rawnet2 code fix

.gitignore

@@ -1,5 +1,6 @@
 # Ignore weights
 checkpoints/RAWNET_ASVSPOOF_FOR_INTHEWILD_PURDUE.pth
+debug.sh
 
 .idea
 

.idea/deployment.xml

@@ -1,6 +1,6 @@
 <?xml version="1.0" encoding="UTF-8"?>
 <project version="4">
-  <component name="PublishConfigData">
+  <component name="PublishConfigData" autoUpload="On explicit save action" autoUploadExternalChanges="true">
     <serverData>
       <paths name="Fermi">
         <serverdata>
@@ -10,5 +10,6 @@
         </serverdata>
       </paths>
     </serverData>
+    <option name="myAutoUpload" value="ON_EXPLICIT_SAVE" />
   </component>
 </project>

.idea/vcs.xml

@@ -2,5 +2,6 @@
 <project version="4">
   <component name="VcsDirectoryMappings">
     <mapping directory="" vcs="Git" />
+    <mapping directory="$PROJECT_DIR$/safe-challenge-practice-dataset" vcs="Git" />
   </component>
 </project>

.idea/workspace.xml

@@ -5,7 +5,19 @@
   </component>
   <component name="ChangeListManager">
     <list default="true" id="23565123-73ab-4f40-a9ef-1086e0c9e1ec" name="Changes" comment="">
+      <change afterPath="$PROJECT_DIR$/Dockerfile" afterDir="false" />
+      <change afterPath="$PROJECT_DIR$/data/06704fa4-5a0c-540c-86e6-c98af1528478.wav" afterDir="false" />
+      <change afterPath="$PROJECT_DIR$/data/07bd0843-74a6-53ec-a3f0-00dfc31d6e2a.wav" afterDir="false" />
+      <change afterPath="$PROJECT_DIR$/data/9712245a-548d-584c-a82d-a543f1ea21ac.wav" afterDir="false" />
+      <change afterPath="$PROJECT_DIR$/data/c3e008aa-e4ba-5d2a-b37e-dd6d0ae640cb.wav" afterDir="false" />
+      <change afterPath="$PROJECT_DIR$/debug.sh" afterDir="false" />
+      <change afterPath="$PROJECT_DIR$/requirements.txt" afterDir="false" />
+      <change beforePath="$PROJECT_DIR$/.gitignore" beforeDir="false" afterPath="$PROJECT_DIR$/.gitignore" afterDir="false" />
+      <change beforePath="$PROJECT_DIR$/.idea/deployment.xml" beforeDir="false" afterPath="$PROJECT_DIR$/.idea/deployment.xml" afterDir="false" />
+      <change beforePath="$PROJECT_DIR$/.idea/vcs.xml" beforeDir="false" afterPath="$PROJECT_DIR$/.idea/vcs.xml" afterDir="false" />
       <change beforePath="$PROJECT_DIR$/.idea/workspace.xml" beforeDir="false" afterPath="$PROJECT_DIR$/.idea/workspace.xml" afterDir="false" />
+      <change beforePath="$PROJECT_DIR$/checkpoints/RAWNET_ASVSPOOF_FOR_INTHEWILD_PURDUE.pth" beforeDir="false" />
+      <change beforePath="$PROJECT_DIR$/preprocess.py" beforeDir="false" afterPath="$PROJECT_DIR$/preprocess.py" afterDir="false" />
       <change beforePath="$PROJECT_DIR$/script.py" beforeDir="false" afterPath="$PROJECT_DIR$/script.py" afterDir="false" />
     </list>
     <option name="SHOW_DIALOG" value="false" />
@@ -38,21 +50,50 @@
     <option name="hideEmptyMiddlePackages" value="true" />
     <option name="showLibraryContents" value="true" />
   </component>
-  <component name="PropertiesComponent">{
-  &quot;keyToString&quot;: {
-    &quot;RunOnceActivity.OpenProjectViewOnStart&quot;: &quot;true&quot;,
-    &quot;RunOnceActivity.ShowReadmeOnStart&quot;: &quot;true&quot;,
-    &quot;WebServerToolWindowFactoryState&quot;: &quot;true&quot;,
-    &quot;last_opened_file_path&quot;: &quot;/Users/davidesalvi/Desktop/PHD/projects/polimi-ispl/ispl_safe&quot;,
-    &quot;node.js.detected.package.eslint&quot;: &quot;true&quot;,
-    &quot;node.js.detected.package.tslint&quot;: &quot;true&quot;,
-    &quot;node.js.selected.package.eslint&quot;: &quot;(autodetect)&quot;,
-    &quot;node.js.selected.package.tslint&quot;: &quot;(autodetect)&quot;,
-    &quot;nodejs_package_manager_path&quot;: &quot;npm&quot;,
