Huzaifa Pardawala commited on
Commit
77de5bf
Β·
1 Parent(s): cfa0c42

editing framework and datasets by domain

Browse files
Files changed (1) hide show
  1. index.html +215 -214
index.html CHANGED
@@ -1063,240 +1063,241 @@
1063
 
1064
  <section id="framework" class="section content-section">
1065
  <div class="container is-max-desktop">
1066
- <!-- Framework -->
 
1067
  <div class="columns is-centered">
1068
  <div class="column is-full">
1069
  <h2 class="title is-3 section-title has-text-centered">Framework & Resources</h2>
1070
-
1071
- <div class="content">
1072
- <!-- Framework Overview -->
1073
- <div class="framework-overview has-text-centered mb-5">
1074
- <div class="box p-5">
1075
- <span class="icon is-large mb-4">
1076
- <i class="fas fa-fire flame-icon fa-3x"></i>
1077
- </span>
1078
- <h3 class="title is-4"><span class="flame">FLaME</span> Framework</h3>
1079
- <p class="subtitle">Open-source, modular toolkit for financial LLM evaluation</p>
1080
-
1081
- <div class="buttons is-centered mt-4">
1082
- <a href="https://github.com/flame-benchmark/flame" target="_blank" class="button is-primary">
1083
- <span class="icon">
1084
- <i class="fab fa-github"></i>
1085
- </span>
1086
- <span>GitHub Repository</span>
1087
- </a>
1088
- <a href="https://huggingface.co/spaces/flame-benchmark/flame" target="_blank" class="button is-info">
1089
- <span class="icon">
1090
- <i class="fab fa-huggingface"></i>
1091
- </span>
1092
- <span>HuggingFace Space</span>
1093
- </a>
1094
- </div>
1095
- </div>
1096
  </div>
1097
-
1098
- <div class="columns is-multiline">
1099
- <!-- Framework Features -->
1100
- <div class="column is-6">
1101
- <div class="card h-100">
1102
- <div class="card-header">
1103
- <p class="card-header-title">
1104
- <span class="icon mr-2"><i class="fas fa-cogs"></i></span>
1105
- Framework Features
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1106
  </p>
 
1107
  </div>
1108
- <div class="card-content">
1109
- <div class="content">
1110
- <div class="feature-list">
1111
- <div class="feature-item mb-3">
1112
- <p class="has-text-weight-bold mb-1">
1113
- <span class="icon has-text-primary">
1114
- <i class="fas fa-check"></i>
1115
- </span>
1116
- Standardized Pipelines
1117
- </p>
1118
- <p class="is-size-7 ml-4">Evaluation pipelines for all major financial NLP tasks</p>
1119
- </div>
1120
-
1121
- <div class="feature-item mb-3">
1122
- <p class="has-text-weight-bold mb-1">
1123
- <span class="icon has-text-primary">
1124
- <i class="fas fa-check"></i>
1125
- </span>
1126
- Customizable Assessment
1127
- </p>
1128
- <p class="is-size-7 ml-4">Configure evaluations across models and datasets</p>
1129
- </div>
1130
-
1131
- <div class="feature-item mb-3">
1132
- <p class="has-text-weight-bold mb-1">
1133
- <span class="icon has-text-primary">
1134
- <i class="fas fa-check"></i>
1135
- </span>
1136
- Reproducible Benchmarking
1137
- </p>
1138
- <p class="is-size-7 ml-4">Consistent metrics and evaluation protocols</p>
1139
- </div>
1140
-
1141
- <div class="feature-item">
1142
- <p class="has-text-weight-bold mb-1">
1143
- <span class="icon has-text-primary">
1144
- <i class="fas fa-check"></i>
1145
- </span>
1146
- Extensible Architecture
1147
- </p>
1148
- <p class="is-size-7 ml-4">Easily add new tasks, models, and datasets</p>
1149
- </div>
1150
- </div>
1151
- </div>
1152
  </div>
1153
- </div>
1154
- </div>
1155
-
1156
- <!-- Datasets -->
1157
- <div class="column is-6">
1158
- <div class="card h-100">
1159
- <div class="card-header">
1160
- <p class="card-header-title">
1161
- <span class="icon mr-2"><i class="fas fa-database"></i></span>
1162
- Datasets by Domain
1163
  </p>
 
