazy_loader" />
+            <item index="121" class="java.lang.String" itemvalue="soxr" />
+            <item index="122" class="java.lang.String" itemvalue="wheel" />
+            <item index="123" class="java.lang.String" itemvalue="gpustat" />
+            <item index="124" class="java.lang.String" itemvalue="s3prl" />
+            <item index="125" class="java.lang.String" itemvalue="blessed" />
+            <item index="126" class="java.lang.String" itemvalue="ffmpeg-python" />
+            <item index="127" class="java.lang.String" itemvalue="pooch" />
+            <item index="128" class="java.lang.String" itemvalue="hydra-core" />
+            <item index="129" class="java.lang.String" itemvalue="future" />
+            <item index="130" class="java.lang.String" itemvalue="cmake" />
+            <item index="131" class="java.lang.String" itemvalue="typing_extensions" />
+            <item index="132" class="java.lang.String" itemvalue="imblearn" />
+            <item index="133" class="java.lang.String" itemvalue="imbalanced-learn" />
+            <item index="134" class="java.lang.String" itemvalue="onnxconverter-common" />
+            <item index="135" class="java.lang.String" itemvalue="panda" />
+            <item index="136" class="java.lang.String" itemvalue="emoji" />
+            <item index="137" class="java.lang.String" itemvalue="nltk" />
+            <item index="138" class="java.lang.String" itemvalue="nvidia-cuda-cupti-cu12" />
+            <item index="139" class="java.lang.String" itemvalue="PyYAML" />
+            <item index="140" class="java.lang.String" itemvalue="nvidia-cufft-cu12" />
+            <item index="141" class="java.lang.String" itemvalue="cycler" />
+            <item index="142" class="java.lang.String" itemvalue="httptools" />
+            <item index="143" class="java.lang.String" itemvalue="frozenlist" />
+            <item index="144" class="java.lang.String" itemvalue="nvidia-cusolver-cu12" />
+            <item index="145" class="java.lang.String" itemvalue="nvidia-curand-cu12" />
+            <item index="146" class="java.lang.String" itemvalue="Pygments" />
+            <item index="147" class="java.lang.String" itemvalue="aliyun-python-sdk-core" />
+            <item index="148" class="java.lang.String" itemvalue="anyio" />
+            <item index="149" class="java.lang.String" itemvalue="multiprocess" />
+            <item index="150" class="java.lang.String" itemvalue="nvidia-cuda-runtime-cu12" />
+            <item index="151" class="java.lang.String" itemvalue="pyparsing" />
+            <item index="152" class="java.lang.String" itemvalue="vocos" />
+            <item index="153" class="java.lang.String" itemvalue="Markdown" />
+            <item index="154" class="java.lang.String" itemvalue="uvloop" />
+            <item index="155" class="java.lang.String" itemvalue="xxhash" />
+            <item index="156" class="java.lang.String" itemvalue="nvidia-cuda-nvrtc-cu12" />
+            <item index="157" class="java.lang.String" itemvalue="kaldifeat" />
+            <item index="158" class="java.lang.String" itemvalue="Werkzeug" />
+            <item index="159" class="java.lang.String" itemvalue="k2" />
+            <item index="160" class="java.lang.String" itemvalue="aiohappyeyeballs" />
+            <item index="161" class="java.lang.String" itemvalue="cryptography" />
+            <item index="162" class="java.lang.String" itemvalue="kiwisolver" />
+            <item index="163" class="java.lang.String" itemvalue="attrs" />