-    &quot;settings.editor.selected.configurable&quot;: &quot;web.server&quot;,
-    &quot;vue.rearranger.settings.migration&quot;: &quot;true&quot;
+  <component name="PropertiesComponent"><![CDATA[{
+  "keyToString": {
+    "RunOnceActivity.OpenProjectViewOnStart": "true",
+    "RunOnceActivity.ShowReadmeOnStart": "true",
+    "WebServerToolWindowFactoryState": "true",
+    "last_opened_file_path": "/Users/davidesalvi/Desktop/PHD/projects/polimi-ispl/ispl_safe",
+    "node.js.detected.package.eslint": "true",
+    "node.js.detected.package.tslint": "true",
+    "node.js.selected.package.eslint": "(autodetect)",
+    "node.js.selected.package.tslint": "(autodetect)",
+    "nodejs_package_manager_path": "npm",
+    "settings.editor.selected.configurable": "com.jetbrains.python.configuration.PyActiveSdkModuleConfigurable",
+    "vue.rearranger.settings.migration": "true"
   }
-}</component>
+}]]></component>
+  <component name="RunManager">
+    <configuration name="script" type="PythonConfigurationType" factoryName="Python" temporary="true" nameIsGenerated="true">
+      <module name="ispl_safe" />
+      <option name="INTERPRETER_OPTIONS" value="" />
+      <option name="PARENT_ENVS" value="true" />
+      <envs>
+        <env name="PYTHONUNBUFFERED" value="1" />
+      </envs>
+      <option name="SDK_HOME" value="" />
+      <option name="WORKING_DIRECTORY" value="$PROJECT_DIR$" />
+      <option name="IS_MODULE_SDK" value="true" />
+      <option name="ADD_CONTENT_ROOTS" value="true" />
+      <option name="ADD_SOURCE_ROOTS" value="true" />
+      <EXTENSION ID="PythonCoverageRunConfigurationExtension" runner="coverage.py" />
+      <option name="SCRIPT_NAME" value="$PROJECT_DIR$/script.py" />
+      <option name="PARAMETERS" value="" />
+      <option name="SHOW_COMMAND_LINE" value="false" />
+      <option name="EMULATE_TERMINAL" value="false" />
+      <option name="MODULE_MODE" value="false" />
+      <option name="REDIRECT_INPUT" value="false" />
+      <option name="INPUT_FILE" value="" />
+      <method v="2" />
+    </configuration>
+    <recent_temporary>
+      <list>
+        <item itemvalue="Python.script" />
+      </list>
+    </recent_temporary>
+  </component>
   <component name="SpellCheckerSettings" RuntimeDictionaries="0" Folders="0" CustomDictionaries="0" DefaultDictionary="application-level" UseSingleDictionary="true" transferred="true" />
   <component name="TaskManager">
     <task active="true" id="Default" summary="Default task">
@@ -62,11 +103,26 @@
       <option name="presentableId" value="Default" />
       <updated>1742573353560</updated>
       <workItem from="1742573355153" duration="587000" />
-      <workItem from="1742806974298" duration="770000" />
+      <workItem from="1742806974298" duration="2741000" />
+      <workItem from="1742810431420" duration="7010000" />
     </task>
     <servers />
   </component>
   <component name="TypeScriptGeneratedFilesManager">
     <option name="version" value="3" />
   </component>
+  <component name="XDebuggerManager">
+    <breakpoint-manager>
+      <breakpoints>
+        <line-breakpoint enabled="true" suspend="THREAD" type="python-line">
+          <url>file://$PROJECT_DIR$/script.py</url>
+          <line>30</line>
+          <option name="timeStamp" value="1" />
+        </line-breakpoint>
+      </breakpoints>
+    </breakpoint-manager>
+  </component>
+  <component name="com.intellij.coverage.CoverageDataManagerImpl">
+    <SUITE FILE_PATH="coverage/ispl_safe$script.coverage" NAME="script Coverage Results" MODIFIED="1742814252166" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$" />
+  </component>
 </project>