1164
  </div>
1165
- <div class="card-content">
1166
- <div class="content">
1167
- <div class="columns is-multiline">
1168
- <div class="column is-6">
1169
- <div class="dataset-category">
1170
- <p class="has-text-weight-bold">
1171
- <span class="icon has-text-primary">
1172
- <i class="fas fa-university"></i>
1173
- </span>
1174
- Banking
1175
- </p>
1176
- <ul style="list-style-type: none;">
1177
- <li>Banking77</li>
1178
- <li>FiQA</li>
1179
- <li>FinRED</li>
1180
- </ul>
1181
- </div>
1182
- </div>
1183
-
1184
- <div class="column is-6">
1185
- <div class="dataset-category">
1186
- <p class="has-text-weight-bold">
1187
- <span class="icon has-text-primary">
1188
- <i class="fas fa-chart-line"></i>
1189
- </span>
1190
- Investment
1191
- </p>
1192
- <ul style="list-style-type: none;">
1193
- <li>FPB</li>
1194
- <li>Headlines</li>
1195
- <li>SubjectiveQA</li>
1196
- </ul>
1197
- </div>
1198
- </div>
1199
-
1200
- <div class="column is-6">
1201
- <div class="dataset-category">
1202
- <p class="has-text-weight-bold">
1203
- <span class="icon has-text-primary">
1204
- <i class="fas fa-calculator"></i>
1205
- </span>
1206
- Accounting
1207
- </p>
1208
- <ul style="list-style-type: none;">
1209
- <li>FinQA</li>
1210
- <li>TaT-QA</li>
1211
- <li>ConvFinQA</li>
1212
- </ul>
1213
- </div>
1214
- </div>
1215
-
1216
- <div class="column is-6">
1217
- <div class="dataset-category">
1218
- <p class="has-text-weight-bold">
1219
- <span class="icon has-text-primary">
1220
- <i class="fas fa-building"></i>
1221
- </span>
1222
- Corporate
1223
- </p>
1224
- <ul style="list-style-type: none;">
1225
- <li>ECTSum</li>
1226
- <li>EDTSum</li>
1227
- <li>FinCausal</li>
1228
- </ul>
1229
- </div>
1230
- </div>
1231
- </div>
1232
- </div>
1233
  </div>
 
1234
  </div>
1235
  </div>
1236
-
1237
- <!-- Getting Started -->
1238
- <div class="column is-full mt-4">
1239
- <div class="card">
1240
- <div class="card-header">
1241
- <p class="card-header-title">
1242
- <span class="icon mr-2"><i class="fas fa-rocket"></i></span>
1243
- Getting Started
1244
- </p>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1245
  </div>
1246
- <div class="card-content">
1247
- <div class="content">
1248
- <div class="steps">
1249
- <div class="step-item">
1250
- <div class="step-marker">1</div>
1251
- <div class="step-details">
1252
- <p class="step-title">Clone Repository</p>
1253
- <p class="is-size-7 has-text-grey">
1254
- <code>git clone https://github.com/flame-benchmark/flame.git</code>
1255
- </p>
1256
- </div>
1257
- </div>
1258
-
1259
- <div class="step-item">
1260
- <div class="step-marker">2</div>
1261
- <div class="step-details">
1262
- <p class="step-title">Install Dependencies</p>
1263
- <p class="is-size-7 has-text-grey">
1264
- <code>pip install -r requirements.txt</code>
1265
- </p>
1266
- </div>
1267
- </div>
1268
-
1269
- <div class="step-item">
1270
- <div class="step-marker">3</div>
1271
- <div class="step-details">
1272
- <p class="step-title">Configure Models</p>
1273
- <p class="is-size-7 has-text-grey">Edit config.yaml with your API keys and model selections</p>
1274
- </div>
1275
- </div>
1276
-
1277
- <div class="step-item">
1278
- <div class="step-marker">4</div>
1279
- <div class="step-details">
1280
- <p class="step-title">Run Evaluations</p>
1281
- <p class="is-size-7 has-text-grey">
1282
- <code>python -m flame.run --task all</code>
1283
- </p>
1284
- </div>
1285
- </div>
1286
- </div>
1287
-
1288
- <div class="notification is-info is-light mt-4">
1289
- <p class="mb-0">For complete documentation and examples, visit our <a href="https://github.com/flame-benchmark/flame" target="_blank">GitHub repository</a>.</p>
1290
- </div>
1291
  </div>
1292
  </div>
1293
  </div>
 