+            <item index="164" class="java.lang.String" itemvalue="coloredlogs" />
+            <item index="165" class="java.lang.String" itemvalue="contourpy" />
+            <item index="166" class="java.lang.String" itemvalue="simplejson" />
+            <item index="167" class="java.lang.String" itemvalue="fonttools" />
+            <item index="168" class="java.lang.String" itemvalue="flatbuffers" />
+            <item index="169" class="java.lang.String" itemvalue="nvidia-cublas-cu12" />
+            <item index="170" class="java.lang.String" itemvalue="nvidia-nvtx-cu12" />
+            <item index="171" class="java.lang.String" itemvalue="cytoolz" />
+            <item index="172" class="java.lang.String" itemvalue="propcache" />
+            <item index="173" class="java.lang.String" itemvalue="oss2" />
+            <item index="174" class="java.lang.String" itemvalue="vector-quantize-pytorch" />
+            <item index="175" class="java.lang.String" itemvalue="encodec" />
+            <item index="176" class="java.lang.String" itemvalue="einx" />
+            <item index="177" class="java.lang.String" itemvalue="nvidia-nvjitlink-cu12" />
+            <item index="178" class="java.lang.String" itemvalue="cffi" />
+            <item index="179" class="java.lang.String" itemvalue="nvidia-cusparse-cu12" />
+            <item index="180" class="java.lang.String" itemvalue="crcmod" />
+            <item index="181" class="java.lang.String" itemvalue="pybase16384" />
+            <item index="182" class="java.lang.String" itemvalue="bitarray" />
+            <item index="183" class="java.lang.String" itemvalue="nvidia-nccl-cu12" />
+            <item index="184" class="java.lang.String" itemvalue="sniffio" />
+            <item index="185" class="java.lang.String" itemvalue="urllib3" />
+            <item index="186" class="java.lang.String" itemvalue="flashlight" />
+            <item index="187" class="java.lang.String" itemvalue="pyarrow" />
+            <item index="188" class="java.lang.String" itemvalue="rich" />
+            <item index="189" class="java.lang.String" itemvalue="nvidia-cudnn-cu12" />
+            <item index="190" class="java.lang.String" itemvalue="addict" />
+            <item index="191" class="java.lang.String" itemvalue="gxl_ai_utils" />
+            <item index="192" class="java.lang.String" itemvalue="jmespath" />
+            <item index="193" class="java.lang.String" itemvalue="toolz" />
+            <item index="194" class="java.lang.String" itemvalue="aliyun-python-sdk-kms" />
+            <item index="195" class="java.lang.String" itemvalue="sanic" />
+            <item index="196" class="java.lang.String" itemvalue="aiohttp" />
+            <item index="197" class="java.lang.String" itemvalue="multidict" />
+            <item index="198" class="java.lang.String" itemvalue="grpcio" />
+            <item index="199" class="java.lang.String" itemvalue="yarl" />
+            <item index="200" class="java.lang.String" itemvalue="frozendict" />
+            <item index="201" class="java.lang.String" itemvalue="pycryptodome" />
+            <item index="202" class="java.lang.String" itemvalue="pytz" />
+            <item index="203" class="java.lang.String" itemvalue="aiosignal" />
+            <item index="204" class="java.lang.String" itemvalue="ujson" />
+            <item index="205" class="java.lang.String" itemvalue="cloudpickle" />
+            <item index="206" class="java.lang.String" itemvalue="ml-dtypes" />
+          </list>
+        </value>
+      </option>
+    </inspection_tool>
+  </profile>
+</component>

