Datasets:
File size: 92,413 Bytes
76d4cf5 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 | [
{
"type": "text",
"text": "SpeechGPT: Empowering Large Language Models with Intrinsic Cross-Modal Conversational Abilities ",
"text_level": 1,
"bbox": [
159,
79,
840,
118
],
"page_idx": 0
},
{
"type": "text",
"text": "Dong Zhang, Shimin Li, Xin Zhang, Jun Zhan, Pengyu Wang, Yaqian Zhou∗, Xipeng Qiu∗ School of Computer Science, Fudan University \nShanghai Key Laboratory of Intelligent Information Processing, Fudan University \n{dongzhang22,xin_zhang22,jzhan22,pywang22}@m.fudan.edu.cn {smli20,zhouyaqian,xpqiu}@fudan.edu.cn ",
"bbox": [
159,
124,
847,
225
],
"page_idx": 0
},
{
"type": "text",
"text": "Abstract ",
"text_level": 1,
"bbox": [
263,
252,
339,
268
],
"page_idx": 0
},
{
"type": "text",
"text": "Multi-modal large language models are regarded as a crucial step towards Artificial General Intelligence (AGI) and have garnered significant interest with the emergence of ChatGPT. However, current speech-language models typically adopt the cascade paradigm, preventing inter-modal knowledge transfer. In this paper, we propose SpeechGPT, a large language model with intrinsic cross-modal conversational abilities, capable of perceiving and generating multi-modal content. With discrete speech representations, we construct SpeechInstruct, the first large-scale crossmodal speech instruction dataset. Additionally, we employ a three-stage training strategy that includes modality-adaptation pretraining, cross-modal instruction fine-tuning, and chain-of-modality instruction fine-tuning. The experimental results demonstrate that SpeechGPT has an impressive capacity to follow cross-modal human instructions and highlight the potential of handling multiple modalities with one model. Code and models are available in https://github.com/ 0nutation/SpeechGPT. Demos are shown in https://0nutation.github. io/SpeechGPT.github.io/. ",
"bbox": [
146,
282,
458,
664
],
"page_idx": 0
},
{
"type": "text",
"text": "1 Introduction ",
"text_level": 1,
"bbox": [
117,
678,
258,
695
],
"page_idx": 0
},
{
"type": "text",
"text": "Large language models (OpenAI, 2023; Touvron et al., 2023) have performed astonishingly on various natural language processing tasks. Meanwhile, multi-modal large language models, such as GPT4, PALM-E (Driess et al., 2023), and LLaVA (Liu et al., 2023), have explored the ability of LLMs to understand multi-modal information. However, a significant gap exists between current LLMs and general artificial intelligence (AGI). First, most current LLMs can only perceive and understand multimodal content but cannot spontaneously generate multi-modal content. Second, continuous signals like images and speech cannot be adapted directly to LLMs that receive discrete tokens. ",
"bbox": [
115,
706,
487,
896
],
"page_idx": 0
},
{
"type": "image",
"img_path": "images/2914927effd25a1062b348dbbc7abea459d3b55a988bd889ae878e885b349122.jpg",
"image_caption": [
"Figure 1: SpeechGPT’s capabilities to tackle multiple cross-modal tasks. "
],
"image_footnote": [],
"bbox": [
514,
252,
878,
390
],
"page_idx": 0
},
{
"type": "text",
"text": "",
"bbox": [
510,
442,
880,
472
],
"page_idx": 0
},
{
"type": "text",
"text": "The current speech-language model mainly adopts a cascading paradigm (Huang et al., 2023a) i.e., the LLM is connected with an automatic speech recognition (ASR) model or a text-tospeech (TTS) model in tandem, or the LLM is employed as a control hub, with several speech processing models (Cheng et al., 2023a,b,c) are integrated to cover multiple audio or speech tasks (Huang et al., 2023a; Shen et al., 2023). Some prior work on generative spoken language models involves encoding the speech signal into a discrete representation (Baevski et al., 2020; Hsu et al., 2021; Zhang et al., 2023a) and modeling it with language models (Lakhotia et al., 2021; Borsos et al., 2022; Zhang et al., 2023d; Wang et al., 2023; Zhang et al., 2023c). ",
"bbox": [
512,
481,
882,
736
],
"page_idx": 0
},
{
"type": "text",
"text": "While capable of perceiving and generating speech, the existing cascaded methods or spoken language models still have several limitations. First, the LLM in the cascaded model only functions as a content generator. Since the representations of speech and text are not aligned, the LLM’s knowledge cannot be transferred to the speech modality. Second, the cascade approach (Shen et al., 2023; Huang et al., 2023a) suffers from the loss of paralinguistic signals such as emotion and prosody. Third, existing spoken language models (Wang et al., 2023; Zhang et al., 2023d) only synthesize speech but fail to comprehend its semantic information, preventing them from achieving true crossmodal perception and generation. ",
"bbox": [
512,
744,
882,
919
],
"page_idx": 0
},
{
"type": "text",
"text": "",
"bbox": [
115,
85,
487,
148
],
"page_idx": 1
},
{
"type": "text",
"text": "In this paper, we propose SpeechGPT, a large language model with intrinsic cross-modal conversational abilities, capable of perceiving and generating multi-modal content. We perform speech discretization with a self-supervised trained speech model to unify the modality between speech and text. The discrete speech tokens are then expanded into the vocabulary of the LLM, thus endowing the model with an inherent competence to perceive and generate the speech. ",
"bbox": [
117,
152,
487,
311
],
"page_idx": 1
},
{
"type": "text",
"text": "To provide the model with the capacity to handle multi-modal instructions, we build the first speech-text cross-modal instruction-following dataset SpeechInstruct. Specifically, we discretize the speech to discrete units (Hsu et al., 2021) and construct the cross-modal unit-text pair based on the existing ASR dataset. Meanwhile, we construct hundreds of instructions for diverse tasks with GPT4 to simulate actual user instructions as illustrated in Appendix B. In addition, to further enhance the model’s cross-modal capability, we designed the Chain-of-Modality instruction data, i.e., the model receives the speech command, thinks about the process in text, and then outputs the response in speech. ",
"bbox": [
115,
315,
485,
555
],
"page_idx": 1
},
{
"type": "text",
"text": "For better cross-modal transfer and efficient training, SpeechGPT undergoes a three-stage training process: modality-adaptation pre-training, cross-modal instruction fine-tuning, and chain-ofmodality instruction fine-tuning. The first stage enables speech comprehension for SpeechGPT with the discrete speech unit continuation task. The second stage employs the SpeechInstruct to improve the model’s cross-modal capabilities. The third stage utilizes parameter-efficient LoRA (Hu et al., 2021) fine-tuning for further modality alignment. ",
"bbox": [
117,
558,
487,
734
],
"page_idx": 1
},
{
"type": "text",
"text": "To evaluate the effectiveness of SpeechGPT, we conduct a wide range of human evaluations and case analyses to estimate the performance of SpeechGPT on textual tasks, speech-text crossmodal tasks, and spoken dialogue tasks. The results demonstrate that SpeechGPT exhibits a strong ability for unimodal and cross-modal instruction following tasks. ",