 
 
 
1294
  </div>
1295
  </div>
1296
  </div>
1297
  </div>
1298
- </div>
1299
- <!--/ Framework -->
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1300
  </div>
1301
  </section>
1302
 
 
1063
 
1064
  <section id="framework" class="section content-section">
1065
  <div class="container is-max-desktop">
1066
+
1067
+ <!-- Framework Overview -->
1068
  <div class="columns is-centered">
1069
  <div class="column is-full">
1070
  <h2 class="title is-3 section-title has-text-centered">Framework & Resources</h2>
1071
+
1072
+ <div class="box has-background-white-ter has-text-centered p-5 mb-5">
1073
+ <span class="icon is-large mb-4">
1074
+ <i class="fas fa-fire flame-icon fa-3x"></i>
1075
+ </span>
1076
+ <h3 class="title is-4"><span class="flame">FLaME</span> Framework</h3>
1077
+ <p class="subtitle">An open-source, modular benchmarking suite for evaluating financial language models.</p>
1078
+
1079
+ <div class="buttons is-centered mt-4">
1080
+ <a href="https://github.com/flame-benchmark/flame" target="_blank" class="button is-primary">
1081
+ <span class="icon"><i class="fab fa-github"></i></span>
1082
+ <span>GitHub Repository</span>
1083
+ </a>
1084
+ <a href="https://huggingface.co/spaces/flame-benchmark/flame" target="_blank" class="button is-info">
1085
+ <span class="icon"><i class="fab fa-huggingface"></i></span>
1086
+ <span>Hugging Face Space</span>
1087
+ </a>
 
 
 
 
 
 
 
 
 
1088
  </div>
1089
+ </div>
1090
+ </div>
1091
+ </div>
1092
+
1093
+ <!-- Features & Getting Started -->
1094
+ <div class="columns is-multiline">
1095
+
1096
+ <!-- πŸ›  Framework Features -->
1097
+ <div class="column is-6">
1098
+ <div class="card h-100">
1099
+ <div class="card-header">
1100
+ <p class="card-header-title">
1101
+ <span class="icon mr-2"><i class="fas fa-cogs"></i></span>
1102
+ Framework Features
1103
+ </p>
1104
+ </div>
1105
+ <div class="card-content">
1106
+ <div class="content">
1107
+ <div class="feature-list">
1108
+
1109
+ <div class="feature-item mb-3">
1110
+ <p class="has-text-weight-bold mb-1">
1111
+ <span class="icon has-text-primary"><i class="fas fa-check"></i></span> πŸ”§ Standardized Pipelines
1112
  </p>
1113
+ <p class="is-size-7 ml-4">Pre-built evaluation pipelines for key financial NLP tasks.</p>
1114
  </div>
1115
+
1116
+ <div class="feature-item mb-3">
1117
+ <p class="has-text-weight-bold mb-1">
1118
+ <span class="icon has-text-primary"><i class="fas fa-check"></i></span> βš™οΈ Customizable Assessments
1119
+ </p>
1120
+ <p class="is-size-7 ml-4">Supports easy model-to-model and dataset-to-dataset comparisons.</p>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1121
  </div>
1122
+
1123
+ <div class="feature-item mb-3">
1124
+ <p class="has-text-weight-bold mb-1">
1125
+ <span class="icon has-text-primary"><i class="fas fa-check"></i></span> πŸ“Š Reproducible Benchmarking
 
 
 
 
 