.idea/inspectionProfiles/profiles_settings.xml

@@ -0,0 +1,6 @@
+<component name="InspectionProjectProfileManager">
+  <settings>
+    <option name="USE_PROJECT_PROFILE" value="false" />
+    <version value="1.0" />
+  </settings>
+</component>

.idea/misc.xml

@@ -0,0 +1,7 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="Black">
+    <option name="sdkName" value="Python 3.12 (OSUM)" />
+  </component>
+  <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.12 (OSUM)" project-jdk-type="Python SDK" />
+</project>

.idea/modules.xml

@@ -0,0 +1,8 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectModuleManager">
+    <modules>
+      <module fileurl="file://$PROJECT_DIR$/.idea/OSUM.iml" filepath="$PROJECT_DIR$/.idea/OSUM.iml" />
+    </modules>
+  </component>
+</project>

.idea/vcs.xml

@@ -0,0 +1,6 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="VcsDirectoryMappings">
+    <mapping directory="" vcs="Git" />
+  </component>
+</project>

app.py

@@ -1,64 +1,257 @@
+import base64
+import json
+import time
+
 import gradio as gr
-from huggingface_hub import InferenceClient
+import os
 
+import sys
+
+
+sys.path.insert(0, '../../../../')
+# from gxl_ai_utils.utils import utils_file
+# from wenet.utils.init_tokenizer import init_tokenizer
+# from gxl_ai_utils.config.gxl_config import GxlNode
+# from wenet.utils.init_model import init_model
+import logging
+# import librosa
+# import torch
+# import torchaudio
+# import numpy as np
+
+# 将图片转换为 Base64
+with open("./实验室.png", "rb") as image_file:
+    encoded_string = base64.b64encode(image_file.read()).decode("utf-8")
+
+# with open("./cat.jpg", "rb") as image_file:
+#     encoded_string = base64.b64encode(image_file.read()).decode("utf-8")
+
+# 自定义CSS样式
+custom_css = """
+/* 自定义CSS样式 */
 """
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
+# 任务提示映射
+TASK_PROMPT_MAPPING = {
+    "ASR (Automatic Speech Recognition)": "执行语音识别任务，将音频转换为文字。",
+    "SRWT (Speech Recognition with Timestamps)": "请转录音频内容，并为每个英文词汇及其对应的中文翻译标注出精确到0.1秒的起止时间，时间范围用<>括起来。",
+    "VED (Vocal Event Detection)(类别:laugh，cough，cry，screaming，sigh，throat clearing，sneeze，other)": "请将音频转录为文字记录，并在记录末尾标注<音频事件>标签，音频事件共8种：laugh，cough，cry，screaming，sigh，throat clearing，sneeze，other。",
+    "SER (Speech Emotion Recognition)(类别:sad，anger，neutral，happy，surprise，fear，disgust，和other)": "请将音频内容转录成文字记录，并在记录末尾标注<情感>标签，情感共8种：sad，anger，neutral，happy，surprise，fear，disgust，和other。",
+    "SSR (Speaking Style Recognition)(类别:新闻科普，恐怖故事，童话故事，客服，诗歌散文，有声书，日常口语，其他)": "请将音频内容进行文字转录，并在最后添加<风格>标签，标签共8种：新闻科普、恐怖故事、童话故事、客服、诗歌散文、有声书、日常口语、其他。",
+    "SGC (Speaker Gender Classification)(类别:female,male)": "请将音频转录为文本，并在文本结尾处标注<性别>标签，性别为female或male。",
+    "SAP (Speaker Age Prediction)(类别:child、adult和old)": "请将音频转录为文本，并在文本结尾处标注<年龄>标签，年龄划分为child、adult和old三种。",
+    "STTC (Speech to Text Chat)": "首先将语音转录为文字，然后对语音内容进行回复，转录和文字之间使用<开始回答>分割。"
+}
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+gpu_id = 4
+# def init_model_my():
+#     logging.basicConfig(level=logging.DEBUG,
+#                         format='%(asctime)s %(levelname)s %(message)s')
+#     config_path = "/home/node54_tmpdata/xlgeng/code/wenet_undersdand_and_speech_xlgeng/examples/wenetspeech/whisper/exp/update_data/epoch_1_with_token/epoch_11.yaml"
+#     #config_path = "/home/work_nfs15/asr_data/ckpt/understanding_model/step_24999.yaml"
+#
+#     checkpoint_path = "/home/node54_tmpdata/xlgeng/code/wenet_undersdand_and_speech_xlgeng/examples/wenetspeech/whisper/exp/update_data/epoch_1_with_token/epoch_11.pt"