"bbox": [
117,
738,
487,
865
],
"page_idx": 1
},
{
"type": "text",
"text": "Our contributions include the following: ",
"bbox": [
132,
869,
433,
884
],
"page_idx": 1
},
{
"type": "text",
"text": "• We build the first multi-modal large language model that can perceive and generate multi",
"bbox": [
119,
888,
487,
919
],
"page_idx": 1
},
{
"type": "text",
"text": "modal contents. \n• We construct and release SpeechInstruct, the first large-scale speech-text cross-modal instructionfollowing dataset. \n• We build the first spoken dialogue LLM with strong human instruction following ability and spoken dialogue ability. \n• We show great potential to incorporate other modalities into LLMs through discrete representations. ",
"bbox": [
514,
85,
884,
247
],
"page_idx": 1
},
{
"type": "text",
"text": "2 Related Work ",
"text_level": 1,
"bbox": [
512,
263,
663,
279
],
"page_idx": 1
},
{
"type": "text",
"text": "Multi-modal Large Language Model Current multi-modal LLMs predominantly focus on the visual domain, feeding continuous representations obtained from pre-trained visual encoders into LLMs, facilitating full-parameter or parameterefficient training on visual-language data (OpenAI, 2023; Huang et al., 2023b; Zhang et al., 2023b). Palm-E (Driess et al., 2023) integrates the 540B PaLM (Chowdhery et al., 2022) and 22B Vision Transformer (Dosovitskiy et al., 2021) into the largest vision-language model. LLaVA (Liu et al., 2023) leverages pre-trained CLIP (Radford et al., 2021) visual encoder and LLaMA (Touvron et al., 2023) and conduct instruct tuning on GPT4- assisted visual instruction data. X-LLM (Chen et al., 2023) converts multi-modalities into representations with X2L interfaces as the inputs of the large language model. However, such structures only enable LLMs to process multi-modal input, without ability to generate multi-modal output. Diverging from prior studies, our approach emphasizes the development of a speech-centric multimodal LLM, endowing it with the proficiency to accommodate both multi-modal input and output. ",
"bbox": [
510,
292,
882,
676
],
"page_idx": 1
},
{
"type": "text",
"text": "Generative Spoken Language Model Discrete self-supervised representation based spoken generative language modeling is making remarkable progress on large-scale speech dataset training (Nguyen et al., 2022). AudioLM (Borsos et al., 2022) proposes to model speech based on audio codecs together with semantic codes, which can synthesize speech in a textlesss setting. VALLE (Wang et al., 2023) builds a generative spoken language model on audio codecs and treat Textto-Speech as a conditional generation task. However, these models are designed for a specific task and failed to benefit from LLMs. SpeechGPT is built upon the foundation of LLM and transfers LLM’s knowledge to speech modality, consequently obtaining better task generalization and human-instruction following ability. ",
"bbox": [
512,
680,
884,
919
],
"page_idx": 1
},
{
"type": "text",
"text": "",
"bbox": [
115,
85,
485,
115
],
"page_idx": 2
},
{
"type": "text",
"text": "Speech-Enabled LLM Interaction Following the emergence of ChatGPT, several studies have concentrated on the integration of expert speech models with LLMs to enable direct speech interaction with LLMs. HuggingGPT (Shen et al., 2023) facilitates task decomposition of human instructions by LLMs and allows the invocation of models from Huggingface to accomplish specific tasks, encompassing a range of automatic speech recognition (ASR) and text-to-speech models. AudioGPT (Huang et al., 2023a) leverages a variety of audio foundation models to process complex audio information and connect LLMs with input/output interface (ASR, TTS) for speech conversations. However, these models exhibit increased complexity, demand extensive resources, and are prone to the unavoidable error accumulation problems. Our approach enables speech interaction with LLMs without relying on ASR or TTS systems, circumventing the aforementioned drawbacks. ",
"bbox": [
115,
117,
487,
438
],
"page_idx": 2
},
{
"type": "text",
"text": "3 SpeechInstruct Construction ",
"text_level": 1,
"bbox": [
117,
451,
396,
468
],
"page_idx": 2
},
{
"type": "text",
"text": "Due to the limitations in publicly available speech data and the lack of variety of speech-text tasks, we construct SpeechInstruct, a speech-text crossmodal instruction-following dataset. This dataset consists of two parts, the first part is called CrossModal Instruction, and the second part is called Chain-of-Modality Instruction. The construction process of SpeechInstruct is illustrated in Figure 2. ",
"bbox": [
115,
479,
487,
607
],
"page_idx": 2
},
{
"type": "text",
"text": "3.1 Cross-modal Instruction ",
"text_level": 1,
"bbox": [
117,
620,
352,
634
],
"page_idx": 2
},
{
"type": "text",
"text": "Data Collection We collect several large-scale English ASR datasets to construct Cross-Modal Instruction, including Gigaspeech (Chen et al., 2021), Common Voice (Ardila et al., 2020), and LibriSpeech (Panayotov et al., 2015). We employ mHuBERT1 as the speech tokenizer to discretize speech data into discrete units and remove the repetitive units of adjacent frames to get reduced units. Ultimately, we obtain 9 million unit-text data pairs. ",
"bbox": [
117,
642,
487,
785
],
"page_idx": 2
},
{
"type": "text",
"text": "Task Description Generation We generate ASR and TTS task descriptions that are compatible with speech-text data pairs. Unlike the Self-Instruct method (Wang et al., 2022), we generate descriptions through a zero-shot approach. Specifically, we directly input the prompts shown in Appendix A into OpenAI GPT-4 to generate task descriptions. Our generation method yields 100 instructions for each task and some examples are shown in Appendix B. ",
"bbox": [
115,
788,
487,
883
],
"page_idx": 2
},
{
"type": "text",
"text": "",
"bbox": [
510,
85,
882,
147
],
"page_idx": 2
},
{
"type": "text",
"text": "Instruction Formatting For a discrete unit sequence $U$ and its associated transcription $T$ , we determine whether it will be used for constructing an ASR task or a TTS task based on the probability $p$ . Subsequently, we randomly select a description $D$ from the corresponding task description. This results in a triplet consisting of the task description, discrete unit sequence, and transcription, denoted as $( D , U , T )$ . Following this, the triplet is assembled into an instruction using the template: [Human]: $\\{ D \\}$ . This is input: $\\{ U \\}$ <eoh>.[SpeechGPT]: $\\{ T \\} { \\ < } { \\bf e } { \\bf 0 } { \\bf s } > .$ .. ",
"bbox": [
512,
151,
882,
343
],
"page_idx": 2
},
{
"type": "text",
"text": "3.2 Chain-of-Modality Instruction ",
"text_level": 1,
"bbox": [
514,
356,
794,
372
],
"page_idx": 2
},
{
"type": "text",