Dockerfile

@@ -0,0 +1,6 @@
+FROM huggingface/competitions:latest
+WORKDIR /tmp
+COPY requirements.txt .
+RUN pip install --progress-bar off -r requirements.txt
+COPY . .
+CMD bash debug.sh

checkpoints/RAWNET_ASVSPOOF_FOR_INTHEWILD_PURDUE.pth

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:abd0803e944f98176c98a8b3d53f7ca0c5623cd4cf0566aba7efb0a09a8bc2f7
-size 70534111

preprocess.py

@@ -1,10 +1,41 @@
 
 import librosa
+import numpy as np
 import torch
 
+def pad_audio(x, max_len=48000):
+    x_len = x.shape[0]
+    if x_len >= max_len:
+        return x[:max_len]
+    # need to pad
+    num_repeats = int(max_len / x_len) + 1
+    padded_x = np.tile(x, (1, num_repeats))[:, :max_len][0]
+    return padded_x
+
+
 def preprocess(audio_file):
+    print(f'Preprocessing {audio_file}')
+    # Load the audio file
+    y, sr = librosa.load(audio_file, sr=16000)
+
+    # Evaluate N windows of the audio file
+    num_eval = 5
+    win_len = int(3*sr)
+    start_sample_list = np.linspace(0, len(y) - win_len, num=num_eval)
+
+    frames = []
+    for start_sample in start_sample_list:
+        frames += [y[int(start_sample):int(start_sample) + win_len]]
+    y_win = np.stack(frames, axis=0)
+    tensor = torch.from_numpy(y_win)
+
+    print(f'preprocessed track - shape {tensor.shape}')
+    return tensor
+
+
+def preprocess_old(audio_file):
     # Load the audio file
     y, sr = librosa.load(audio_file, sr=None)
     mfccs = librosa.feature.mfcc(y=y, sr=sr, n_mfcc=13)
-    tensor  = torch.from_numpy(mfccs)[None]
+    tensor = torch.from_numpy(mfccs)[None]
     return tensor

requirements.txt

@@ -0,0 +1,14 @@
+--extra-index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/onnxruntime-cuda-12/pypi/simple/
+onnxruntime-gpu==1.18.0
+av==14.2.0
+tqdm==4.66.4
+torchaudio==2.4.0
+transformers==4.44.1
+h5py==3.13.0
+webrtcvad==2.0.10
+scipy==1.14.1
+torch==2.4.0
+soundfile==0.13.1
+librosa==0.10.2.post1
+omegaconf==2.3.0
+hydra-core==1.3.2

script.py

@@ -2,19 +2,22 @@ import io
 import time
 import os
 
+os.environ['OMP_NUM_THREADS'] = '1'
+os.environ['OPENBLAS_NUM_THREADS'] = '1'
+os.environ['MKL_NUM_THREADS'] = '1'
+os.environ['NUMEXPR_NUM_THREADS'] = '1'
+
 import torch
 import tqdm.auto as tqdm
+from glob import glob
 from datasets import load_dataset
 
 import numpy as np
 import pandas as pd
 
-from models import Model
+# from models import Model
 from preprocess import preprocess
-from pathlib import Path
-
 from src.rawnet_model import RawNet
-# from src.utils import *
 
 
 # Import your model and anything else you want
@@ -25,20 +28,18 @@ from src.rawnet_model import RawNet
 # For testing, this is the docker image that will be used https://github.com/huggingface/competitions/blob/main/Dockerfile
 # It can be pulled here https://hub.docker.com/r/huggingface/competitions/tags
 
-
+print('imported packages')
 # load the dataset. dataset will be automatically downloaded to /tmp/data during evaluation
 DATASET_PATH = "/tmp/data"
-dataset_remote = load_dataset(DATASET_PATH,split = "test",streaming = True)
-
+dataset_remote = glob(os.path.join(DATASET_PATH, '*'))
+# dataset_remote = load_dataset('safe-challenge/safe-challenge-practice-dataset', split="test", streaming=True)
+print('loaded dataset')
 