+#     checkpoint_path = "/home/work_nfs15/asr_data/ckpt/understanding_model/epoch4/step_21249.pt"
+#     checkpoint_path = "/home/work_nfs15/asr_data/ckpt/understanding_model/epoch_13_with_asr-chat_full_data/step_32499/step_32499.pt"
+#     args = GxlNode({
+#         "checkpoint": checkpoint_path,
+#     })
+#     configs = utils_file.load_dict_from_yaml(config_path)
+#     model, configs = init_model(args, configs)
+#     model = model.cuda(gpu_id)
+#     tokenizer = init_tokenizer(configs)
+#     print(model)
+#     return model, tokenizer
+#
+# model, tokenizer = init_model_my()
+#
+# def do_resample(input_wav_path, output_wav_path):
+#     """"""
+#     print(f'input_wav_path: {input_wav_path}, output_wav_path: {output_wav_path}')
+#     waveform, sample_rate = torchaudio.load(input_wav_path)
+#     # 检查音频的维度
+#     num_channels = waveform.shape[0]
+#     # 如果音频是多通道的，则进行通道平均
+#     if num_channels > 1:
+#         waveform = torch.mean(waveform, dim=0, keepdim=True)
+#     waveform = torchaudio.transforms.Resample(
+#         orig_freq=sample_rate, new_freq=16000)(waveform)
+#     utils_file.makedir_for_file(output_wav_path)
+#     torchaudio.save(output_wav_path, waveform, 16000)
+#
+# def true_decode_fuc(input_wav_path, input_prompt):
+#     # input_prompt = TASK_PROMPT_MAPPING.get(input_prompt, "未知任务类型")
+#     print(f"wav_path: {input_wav_path}, prompt:{input_prompt}")
+#     timestamp_ms = int(time.time() * 1000)
+#     now_file_tmp_path_resample = f'/home/xlgeng/.cache/.temp/{timestamp_ms}_resample.wav'
+#     do_resample(input_wav_path, now_file_tmp_path_resample)
+#     # tmp_vad_path = f'/home/xlgeng/.cache/.temp/{timestamp_ms}_vad.wav'
+#     # remove_silence_torchaudio_ends(now_file_tmp_path_resample, tmp_vad_path)
+#     # input_wav_path  = tmp_vad_path
+#     input_wav_path = now_file_tmp_path_resample
+#     waveform, sample_rate = torchaudio.load(input_wav_path)
+#     waveform = waveform.squeeze(0)  # (channel=1, sample) -> (sample,)
+#     print(f'wavform shape: {waveform.shape}, sample_rate: {sample_rate}')
+#     window = torch.hann_window(400)
+#     stft = torch.stft(waveform,
+#                       400,
+#                       160,
+#                       window=window,
+#                       return_complex=True)
+#     magnitudes = stft[..., :-1].abs() ** 2
+#
+#     filters = torch.from_numpy(
+#         librosa.filters.mel(sr=sample_rate,
+#                             n_fft=400,
+#                             n_mels=80))
+#     mel_spec = filters @ magnitudes
+#
+#     # NOTE(xcsong): https://github.com/openai/whisper/discussions/269
+#     log_spec = torch.clamp(mel_spec, min=1e-10).log10()
+#     log_spec = torch.maximum(log_spec, log_spec.max() - 8.0)
+#     log_spec = (log_spec + 4.0) / 4.0
+#     feat = log_spec.transpose(0, 1)
+#     feat_lens = torch.tensor([feat.shape[0]], dtype=torch.int64).to(gpu_id)
+#     feat = feat.unsqueeze(0).to(gpu_id)
+#     # feat = feat.half()
+#     # feat_lens = feat_lens.half()
+#     res_text = model.generate(wavs=feat, wavs_len=feat_lens, prompt=input_prompt)[0]
+#     print("耿雪龙哈哈：", res_text)
+#     return res_text, now_file_tmp_path_resample
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+def do_decode(input_wav_path, input_prompt):
+    print(f'input_wav_path= {input_wav_path}, input_prompt= {input_prompt}')
+    # 省略处理逻辑
+    # output_res, now_file_tmp_path_resample= true_decode_fuc(input_wav_path, input_prompt)
+    output_res = f"耿雪龙哈哈：测试结果, input_wav_path= {input_wav_path}, input_prompt= {input_prompt}"
+    return output_res
 