"text": "Speech Instruction Generation Due to the lack of instruction data with speech input and speech output, we trained a text-to-unit generator to convert text instruction data into speech instruction data. Specifically, the text-to-unit generator adopts a Transformer encoder-decoder architecture. We trained it on LibriSpeech unit-text pairs in Crossmodal Instruction. We select 37,969 samples from the moss-002-sft-data dataset 2 whose response length is shorter than 35 words. And we convert both their instructions and responses into unit sequences through the text-to-unit generator. As a result, we obtained 37,969 quadruplets composed of speech instructions, text instructions, text responses, and speech responses, denoted as (SpeechI, T extI, T extR, SpeechR). ",
"bbox": [
512,
380,
882,
634
],
"page_idx": 2
},
{
"type": "text",
"text": "Instruction Formatting Using the above quadruplets, we could construct chain-of-thought style instructions for four input-output formats, namely Speech Instruction-Speech Response, Speech Instruction-Text Response, Text Instruction-Speech Response, and Text Instruction-Text Response. Their corresponding templates can be found in Appendix C. ",
"bbox": [
512,
638,
882,
765
],
"page_idx": 2
},
{
"type": "text",
"text": "3.3 SpeechInstruct Evaluation Set ",
"text_level": 1,
"bbox": [
514,
780,
791,
795
],
"page_idx": 2
},
{
"type": "text",
"text": "We constructed cross-modal dialogue datasets under different scenarios to evaluate whether SpeechGPT could take on various roles. Specifically, these included a talking encyclopedia, personal assistant, chat partner, poet, psychologist, and educational assistant. For each role, we provide 10 manually authored instruction-response pairs written by ourselves. We use a pre-trained text-to-speech model 3 to convert the text into corresponding speech. We then employ mHuBERT to discretize speech data into discrete units as described in Section 3.1. Ultimately, for each role, we obtained 10 quadruplets composed of speech instructions, text instructions, text responses, and speech responses. ",
"bbox": [
512,
802,
882,
882
],
"page_idx": 2
},
{
"type": "image",
"img_path": "images/18e296fb0af1200b4362fce127e53612728fab4197c4d19c9740c246b9e519ce.jpg",
"image_caption": [
"Figure 2: Left: An overview of SpeechInstruct construction process. The SpeechInstruct dataset consists of two parts: Cross-modal Instruction data and Chain-of-Modality Instruction data. T emplate1 is shown in 3.1. T emplate2 is shown in Appendix C. Right: An illustration of SpeechGPT model structure. "
],
"image_footnote": [],
"bbox": [
121,
82,
884,
286
],
"page_idx": 3
},
{
"type": "text",
"text": "",
"bbox": [
117,
366,
487,
526
],
"page_idx": 3
},
{
"type": "text",
"text": "4 SpeechGPT ",
"text_level": 1,
"bbox": [
117,
541,
250,
557
],
"page_idx": 3
},
{
"type": "text",
"text": "4.1 Model Structure ",
"text_level": 1,
"bbox": [
117,
569,
289,
585
],
"page_idx": 3
},
{
"type": "text",
"text": "A unified framework is designed to provide architecture compatibility across different modalities. As shown in Figure 2, our model consists of three main components: discrete unit extractor, large language modal and unit vocoder. Under this architecture, LLM can perceive multi-modal inputs and generate multi-modal outputs. ",
"bbox": [
117,
592,
487,
703
],
"page_idx": 3
},
{
"type": "text",
"text": "Discrete Unit Extractor The discrete unit extractor utilizes the Hidden-unit BERT (HuBERT) model (Hsu et al., 2021) to transform continuous speech signals into a sequence of discrete units, . HuBERT is a self-supervised model that learns by predicting discrete labels for masked audio segments based on $\\mathbf { k }$ -means clustering applied to the model’s intermediate representations. It features a combination of 1-D convolutional layers and a Transformer encoder to encode speech into continuous intermediate representations, with a kmeans model further converting these representations into a sequence of cluster indices. Subsequently, adjacent duplicate indices are removed, resulting in a discrete units sequence represented as $U = ( u _ { 1 } , u _ { 2 } , . . . , u _ { T } )$ , $u _ { i } \\in { 0 , 1 , . . . , K - 1 }$ , $\\forall 1 \\leq i \\leq T$ , with $K$ denoting the total number of clusters. ",
"bbox": [
115,
706,
487,
881
],
"page_idx": 3
},
{
"type": "text",
"text": "",
"bbox": [
512,
366,
884,
478
],
"page_idx": 3
},
{
"type": "text",
"text": "Large Language Model We employ the Meta AI LLaMA (Touvron et al., 2023) model as our Large Language Model. LLaMA comprises an embedding layer, multiple transformer blocks, and an LM head layer. The total number of parameters in LLaMA ranges from 7B to 65B. Drawing from an extensive training dataset of 1.0 trillion tokens, LLaMA demonstrates competitive performance compared to the substantially larger 175B GPT-3 across various NLP benchmarks. ",
"bbox": [
512,
499,
882,
657
],
"page_idx": 3
},
{
"type": "text",
"text": "Unit Vocoder Due to limition of single speaker unit vocoder in (Polyak et al., 2021), we train a multi-speaker unit HiFi-GAN to decode the speech signal from the discrete representation. The HiFiGAN architecture consists of a generator $\\mathbf { G }$ and multiple discriminators D. The generator uses look-up tables (LUT) to embed discrete representations and the embedding sequences are up-sampled by a series of blocks composed of transposed convolution and a residual block with dilated layers. The speaker embedding is concatenated to each frame in the up-sampled sequence. The discriminator features a Multi-Period Discriminator (MPD) and a Multi-Scale Discriminator (MSD), which have the same architecture as (Polyak et al., 2021). ",
"bbox": [
512,
680,
882,
919
],
"page_idx": 3
},
{
"type": "text",
"text": "4.2 Training ",
"text_level": 1,
"bbox": [
117,
84,
228,
99
],
"page_idx": 4
},
{
"type": "text",
"text": "To incorporate speech discrete representation into LLM, we expand the vocabulary and corresponding embedding matrix first. We divide the training process into three stages. The first stage is ModalityAdaptation Pre-training on unpaired speech data. The second stage is Cross-modal Instruction FineTuning. The third stage is Chain-of-Modality Instruction Fine-Tuning. ",
"bbox": [
115,
107,
487,
236
],
"page_idx": 4
},
{
"type": "text",
"text": "Expanding Vocabulary Given original LLM vocabulary $V$ of size $| V |$ , to integrate speech discrete representations into LLM, we expand the vocabulary with an additional set of unit tokens $V ^ { \\prime }$ , of size $| V ^ { \\prime } | = K$ . The expanded vocabulary $V ^ { \\prime \\prime }$ is the union of the original vocabulary $V$ and the new words $V ^ { \\prime }$ : ",
"bbox": [
115,
237,
487,
348
],
"page_idx": 4
},
{
"type": "equation",
"img_path": "images/ecf3e6ed2da4cca7d61a363613f9a1609345ce3c1d02ce37899c0cd76172bc7f.jpg",
"text": "$$\nV ^ { \\prime \\prime } = V \\cup V ^ { \\prime }\n$$",
"text_format": "latex",
"bbox": [
247,
367,
356,
385
],
"page_idx": 4
},
{
"type": "text",