-# load your model
-device = "cuda:0"
 
-# # DEFINE RAWNET2 MODEL
-this_folder = Path(__file__).parent
-# config_path = this_folder / 'config' / 'rawnet_config.yaml'
-# config = read_yaml(config_path)
+# device = "cuda:0"
+device = "cpu"
 
+# DEFINE RAWNET2 MODEL
 config = {
     "first_conv": 1024,
     "in_channels": 1,
@@ -50,19 +51,20 @@ config = {
     "nb_classes": 2
 }
 
-model_rawnet = RawNet(config['model'], device)
+model = RawNet(config, device)
 
-model_path = this_folder / 'checkpoints' / 'RAWNET_ASVSPOOF_FOR_INTHEWILD_PURDUE.pth'
-model_rawnet.load_state_dict(torch.load(model_path, map_location=device))
+model_path = '/tmp/checkpoints/RAWNET_ASVSPOOF_FOR_INTHEWILD_PURDUE.pth'
+model.load_state_dict(torch.load(model_path, map_location=device))
+# model = model.float()
 
-del model_rawnet
+print('Loaded RawNet2 Weights')
 
-# Define old model
-model = Model().to(device)
+# model = Model().to(device) # OLD MODEL
 
 # iterate over the dataset
 out = []
-for el in tqdm.tqdm(dataset_remote):
+# for el in tqdm.tqdm(dataset_remote):
+for el in dataset_remote:
 
     start_time = time.time()
 
@@ -72,27 +74,38 @@ for el in tqdm.tqdm(dataset_remote):
     # el["audio"]["path"] containts the filename. This is just for reference and you cant actually load it
 
     # if you are using libraries that expect a file. You can use BytesIO object
-    try:
-        file_like = io.BytesIO(el["audio"]["bytes"])
-        tensor = preprocess(file_like)
 
-        with torch.no_grad():
-            # soft decision (such as log likelihood score)
-            # positive score correspond to synthetic prediction
-            # negative score correspond to pristine prediction
-            score = model(tensor.to(device)).cpu().item()
+    # try:
+
+    # file_like = io.BytesIO(el["audio"]["bytes"])
+    # tensor = preprocess(file_like)
+    tensor = preprocess(el)
+
+    with torch.no_grad():
+        # soft decision (such as log likelihood score)
+        # positive score correspond to synthetic prediction
+        # negative score correspond to pristine prediction
+        # score = model(tensor.to(device)).cpu().item()
+        score = model(tensor.to(device))[:, 1].cpu()
+        print(f'SCORE OUT: {score}')
+        score = score.mean().item()
+        print(f'SCORE FINAL: {score}')
+
+        # we require a hard decision to be submited. so you need to pick a threshold
+        pred = "generated" if score > model.threshold else "pristine"
 
-            # we require a hard decision to be submited. so you need to pick a threshold
-            pred = "generated" if score > model.threshold else "pristine"
+    # append your prediction
+    # "id" and "pred" are required. "score" will not be used in scoring but we encourage you to include it. We'll use it for analysis of the results
 
-        # append your prediction
-        # "id" and "pred" are required. "score" will not be used in scoring but we encourage you to include it. We'll use it for analysis of the results
+    # out.append(dict(id = el["id"], pred = pred, score = score, time = time.time() - start_time))
+    out.append(dict(id=el, pred=pred, score=score, time=time.time() - start_time))
 
-        out.append(dict(id = el["id"], pred = pred, score = score, time = time.time() - start_time))
-    except Exception as e:
-        print(e)
-        print("failed", el["id"])
-        out.append(dict(id = el["id"], pred = "none", score = None))
+    # except Exception as e:
+    #     print(e)
+    #     # print("failed", el["id"])
+    #     # out.append(dict(id = el["id"], pred = "none", score = None))
+    #     print("failed", el)
+    #     out.append(dict(id = el, pred = "none", score = None))
 
 # save the final result and that's it
-pd.DataFrame(out).to_csv("submission.csv",index = False)
+pd.DataFrame(out).to_csv("submission.csv", index=False)