 
1126
  </p>
1127
+ <p class="is-size-7 ml-4">Ensures consistent evaluation metrics and transparent methodology.</p>
1128
  </div>
1129
+
1130
+ <div class="feature-item">
1131
+ <p class="has-text-weight-bold mb-1">
1132
+ <span class="icon has-text-primary"><i class="fas fa-check"></i></span> πŸ”— Extensible Architecture
1133
+ </p>
1134
+ <p class="is-size-7 ml-4">Easily integrates new tasks, datasets, and evaluation modules.</p>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1135
  </div>
1136
+
1137
  </div>
1138
  </div>
1139
+ </div>
1140
+ </div>
1141
+ </div>
1142
+
1143
+ <!-- πŸš€ Getting Started -->
1144
+ <div class="column is-6">
1145
+ <div class="card h-100">
1146
+ <div class="card-header">
1147
+ <p class="card-header-title">
1148
+ <span class="icon mr-2"><i class="fas fa-rocket"></i></span>
1149
+ Getting Started
1150
+ </p>
1151
+ </div>
1152
+ <div class="card-content">
1153
+ <div class="content">
1154
+ <div class="steps">
1155
+ <div class="step-item">
1156
+ <div class="step-marker">1</div>
1157
+ <div class="step-details">
1158
+ <p class="step-title">Clone Repository</p>
1159
+ <p class="is-size-7 has-text-grey">
1160
+ <code>git clone https://github.com/flame-benchmark/flame.git</code>
1161
+ </p>
1162
+ </div>
1163
  </div>
1164
+
1165
+ <div class="step-item">
1166
+ <div class="step-marker">2</div>
1167
+ <div class="step-details">
1168
+ <p class="step-title">Install Dependencies</p>
1169
+ <p class="is-size-7 has-text-grey">
1170
+ <code>pip install -r requirements.txt</code>
1171
+ </p>
1172
+ </div>
1173
+ </div>
1174
+
1175
+ <div class="step-item">
1176
+ <div class="step-marker">3</div>
1177
+ <div class="step-details">
1178
+ <p class="step-title">Configure Models</p>
1179
+ <p class="is-size-7 has-text-grey">Edit <code>config.yaml</code> with API keys and model settings.</p>
1180
+ </div>
1181
+ </div>
1182
+
1183
+ <div class="step-item">
1184
+ <div class="step-marker">4</div>
1185
+ <div class="step-details">
1186
+ <p class="step-title">Run Evaluations</p>
1187
+ <p class="is-size-7 has-text-grey">
1188
+ <code>python -m flame.run --task all</code>
1189
+ </p>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1190
  </div>
1191
  </div>
1192
  </div>
1193
+
1194
+ <div class="notification is-info is-light mt-4">
1195
+ <p class="mb-0">For detailed documentation, visit our <a href="https://github.com/flame-benchmark/flame" target="_blank">GitHub repository</a>.</p>
1196
+ </div>
1197
  </div>
1198
  </div>
1199
  </div>
1200
  </div>
1201
+
1202
+ </div> <!-- End Columns -->
1203
+
1204
+ </div>
1205
+ </section>
1206
+
1207
+ <section id="datasets" class="section content-section">
1208
+ <div class="container is-max-desktop">
1209
+ <h2 class="title is-3 section-title has-text-centered">Datasets & Domains</h2>
1210
+
1211
+ <div class="columns is-multiline">
1212
+
1213
+ <!-- πŸ“Š Numerical Reasoning & Question Answering -->
1214
+ <div class="column is-6">
1215
+ <div class="dataset-category box">
1216
+ <p class="has-text-weight-bold">
1217
+ <span class="icon has-text-primary"><i class="fa-solid fa-calculator"></i></span> πŸ“Š Numerical Reasoning & Question Answering
1218
+ </p>
1219
+ <ul>
1220
+ <li><strong>FinQA</strong> – Multi-step financial numerical reasoning.</li>
1221
+ <li><strong>ConvFinQA</strong> – Conversational numerical reasoning.</li>
1222