"text": "We denote the original word embedding matrix as $E \\in \\mathbb { R } ^ { | V | \\times d }$ , where $d$ is the dimension of word embeddings. To accommodate the expanded vocabulary, we need to create a randomly initialized word embedding matrix $E ^ { \\prime } \\in \\mathbb { R } ^ { | V ^ { \\prime \\prime } | \\times d }$ . We preserve the original word embeddings by copying the values of $E$ to the first $| V |$ rows of $E ^ { \\prime }$ : ",
"bbox": [
115,
397,
487,
510
],
"page_idx": 4
},
{
"type": "equation",
"img_path": "images/2a5ff459674a5e6a9098d06007bb122a5347290ea0bb0eb37dcdfe681b740ee4.jpg",
"text": "$$\nE ^ { \\prime } [ 0 : | V | , : ] = E\n$$",
"text_format": "latex",
"bbox": [
233,
527,
369,
546
],
"page_idx": 4
},
{
"type": "text",
"text": "Finally, we replace the original vocabulary and word embedding matrix with the new vocabulary $V ^ { \\prime \\prime }$ and the word embedding matrix $E ^ { \\prime }$ . ",
"bbox": [
115,
557,
487,
606
],
"page_idx": 4
},
{
"type": "text",
"text": "Stage 1: Modality-Adaptation Pre-training To enable LLM to handle discrete units modality, we utilize an unlabeled speech corpus to train LLM in a next-token prediction task. This approach aligns with the text pre-training objective of LLM. Given unlabeled speech corpus $C$ consisting of speech $U _ { 1 } , U _ { 2 } , \\dots , U _ { m }$ and LLM denoted as $L _ { 1 }$ , the negative log-likelihood loss can be formulated as: ",
"bbox": [
115,
607,
487,
736
],
"page_idx": 4
},
{
"type": "equation",
"img_path": "images/5c24d3200a6c1c1a8ab469f2b58bd7a939e61229813d75742509df2d4030eccd.jpg",
"text": "$$\n\\mathcal { L } ( L | C ) = - \\sum _ { j = 1 } ^ { m } \\sum _ { i = 1 } ^ { n _ { j } } \\log P ( u _ { i , j } | u _ { < i , j } ; L )\n$$",
"text_format": "latex",
"bbox": [
136,
763,
447,
809
],
"page_idx": 4
},
{
"type": "text",
"text": "where $m$ is the number of speech in dataset $C$ , $n _ { j }$ is the number of discrete unit token in speech $U _ { j }$ , and $u _ { i , j }$ represents the i-th unit token in the $\\mathrm { j } \\cdot$ -th speech. ",
"bbox": [
115,
822,
487,
887
],
"page_idx": 4
},
{
"type": "text",
"text": "Stage 2: Cross-modal Instruction FineTuning In this stage, we align speech and text modalities utilizing paired data. We mix Crossmodal Instruction in SpeechInstruct with moss-002- sft dataset to derive mix dataset $I$ , which consists of samples $T _ { 1 } , T _ { 2 } , \\dots , T _ { x }$ . We fine-tune the model $L$ obtained from the first stage on $I$ . ",
"bbox": [
115,
888,
487,
920
],
"page_idx": 4
},
{
"type": "text",
"text": "",
"bbox": [
510,
84,
884,
164
],
"page_idx": 4
},
{
"type": "text",
"text": "Each sample $T _ { j }$ consisting of $t _ { 1 } , t _ { 2 } , \\ldots , t _ { n _ { j } }$ is formed by concatenating a prefix and a text. The training objective is to minimize the negative loglikelihood and the loss calculation only considers the text part, ignoring the prefix, which can be formated as: ",
"bbox": [
512,
166,
884,
260
],
"page_idx": 4
},
{
"type": "equation",
"img_path": "images/736c4a0fae6a175438d4a57d6bf4d60b0599e47380be66ebaf270706a95705b4.jpg",
"text": "$$\n\\mathcal { L } ( L | I ) = - \\sum _ { j = 1 } ^ { x } \\sum _ { i = p _ { j } + 1 } ^ { y _ { j } } \\log P ( t _ { i , j } | t _ { < i , j } ; L )\n$$",
"text_format": "latex",
"bbox": [
526,
285,
847,
334
],
"page_idx": 4
},
{
"type": "text",
"text": "where $x$ is the number of samples in corpus $I$ , $y _ { j }$ is the total number of tokens in sample $T _ { j } , p _ { j }$ is the number of tokens in the prefix part of $T _ { j }$ , and $t _ { i , j }$ represents the i-th word in $T _ { j }$ . ",
"bbox": [
510,
344,
882,
409
],
"page_idx": 4
},
{
"type": "text",
"text": "Stage 3: Chain-of-Modality Instruction FineTuning After obtaining the model in stage 2, we utilizes parameter-efficient Low-Rank Adaptation (LoRA) (Hu et al., 2021) to fine-tune it on Chain-of-Modality Instruction in SpeechInstruct. We add LoRA weights (adapters) to the attention mechanisms and train the newly added LoRA parameters. We adopt the same loss function as stage 2. ",
"bbox": [
510,
410,
884,
551
],
"page_idx": 4
},
{
"type": "text",
"text": "5 Experiments ",
"text_level": 1,
"bbox": [
512,
567,
653,
583
],
"page_idx": 4
},
{
"type": "text",
"text": "5.1 Experimental Setups ",
"text_level": 1,
"bbox": [
512,
592,
719,
608
],
"page_idx": 4
},
{
"type": "text",
"text": "Datasets For modality-adaption pre-training, we use LibriLight (Kahn et al., 2020) which contains 60K hours of unlabelled English audiobook speech. For cross-modal instruction fine-tuning stage, we use Gigaspeech (Chen et al., 2021), Common voice (Ardila et al., 2020) and LibriSpeech (Panayotov et al., 2015) dataset and moss-002-sft-data dataset, which is illustrated in detail in 3.1. For chain-of-modality instruction fine-tuning stage, we use moss-002-sft-data dataset, which is illustrated in detail in 3.2. ",
"bbox": [
510,
614,
882,
790
],
"page_idx": 4
},
{
"type": "text",
"text": "Configuration We employ LLaMA-13B (Touvron et al., 2023) as our backbone model for a trade-off between performance and computational resources available. For stage 1, we use 96 A100 GPUs and train for 900 steps with batch size 768. For stage 2, we use 96 A100 GPUs and train for 2100 steps with batch size 1536. For stage 3, we use 8 A100 GPUs and train for 4200 steps with batch size 128. ",
"bbox": [
510,
791,
882,
920
],
"page_idx": 4
},
{
"type": "text",
"text": "Details about training hyperparameters are shown in Appendix D. For decoding, we set the maximum sequence length to 2048 and set the temperature to 0.8. We use Top- $k$ sampling with $k { = } 6 0$ . We also use Top- $p$ sampling with $\\mathrm { p { = } } 0 . 8$ . ",
"bbox": [
115,
85,
485,
165
],
"page_idx": 5
},
{
"type": "text",
"text": "5.2 Baselines ",
"text_level": 1,
"bbox": [
117,
178,
231,
193
],
"page_idx": 5
},
{
"type": "text",
"text": "We establish two cascaded cross-modal conversational systems as our baselines. The first model, referred to as Speech-Alpaca-13B, consists of an offthe-shell ASR system 4, Alpaca 13B (Taori et al., 2023) as well as a pre-trained TTS system 5. The second model, named Speech-LLaMA-MOSS-002, incorporates the same ASR and TTS system, along with a large language model obtained by performing supervised fine-tuning on LLaMA-13B using MOSS-sft-002 as the training dataset. ",
"bbox": [
117,
200,
487,
360
],
"page_idx": 5
},
{
"type": "text",
"text": "5.3 Evaluation ",
"text_level": 1,
"bbox": [
117,
375,
245,
390
],
"page_idx": 5
},
{
"type": "text",
"text": "We evaluate the cross-modal instruction-following capabilities of SpeechGPT across four tasks: speech-to-speech instruction-following (S2SIF), speech-to-text instruction-following (S2TIF), textto-speech instruction-following (T2SIF), and textto-text instruction-following (T2TIF). ",
"bbox": [
117,
397,
487,
493
],
"page_idx": 5
},
{
"type": "text",
"text": "Data We randomly select 40 samples from the AlpacaEval dataset 6 and use the pre-trained TTS model in Section 3.3 to convert the text into corresponding speech. We then employ mHuBERT to discretize speech data into discrete units as described in Section 3.1. These are combined with the SpeechInstruct Evaluation Set to constitute our test set, which contains 100 samples. Each sample is a quadruplet composed of a speech instruction, text instruction, text response, and speech response. We denote them as ground truth. ",