-    messages.append({"role": "user", "content": message})
+def save_to_jsonl(if_correct, wav, prompt, res):
+    data = {
+        "if_correct": if_correct,
+        "wav": wav,
+        "task": prompt,
+        "res": res
+    }
+    with open("results.jsonl", "a", encoding="utf-8") as f:
+        f.write(json.dumps(data, ensure_ascii=False) + "\n")
 
-    response = ""
+def handle_submit(input_wav_path, input_prompt):
+    output_res = do_decode(input_wav_path, input_prompt)
+    return output_res
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+def download_audio(input_wav_path):
+    if input_wav_path:
+        # 返回文件路径供下载
+        return input_wav_path
+    else:
+        return None
 
-        response += token
-        yield response
+# 创建Gradio界面
+with gr.Blocks(css=custom_css) as demo:
+    # 添加标题
+    gr.Markdown(
+        f"""
+        <div style="display: flex; align-items: center; justify-content: center; text-align: center;">
+            <h1 style="font-family: 'Arial', sans-serif; color: #014377; font-size: 32px; margin-bottom: 0; display: inline-block; vertical-align: middle;">
+                OSUM Speech Understanding Model Test
+            </h1>
+        </div>
+        """
+    )
 
+    # 添加音频输入和任务选择
+    with gr.Row():
+        with gr.Column(scale=1):
+            audio_input = gr.Audio(label="录音", type="filepath")
+        with gr.Column(scale=1, min_width=300):  # 给输出框设置最小宽度，确保等高对齐
+            output_text = gr.Textbox(label="输出结果", lines=8, placeholder="生成的结果将显示在这里...", interactive=False)
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
-)
-
-
-if __name__ == "__main__":
-    demo.launch()
+    # 添加任务选择和自定义输入框
+    with gr.Row():
+        task_dropdown = gr.Dropdown(
+            label="任务",
+            choices=list(TASK_PROMPT_MAPPING.keys()) + ["自主输入文本"],  # 新增选项
+            value="ASR (Automatic Speech Recognition)"
+        )
+        custom_prompt_input = gr.Textbox(label="自定义任务提示", placeholder="请输入自定义任务提示...", visible=False)  # 新增文本输入框
+
+    # 添加按钮（下载按钮在左边，开始处理按钮在右边）
+    with gr.Row():
+        download_button = gr.DownloadButton("下载音频", variant="secondary", elem_classes=["button-height", "download-button"])
+        submit_button = gr.Button("开始处理", variant="primary", elem_classes=["button-height", "submit-button"])
+
+    # 添加确认组件
+    with gr.Row(visible=False) as confirmation_row:
+        gr.Markdown("请判断结果是否正确：")
+        confirmation_buttons = gr.Radio(
+            choices=["正确", "错误"],
+            label="",
+            interactive=True,
+            container=False,
+            elem_classes="confirmation-buttons"
+        )
+        save_button = gr.Button("提交反馈", variant="secondary")
+
+    # 添加底部内容
+    with gr.Row():
+        # 底部内容容器
+        with gr.Column(scale=1, min_width=800):  # 设置最小宽度以确保内容居中
+            gr.Markdown(
+                f"""
+                <div style="position: fixed; bottom: 20px; left: 50%; transform: translateX(-50%); display: flex; align-items: center; justify-content: center; gap: 20px;">
+                    <div style="text-align: center;">
+                        <p style="margin: 0;"><strong>Audio, Speech and Language Processing Group (ASLP@NPU),</strong></p>
+                        <p style="margin: 0;"><strong>Northwestern Polytechnical University</strong></p>
+                    </div>
+                    <img src="data:image/png;base64,{encoded_string}" alt="OSUM Logo" style="height: 80px; width: auto;">
+                </div>
+                """
+            )
+
+    # 绑定事件
+    def show_confirmation(output_res, input_wav_path, input_prompt):
+        return gr.update(visible=True), output_res, input_wav_path, input_prompt
+
+    def save_result(if_correct, wav, prompt, res):
+        save_to_jsonl(if_correct, wav, prompt, res)
+        return gr.update(visible=False)
+
+    def handle_submit(input_wav_path, task_choice, custom_prompt):
+        if task_choice == "自主输入文本":
+            input_prompt = custom_prompt  # 使用用户输入的自定义文本
+        else:
+            input_prompt = TASK_PROMPT_MAPPING.get(task_choice, "未知任务类型")  # 使用预定义的提示
+        output_res = do_decode(input_wav_path, input_prompt)
+        return output_res
+
+    task_dropdown.change(
+        fn=lambda choice: gr.update(visible=choice == "自主输入文本"),
+        inputs=task_dropdown,
+        outputs=custom_prompt_input
+    )
+
+    submit_button.click(
+        fn=handle_submit,
+        inputs=[audio_input, task_dropdown, custom_prompt_input],
+        outputs=output_text
+    ).then(
+        fn=show_confirmation,
+        inputs=[output_text, audio_input, task_dropdown],
+        outputs=[confirmation_row, output_text, audio_input, task_dropdown]
+    )
+
+    download_button.click(
+        fn=download_audio,
+        inputs=[audio_input],
+        outputs=[download_button]  # 输出到 download_button
+    )
+
+    save_button.click(
+        fn=save_result,
+        inputs=[confirmation_buttons, audio_input, task_dropdown, output_text],
+        outputs=confirmation_row
+    )
+
+if __name__== "__main__":
+    demo.launch()

app_old.py

@@ -0,0 +1,64 @@
+import gradio as gr
+from huggingface_hub import InferenceClient
+
+"""
+For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
+"""
+client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+
+
+def respond(
+    message,
+    history: list[tuple[str, str]],
+    system_message,
+    max_tokens,
+    temperature,
+    top_p,
+):
+    messages = [{"role": "system", "content": system_message}]
+
+    for val in history:
+        if val[0]:
+            messages.append({"role": "user", "content": val[0]})
+        if val[1]:
+            messages.append({"role": "assistant", "content": val[1]})
+
+    messages.append({"role": "user", "content": message})
+
+    response = ""
+
+    for message in client.chat_completion(
+        messages,
+        max_tokens=max_tokens,
+        stream=True,
+        temperature=temperature,
+        top_p=top_p,
+    ):
+        token = message.choices[0].delta.content
+
+        response += token
+        yield response
+
+
+"""
+For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
+"""
+demo = gr.ChatInterface(
+    respond,
+    additional_inputs=[
+        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
+        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
+        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
+        gr.Slider(
+            minimum=0.1,
+            maximum=1.0,
+            value=0.95,
+            step=0.05,
+            label="Top-p (nucleus sampling)",
+        ),
+    ],
+)
+
+
+if __name__ == "__main__":
+    demo.launch()