+ <li><strong>TaT-QA</strong> – Hybrid table-text question answering.</li>
1223
+ </ul>
1224
+ </div>
1225
+ </div>
1226
+
1227
+ <!-- πŸ“ Text Summarization -->
1228
+ <div class="column is-6">
1229
+ <div class="dataset-category box">
1230
+ <p class="has-text-weight-bold">
1231
+ <span class="icon has-text-primary"><i class="fa-solid fa-file-lines"></i></span> πŸ“ Text Summarization
1232
+ </p>
1233
+ <ul>
1234
+ <li><strong>ECTSum</strong> – Earnings call transcript summarization.</li>
1235
+ <li><strong>EDTSum</strong> – Financial news summarization.</li>
1236
+ </ul>
1237
+ </div>
1238
+ </div>
1239
+
1240
+ <!-- πŸ”Ž Information Retrieval -->
1241
+ <div class="column is-6">
1242
+ <div class="dataset-category box">
1243
+ <p class="has-text-weight-bold">
1244
+ <span class="icon has-text-primary"><i class="fa-solid fa-search"></i></span> πŸ”Ž Information Retrieval
1245
+ </p>
1246
+ <ul>
1247
+ <li><strong>FiNER-ORD</strong> – Named entity recognition for financial documents.</li>
1248
+ <li><strong>FinEntity</strong> – Entity-based sentiment classification.</li>
1249
+ <li><strong>Financial Numeric Extreme Labeling (FNXL)</strong> – Automated numeral annotation in financial reports.</li>
1250
+ <li><strong>FinRED</strong> – Relation extraction in finance.</li>
1251
+ <li><strong>REFinD</strong> – Relation extraction from SEC filings.</li>
1252
+ </ul>
1253
+ </div>
1254
+ </div>
1255
+
1256
+ <!-- 😐 Sentiment Analysis -->
1257
+ <div class="column is-6">
1258
+ <div class="dataset-category box">
1259
+ <p class="has-text-weight-bold">
1260
+ <span class="icon has-text-primary"><i class="fa-solid fa-comment-alt"></i></span> 😐 Sentiment Analysis
1261
+ </p>
1262
+ <ul>
1263
+ <li><strong>FiQA (Task 1)</strong> – Aspect-based sentiment analysis.</li>
1264
+ <li><strong>FiQA (Task 2)</strong> – Opinion-based financial QA.</li>
1265
+ <li><strong>Financial Phrase Bank (FPB)</strong> – Market sentiment classification.</li>
1266
+ <li><strong>SubjECTive-QA</strong> – Subjectivity detection in earnings call Q&A.</li>
1267
+ </ul>
1268
+ </div>
1269
+ </div>
1270
+
1271
+ <!-- 🏷️ Text Classification -->
1272
+ <div class="column is-6">
1273
+ <div class="dataset-category box">
1274
+ <p class="has-text-weight-bold">
1275
+ <span class="icon has-text-primary"><i class="fa-solid fa-tags"></i></span> 🏷️ Text Classification
1276
+ </p>
1277
+ <ul>
1278
+ <li><strong>Numerical Claim Detection</strong> – Fine-grained investor claim detection.</li>
1279
+ <li><strong>News Headline Classification</strong> – Market sentiment and trend detection.</li>
1280
+ <li><strong>FOMC Dataset</strong> – Hawkish-Dovish stance classification.</li>
1281
+ <li><strong>Banking77</strong> – Fine-grained intent detection.</li>
1282
+ <li><strong>FinBench</strong> – Financial risk classification (default, fraud, churn).</li>
1283
+ </ul>
1284
+ </div>
1285
+ </div>
1286
+
1287
+ <!-- Causal Analysis -->
1288
+ <div class="column is-6">
1289
+ <div class="dataset-category box">
1290
+ <p class="has-text-weight-bold">
1291
+ <span class="icon"><i class="fa-solid fa-brain"></i></span> 🧠 Causal Analysis
1292
+ </p>
1293
+ <ul>
1294
+ <li><strong>FinCausal</strong> – Causal reasoning in financial news.</li>
1295
+ <li><strong>FinCausal-SC</strong> – Cause-effect span extraction.</li>
1296
+ </ul>
1297
+ </div>
1298
+ </div>
1299
+
1300
+ </div> <!-- End Columns -->
1301
  </div>
1302
  </section>
1303