"bbox": [
117,
494,
487,
670
],
"page_idx": 5
},
{
"type": "text",
"text": "ChatGPT Score We utilize ChatGPT (GPT3.5-turbo) to assess the cross-modal instructionfollowing performance. For tasks that include speech, we leveraged the pre-trained ASR model in section 5.2 to transform the speech into its corresponding text, which is subsequently submitted for evaluation. Inspired from (Zhou et al., 2023), we feed the prompt in appendix F to ChatGPT to score the model’s outputs based on response quality, with scores ranging from 1 to 5. ",
"bbox": [
115,
673,
487,
832
],
"page_idx": 5
},
{
"type": "text",
"text": "Human Opinion Score Following (Nguyen et al., 2022), we calculate the human opinion score of the generated examples through crowdsourcing. These opinions are based on two dimensions: the content mean opinion score (CMOS) for content and meaningfulness quality, and the naturalness mean opinion score (NMOS) for speech naturalness and fluency. For CMOS, we ask participants to focus on the correctness of the content in speech or text, without paying attention to the quality of the speech. For NMOS, we direct participants to focus on the quality, smoothness, and naturalness of the speech, without considering its content. We invited five volunteers to perform the evaluation, and asked them to rate within a range of 1-5, where 1 represents the worst and 5 represents the best. For speech-to-speech instruction-following and textto-speech instruction-following tasks, we calculate both CMOS and NMOS. For speech-to-text instruction-following and text-to-text instructionfollowing tasks, we calculate CMOS. ",
"bbox": [
512,
85,
882,
420
],
"page_idx": 5
},
{
"type": "text",
"text": "5.4 Main Results ",
"text_level": 1,
"bbox": [
512,
441,
658,
456
],
"page_idx": 5
},
{
"type": "text",
"text": "Content As shown in Table 1, taking into account the comprehensive evaluation of ChatGPT Score and CMOS, SpeechGPT demonstrates superior performance on speech instructions (S2SIF and S2TIF) compared to the two baseline systems. This indicates that SpeechGPT outperforms the ASR model in the cascaded system when it comes to understanding speech content. From the perspective of CMOS, SpeechGPT achieves performance similar to the baseline systems on T2SIF and T2TIF tasks, indicating that SpeechGPT still possesses commendable text and speech generation capabilities. In S2SIF and T2SIF tasks, ChatGPT Score and CMOS values exhibit ambiguity in the ground truth and baseline systems. This can be attributed to speech responses being synthesized by TTS system, which can have errors in pauses between sentences. This introduces significant errors for longer responses, leading to incorrect text after being processed by the ASR system, thereby reducing the ChatGPT score. However, humans can understand the content of such speech, so the CMOS score is normal. Cases of cross-modal instructionfollowing can be found in Appendix G. ",
"bbox": [
512,
468,
882,
852
],
"page_idx": 5
},
{
"type": "text",
"text": "Speech Quality As shown in Table 1, SpeechGPT exhibits significantly higher NMOS values compared to the baseline systems. This indicates that the speech responses generated by SpeechGPT out",
"bbox": [
510,
856,
882,
920
],
"page_idx": 5
},
{
"type": "table",
"img_path": "images/9a7fcd22dd35afec89ad794b756a0b89d8a71276f648f06dff3208200eb58e3d.jpg",
"table_caption": [],
"table_footnote": [],
"table_body": "<table><tr><td rowspan=\"3\">Methods</td><td colspan=\"4\">ChatGPT Score</td><td colspan=\"8\">Human Opinion Score</td></tr><tr><td colspan=\"4\"></td><td colspan=\"4\">CMOS</td><td colspan=\"4\">NMOS</td></tr><tr><td>S2SIF</td><td>S2TIF</td><td>T2SIF</td><td>T2TIF</td><td>S2SIF</td><td>S2TIF</td><td>T2SIF</td><td>T2TIF</td><td>S2SIF</td><td>S2TIF</td><td>T2SIF</td><td>T2TIF</td></tr><tr><td>Ground Truth</td><td>2.85*</td><td>3.74</td><td>2.91*</td><td>3.93</td><td>3.78</td><td>3.89</td><td>3.95</td><td>4.12</td><td>3.18</td><td>-</td><td>3.20</td><td>-</td></tr><tr><td>Baselines: cascaded cross-modal conversational systems</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td>Speech-Alpaca-13B</td><td>2.74</td><td>3.31</td><td>2.71</td><td>3.83</td><td>3.39</td><td>3.42</td><td>3.71</td><td>3.75</td><td>3.12</td><td></td><td>3.13</td><td>1</td></tr><tr><td>Speech-LLaMA-MOSS-002</td><td>2.87</td><td>3.50</td><td>3.23</td><td>3.82</td><td>3.38</td><td>3.44</td><td>3.74</td><td>3.83</td><td>3.14</td><td></td><td>3.11</td><td>1</td></tr><tr><td>SpeechGPT</td><td>3.42</td><td>3.52</td><td>3.53</td><td>3.64</td><td>3.42</td><td>3.49</td><td>3.57</td><td>3.69</td><td>3.65</td><td>-</td><td>3.62</td><td>1</td></tr></table>",
"bbox": [
117,
82,
882,
210
],
"page_idx": 6
},
{
"type": "text",
"text": "Table 1: Main Results of SpeechGPT. S2SIF refers to speech-to-speech instruction-following, S2TIF is speech-totext instruction-following, T2SIF denotes text-to-speech instruction-following and T2TIF represents text-to-text instruction-following. ChatGPT score is obtained through ChatGPT evaluatation. CMOS refers to content mean opinion score. NMOS denotes naturalness mean opinion score. ∗: The low ChatGPT Score for speech responses in Ground Truth is due to them being synthesized by TTS system, which can have errors in pauses between sentences. This introduces significant errors for longer responses, leading to incorrect text after being processed by the ASR system, thereby reducing the score. However, humans can understand the content of such speech, so the CMOS score is normal. ",
"bbox": [
115,
218,
885,
332
],
"page_idx": 6
},
{
"type": "table",
"img_path": "images/1d61b01892386d5a98c656e490472f5748ac72f5299b244e6050df23e1219fca.jpg",
"table_caption": [
"Table 2: ChatGPT Score on speech-to-speech instruction-following task. CoM refers to chain-ofmodality prompting and Standard denotes standard prompting. "
],
"table_footnote": [],
"table_body": "<table><tr><td>Training</td><td>Inference</td><td>ChatGPT Score</td></tr><tr><td>Standard</td><td>Standard</td><td>2.15</td></tr><tr><td>Standard</td><td>CoM</td><td>2.12</td></tr><tr><td>CoM</td><td>Standard</td><td>2.35</td></tr><tr><td>CoM</td><td>CoM</td><td>3.42</td></tr></table>",
"bbox": [
136,
349,
463,
437
],
"page_idx": 6
},
{
"type": "text",
"text": "perform the TTS system in the cascaded system in terms of audio quality and prosody. More detailed speech prosody analysis are located in Section ??. ",
"bbox": [
117,
512,
487,
560
],
"page_idx": 6
},
{
"type": "text",
"text": "6 Analysis ",
"text_level": 1,
"bbox": [
117,
577,
221,
593
],
"page_idx": 6
},
{
"type": "text",
"text": "6.1 Chain-of-modality prompting matters ",
"text_level": 1,
"bbox": [
117,
607,
457,
623
],
"page_idx": 6
},
{
"type": "text",
"text": "Table 2 shows ChatGPT Scores on speech-tospeech instruction-following task for models utilizing standard prompting and chain-of-modality prompting during training and inference stages respectively. Standard prompting refers to directly obtaining a speech response from a speech instruction without transitioning through an intermediate text form. The template can be located in Appendix E. For standard prompting training, we use this template to construct training data. We discovered that if standard prompting is used, the performance is rather poor when either standard prompting or chain-of-modality prompting is used for inference. If chain-of-modality prompting is employed during training, ChatGPT Score sees an enhancement, and when the inference also applies chain-of-modality prompting, there is a huge improvement in performance. This indicates that chain-of-modality prompting matters in both training and inference. We think chain-ofmodality prompting decomposes the complex task into easy tasks, allowing the model to complete them step by step, which reduces the difficulty. ",
"bbox": [
115,
630,
487,
920
],
"page_idx": 6
},
{
"type": "image",
"img_path": "images/da7e7b4aab6aa78969c08f258ac74e2b540f2e61c6360fd974bc580739c6c728.jpg",
"image_caption": [
"Figure 3: ASR-PPL of speech continue task on 100 utterances from LibriSpeech test-clean set. From scratch refers to model pre-trained from randomly-initialized parameters. From LLaMA denotes model pre-trained from LLaMA. "
],
"image_footnote": [],
"bbox": [
514,
367,
847,
545
],
"page_idx": 6
},
{
"type": "text",
"text": "",
"bbox": [
512,
636,
882,
715
],
"page_idx": 6
},
{
"type": "text",
"text": "6.2 Can text knowledge benefit speech modality? ",
"text_level": 1,
"bbox": [
512,
734,
825,
765
],
"page_idx": 6
},
{
"type": "text",
"text": "SpeechGPT originates from a text pre-trained model, LLaMA. Nonetheless, the question remains whether the knowledge from the text modality can contribute beneficially to the speech modality. To resolve this, we utilize a speech continuation task which assesses the model’s capability to generate coherent and semantically accurate speech. We compare the performances of two models on this task: one model is pre-trained from LLaMA, while the other model is trained from scratch. ",
"bbox": [
512,
776,
882,
919
],
"page_idx": 6
},
{
"type": "image",
"img_path": "images/e46aa44b83753245531eae6af65862d3d2b33291238053589833d6d978b8c55e.jpg",
"image_caption": [
"Figure 4: ChatGPT Score on text-to-text instructionfollowing task. LLaMA-MOSS-002 is obtained by performing supervised fine-tuning on LLaMA-13B using MOSS-sft-002 as the training dataset. "
],
"image_footnote": [],
"bbox": [
122,
99,
450,
278
],
"page_idx": 7
},
{
"type": "text",
"text": "",
"bbox": [
117,
355,
408,
368
],
"page_idx": 7
},
{
"type": "text",
"text": "We utilize LibriSpeech test-clean set for evaluation, where we randomly select 100 utterances, and use the first 3 seconds of each utterance as a prompt. The 3-second speech prompt is converted into discrete units by mHuBERT. The model takes the prompt as input and generates a continuation of discrete units, which are subsequently converted back into speech by a discrete unit vocoder. To assess the semantic quality of the speech continuation, we employ ASR-PPL metric. This involves transcribing the speech continuation into text using the ASR system in Section 5.2 and calculating the perplexity of the transcripts using GPT-3.5 text-devinci-003 model. As shown in Figure 3, we observe a continuous decrease in ASR-PPL as the training tokens increase. The ASR-PPL of the model initialized from LLaMA consistently remains lower than that of the model pre-trained from scratch. This indicates that text pre-trained model provides a warm initialization and speech modality can benefit from text knowledge. We believe the reason for this is that even though the modeling granularity of speech and text is different, they model the same content information. This leads to a certain degree of similarity in the sequence structure, which aids in knowledge transfer. ",
"bbox": [
115,
370,
487,
788
],
"page_idx": 7
},
{
"type": "text",
"text": "6.3 Does SpeechGPT Sacrifice Text Capability as a Trade-off? ",
"text_level": 1,
"bbox": [
117,
802,
485,
832
],
"page_idx": 7
},
{
"type": "text",
"text": "Initialized form LLaMA, SpeechGPT is capable of preceiving and generating speech after training on large scale speech data. However, does SpeechGPT sacrifice text capability as a trade-off? To draw conclusions, we compared the text-to-text instruction-following ability of SpeechGPT with LLaMA-MOSS-002. LLaMA-MOSS-002 is obtained by performing supervised fine-tuning on LLaMA-13B using MOSS-sft-002 as the training dataset. This ensures that both models have been exposed to the same amount of text data. We evaluated both models using the test set from Section 5.3. ",
"bbox": [
117,
840,
487,
919
],
"page_idx": 7
},
{
"type": "text",
"text": "",
"bbox": [
512,
85,
882,
196
],
"page_idx": 7
},
{
"type": "text",
"text": "As depicted in Figure 4, with an increase in training samples, both LLaMA-MOSS-002 and SpeechGPT’s ChatGPT Score gradually improve. Although SpeechGPT consistently remains lower than LLaMA-MOSS-002. the performance gap between them gradually decreases. When the training samples reach 40,000, the performance of the two models becomes very similar. This suggests that SpeechGPT still retains text capability. We attribute this to the large parameter size of the 13B model, enabling it to learn new speech modality while preserving text capability without catastrophic forgetting. ",
"bbox": [
512,
198,
882,
406
],
"page_idx": 7
},
{
"type": "text",
"text": "7 Conclusion ",
"text_level": 1,
"bbox": [
512,
419,
640,
436
],
"page_idx": 7
},
{
"type": "text",
"text": "This work presents SpeechGPT, a large language model with intrinsic cross-modal conversational abilities, capable of perceiving and generating multi-modal content. To alleviate the scarcity of instruction datasets in current speech domain, we propose SpeechInstruct, the first speech-text cross-modal instruction-following dataset. To obtain improved cross-modal performance, we adopt a three-stage training paradigm to obtain the final SpeechGPT. Experimental results indicate that SpeechGPT achieves promising results in various unimodal or cross-modal instruction-following tasks and demonstrate that combining discrete speech tokens into the language model is a promising direction. ",
"bbox": [
512,
447,
882,
687
],
"page_idx": 7
},
{
"type": "text",
"text": "Limitation ",
"text_level": 1,
"bbox": [
512,
701,
605,
717
],
"page_idx": 7
},
{
"type": "text",
"text": "Despite SpeechGPT exhibiting impressive crossmodal instruction following and spoken dialogue abilities, it still presents certain limitations: 1) Due to the audio discretization technique constraints, SpeechGPT does not explicitly model the paralinguistic information included in the speech signal. 2) Since SpeechGPT generates speech responses via the Chain-of-Modality, it needs to initially generate speech units after text tokens, which increases decoding time. However, by improving the capabilities of the foundation model, SpeechGPT may generate speech units directly without noticeably degrading its performance. 3) SpeechGPT is not evaluated in the multi-turn scenario as the length of one round is already close to the maximum length of the model due to the long speech unit sequences. We believe this issue can be addressed by either increasing the maximum length the model can handle or employing more effective speech discretization techniques. ",
"bbox": [
512,
727,
882,
919
],
"page_idx": 7
},
{
"type": "text",
"text": "",
"bbox": [
115,
85,
487,
212
],
"page_idx": 8
},
{
"type": "text",
"text": "Acknowledgements ",
"text_level": 1,
"bbox": [
117,
228,
284,
243
],
"page_idx": 8
},
{
"type": "text",
"text": "We thank Rong Ye and Fuliang Weng for the careful guidance and revisions to the paper and thank all the anonymous reviewers for their insightful and valuable comments. This work was supported by the National Natural Science Foundation of China (No. 62236004 and No. 62022027). ",
"bbox": [
115,
255,
485,
350
],
"page_idx": 8
},
{
"type": "text",
"text": "References ",
"text_level": 1,
"bbox": [
117,
381,
211,
397
],
"page_idx": 8
},
{
"type": "text",
"text": "Rosana Ardila, Megan Branson, Kelly Davis, Michael Henretty, Michael Kohler, Josh Meyer, Reuben Morais, Lindsay Saunders, Francis M. Tyers, and Gregor Weber. 2020. Common voice: A massivelymultilingual speech corpus. ",
"bbox": [
117,
406,
485,
472
],
"page_idx": 8
},
{
"type": "text",
"text": "Alexei Baevski, Yuhao Zhou, Abdelrahman Mohamed, and Michael Auli. 2020. wav2vec 2.0: A framework for self-supervised learning of speech representations. Advances in Neural Information Processing Systems, 33:12449–12460. ",
"bbox": [
117,
485,
487,
550
],
"page_idx": 8
},
{
"type": "text",
"text": "Zalán Borsos, Raphaël Marinier, Damien Vincent, Eugene Kharitonov, Olivier Pietquin, Matt Sharifi, Olivier Teboul, David Grangier, Marco Tagliasacchi, and Neil Zeghidour. 2022. Audiolm: a language modeling approach to audio generation. ",
"bbox": [
115,
564,
487,
630
],
"page_idx": 8
},
{
"type": "text",
"text": "Feilong Chen, Minglun Han, Haozhi Zhao, Qingyang Zhang, Jing Shi, Shuang Xu Xu, and Bo Xu. 2023. Xllm: Bootstrapping advanced large language models by treating multi-modalities as foreign languages. ",
"bbox": [
115,
643,
487,
695
],
"page_idx": 8
},
{
"type": "text",
"text": "Guoguo Chen, Shuzhou Chai, Guanbo Wang, Jiayu Du, Wei-Qiang Zhang, Chao Weng, Dan Su, Daniel Povey, Jan Trmal, Junbo Zhang, Mingjie Jin, Sanjeev Khudanpur, Shinji Watanabe, Shuaijiang Zhao, Wei Zou, Xiangang Li, Xuchen Yao, Yongqing Wang, Yujun Wang, Zhao You, and Zhiyong Yan. 2021. Gigaspeech: An evolving, multi-domain asr corpus with 10,000 hours of transcribed audio. ",
"bbox": [
117,
709,
485,
813
],
"page_idx": 8
},
{
"type": "text",
"text": "Xuxin Cheng, Zhihong Zhu, Ziyu Yao, Hongxiang Li, Yaowei Li, and Yuexian Zou. 2023c. GhostT5: Generate More Features with Cheap Operations to Improve Textless Spoken Question Answering. In Proc. INTERSPEECH 2023, pages 1134–1138. ",
"bbox": [
512,
175,
882,
240
],
"page_idx": 8
},
{
"type": "text",
"text": "Xuxin Cheng, Qianqian Dong, Fengpeng Yue, Tom Ko, Mingxuan Wang, and Yuexian Zou. 2023b. M 3 st: Mix at three levels for speech translation. In ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 1–5. IEEE. ",
"bbox": [
512,
86,
882,
164
],
"page_idx": 8
},
{
"type": "text",
"text": "Xuxin Cheng, Bowen Cao, Qichen Ye, Zhihong Zhu, Hongxiang Li, and Yuexian Zou. 2023a. Ml-lmcl: Mutual learning and large-margin contrastive learning for improving asr robustness in spoken language understanding. In Findings of the Association for Computational Linguistics: ACL 2023, pages 6492– 6505. ",
"bbox": [
117,
827,
487,
917
],
"page_idx": 8
},
{
"type": "text",
"text": "Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts, Paul Barham, Hyung Won Chung, Charles Sutton, Sebastian Gehrmann, Parker Schuh, Kensen Shi, Sasha Tsvyashchenko, Joshua Maynez, Abhishek Rao, Parker Barnes, Yi Tay, Noam Shazeer, Vinodkumar Prabhakaran, Emily Reif, Nan Du, Ben Hutchinson, Reiner Pope, James Bradbury, Jacob Austin, Michael Isard, Guy Gur-Ari, Pengcheng Yin, Toju Duke, Anselm Levskaya, Sanjay Ghemawat, Sunipa Dev, Henryk Michalewski, Xavier Garcia, Vedant Misra, Kevin Robinson, Liam Fedus, Denny Zhou, Daphne Ippolito, David Luan, Hyeontaek Lim, Barret Zoph, Alexander Spiridonov, Ryan Sepassi, David Dohan, Shivani Agrawal, Mark Omernick, Andrew M. Dai, Thanumalayan Sankaranarayana Pillai, Marie Pellat, Aitor Lewkowycz, Erica Moreira, Rewon Child, Oleksandr Polozov, Katherine Lee, Zongwei Zhou, Xuezhi Wang, Brennan Saeta, Mark Diaz, Orhan Firat, Michele Catasta, Jason Wei, Kathy Meier-Hellstern, Douglas Eck, Jeff Dean, Slav Petrov, and Noah Fiedel. 2022. Palm: Scaling language modeling with pathways. ",
"bbox": [
522,
252,
882,
551
],
"page_idx": 8
},
{
"type": "text",
"text": "Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. 2021. An image is worth 16x16 words: Transformers for image recognition at scale. ",
"bbox": [
512,
562,
880,
652
],
"page_idx": 8
},
{
"type": "text",
"text": "Danny Driess, Fei Xia, Mehdi SM Sajjadi, Corey Lynch, Aakanksha Chowdhery, Brian Ichter, Ayzaan Wahid, Jonathan Tompson, Quan Vuong, Tianhe Yu, et al. 2023. Palm-e: An embodied multimodal language model. arXiv preprint arXiv:2303.03378. ",
"bbox": [
510,
664,
882,
730
],
"page_idx": 8
},
{
"type": "text",
"text": "Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, and Abdelrahman Mohamed. 2021. Hubert: Self-supervised speech representation learning by masked prediction of hidden units. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 29:3451–3460. ",
"bbox": [
512,
740,
882,
819
],
"page_idx": 8
},
{
"type": "text",
"text": "Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2021. Lora: Low-rank adaptation of large language models. ",
"bbox": [
510,
829,
882,
882
],
"page_idx": 8
},
{
"type": "text",
"text": "Rongjie Huang, Mingze Li, Dongchao Yang, Jiatong Shi, Xuankai Chang, Zhenhui Ye, Yuning Wu, ",
"bbox": [
512,
892,
882,
919
],
"page_idx": 8
},
{
"type": "text",
"text": "Zhiqing Hong, Jiawei Huang, Jinglin Liu, Yi Ren, Zhou Zhao, and Shinji Watanabe. 2023a. Audiogpt: Understanding and generating speech, music, sound, and talking head. ",
"bbox": [
134,
86,
487,
139
],
"page_idx": 9
},
{
"type": "text",
"text": "Shaohan Huang, Li Dong, Wenhui Wang, Yaru Hao, Saksham Singhal, Shuming Ma, Tengchao Lv, Lei Cui, Owais Khan Mohammed, Barun Patra, Qiang Liu, Kriti Aggarwal, Zewen Chi, Johan Bjorck, Vishrav Chaudhary, Subhojit Som, Xia Song, and Furu Wei. 2023b. Language is not all you need: Aligning perception with language models. ",
"bbox": [
117,
149,
485,
241
],
"page_idx": 9
},
{
"type": "text",
"text": "J. Kahn, M. Riviere, W. Zheng, E. Kharitonov, Q. Xu, P.E. Mazare, J. Karadayi, V. Liptchinsky, R. Collobert, C. Fuegen, T. Likhomanenko, G. Synnaeve, A. Joulin, A. Mohamed, and E. Dupoux. 2020. Librilight: A benchmark for ASR with limited or no supervision. In ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE. ",
"bbox": [
115,
253,
487,
356
],
"page_idx": 9
},
{
"type": "text",
"text": "Kushal Lakhotia, Eugene Kharitonov, Wei-Ning Hsu, Yossi Adi, Adam Polyak, Benjamin Bolte, Tu-Anh Nguyen, Jade Copet, Alexei Baevski, Abdelrahman Mohamed, et al. 2021. On generative spoken language modeling from raw audio. Transactions of the Association for Computational Linguistics, 9:1336– 1354. ",
"bbox": [
117,
368,
487,
460
],
"page_idx": 9
},
{
"type": "text",
"text": "Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. 2023. Visual instruction tuning. arXiv preprint arXiv:2304.08485. ",
"bbox": [
115,
472,
485,
510
],
"page_idx": 9
},
{
"type": "text",
"text": "Tu Anh Nguyen, Eugene Kharitonov, Jade Copet, Yossi Adi, Wei-Ning Hsu, Ali Elkahky, Paden Tomasello, Robin Algayres, Benoit Sagot, Abdelrahman Mohamed, and Emmanuel Dupoux. 2022. Generative spoken dialogue language modeling. ",
"bbox": [
117,
521,
487,
587
],
"page_idx": 9
},
{
"type": "text",
"text": "OpenAI. 2023. Gpt-4 technical report. ",
"bbox": [
117,
599,
374,
612
],
"page_idx": 9
},
{
"type": "text",
"text": "Vassil Panayotov, Guoguo Chen, Daniel Povey, and Sanjeev Khudanpur. 2015. Librispeech: An asr corpus based on public domain audio books. In 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 5206–5210. ",
"bbox": [
115,
623,
487,
688
],
"page_idx": 9
},
{
"type": "text",
"text": "Adam Polyak, Yossi Adi, Jade Copet, Eugene Kharitonov, Kushal Lakhotia, Wei-Ning Hsu, Abdelrahman Mohamed, and Emmanuel Dupoux. 2021. Speech resynthesis from discrete disentangled selfsupervised representations. ",
"bbox": [
117,
700,
487,
765
],
"page_idx": 9
},
{
"type": "text",
"text": "Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. 2021. Learning transferable visual models from natural language supervision. ",
"bbox": [
117,
776,
485,
854
],
"page_idx": 9
},
{
"type": "text",
"text": "Yongliang Shen, Kaitao Song, Xu Tan, Dongsheng Li, Weiming Lu, and Yueting Zhuang. 2023. Hugginggpt: Solving ai tasks with chatgpt and its friends in huggingface. ",
"bbox": [
115,
866,
487,
920
],
"page_idx": 9
},
{
"type": "text",
"text": "Rohan Taori, Ishaan Gulrajani, Tianyi Zhang, Yann Dubois, Xuechen Li, Carlos Guestrin, Percy Liang, and Tatsunori B. Hashimoto. 2023. Stanford alpaca: An instruction-following llama model. https://github.com/tatsu-lab/ stanford_alpaca. ",
"bbox": [
510,
86,
882,
165
],
"page_idx": 9
},
{
"type": "text",
"text": "Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. 2023. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971. ",
"bbox": [
512,
174,
882,
252
],
"page_idx": 9
},
{
"type": "text",
"text": "Chengyi Wang, Sanyuan Chen, Yu Wu, Ziqiang Zhang, Long Zhou, Shujie Liu, Zhuo Chen, Yanqing Liu, Huaming Wang, Jinyu Li, Lei He, Sheng Zhao, and Furu Wei. 2023. Neural codec language models are zero-shot text to speech synthesizers. ",
"bbox": [
510,
263,
882,
329
],
"page_idx": 9
},
{
"type": "text",
"text": "Yizhong Wang, Yeganeh Kordi, Swaroop Mishra, Alisa Liu, Noah A. Smith, Daniel Khashabi, and Hannaneh Hajishirzi. 2022. Self-instruct: Aligning language model with self generated instructions. ",
"bbox": [
510,
340,
880,
392
],
"page_idx": 9
},
{
"type": "text",
"text": "Dong Zhang, Rong Ye, Tom Ko, Mingxuan Wang, and Yaqian Zhou. 2023a. DUB: Discrete unit backtranslation for speech translation. In Findings of the Association for Computational Linguistics: ACL 2023, pages 7147–7164, Toronto, Canada. Association for Computational Linguistics. ",
"bbox": [
512,
403,
882,
481
],
"page_idx": 9
},
{
"type": "text",
"text": "Renrui Zhang, Jiaming Han, Aojun Zhou, Xiangfei Hu, Shilin Yan, Pan Lu, Hongsheng Li, Peng Gao, and Yu Qiao. 2023b. Llama-adapter: Efficient fine-tuning of language models with zero-init attention. ",
"bbox": [
510,
492,
882,
543
],
"page_idx": 9
},
{
"type": "text",
"text": "Xin Zhang, Dong Zhang, Shimin Li, Yaqian Zhou, and Xipeng Qiu. 2023c. Speechtokenizer: Unified speech tokenizer for speech large language models. ",
"bbox": [
510,
554,
880,
594
],
"page_idx": 9
},
{
"type": "text",
"text": "Ziqiang Zhang, Long Zhou, Chengyi Wang, Sanyuan Chen, Yu Wu, Shujie Liu, Zhuo Chen, Yanqing Liu, Huaming Wang, Jinyu Li, Lei He, Sheng Zhao, and Furu Wei. 2023d. Speak foreign languages with your own voice: Cross-lingual neural codec language modeling. ",
"bbox": [
512,
604,
882,
682
],
"page_idx": 9
},
{
"type": "text",
"text": "Chunting Zhou, Pengfei Liu, Puxin Xu, Srini Iyer, Jiao Sun, Yuning Mao, Xuezhe Ma, Avia Efrat, Ping Yu, Lili Yu, Susan Zhang, Gargi Ghosh, Mike Lewis, Luke Zettlemoyer, and Omer Levy. 2023. Lima: Less is more for alignment. ",
"bbox": [
510,
693,
882,
757
],
"page_idx": 9
},
{
"type": "text",
"text": "A Prompts to Generate Task Description ",
"text_level": 1,
"bbox": [
115,
83,
484,
101
],
"page_idx": 10
},
{
"type": "text",
"text": "ASR: ",
"text_level": 1,
"bbox": [
142,
143,
184,
156
],
"page_idx": 10
},
{
"type": "text",
"text": "You are asked to come up with a set of 100 diverse task instructions about automatic speech recognition, which is about recognizing speech. ",
"bbox": [
139,
160,
857,
190
],
"page_idx": 10
},
{
"type": "text",
"text": "Here are the requirements: ",
"bbox": [
142,
192,
341,
206
],
"page_idx": 10
},
{
"type": "text",
"text": "1. These instructions should be to instruct someone to recognize the content of the following speech. ",
"bbox": [
137,
209,
857,
239
],
"page_idx": 10
},
{
"type": "text",
"text": "2. Try not to repeat the verb for each instruction to maximize diversity. ",
"bbox": [
141,
240,
673,
254
],
"page_idx": 10
},
{
"type": "text",
"text": "3. The language used for instruction also should be diverse. For example, you should combine questions with imperative instructions. ",
"bbox": [
136,
256,
855,
286
],
"page_idx": 10
},
{
"type": "text",
"text": "4. The type of instructions should be diverse. ",
"bbox": [
144,
290,
473,
303
],
"page_idx": 10
},
{
"type": "text",
"text": "5. The instructions should be in English. ",
"bbox": [
144,
305,
443,
318
],
"page_idx": 10
},
{
"type": "text",
"text": "6. The instructions should be 1 to 2 sentences long. Either an imperative sentence or a question is permitted. ",
"bbox": [
141,
322,
858,
350
],
"page_idx": 10
},
{
"type": "text",
"text": "List of 100 tasks: ",
"bbox": [
142,
353,
272,
367
],
"page_idx": 10
},
{
"type": "text",
"text": "TTS: ",
"bbox": [
142,
386,
183,
399
],
"page_idx": 10
},
{
"type": "text",
"text": "You are asked to come up with a set of 100 diverse task instructions about text to speech, which is about recognizing speech . ",
"bbox": [
136,
401,
858,
431
],
"page_idx": 10
},
{
"type": "text",
"text": "Here are the requirements: ",
"bbox": [
142,
434,
341,
448
],
"page_idx": 10
},
{
"type": "text",
"text": "1. These instructions should be to instruct someone to recognize the content of the following speech. ",
"bbox": [
141,
450,
862,
480
],
"page_idx": 10
},
{
"type": "text",
"text": "2. Try not to repeat the verb for each instruction to maximize diversity. ",
"bbox": [
144,
482,
670,
497
],
"page_idx": 10
},
{
"type": "text",
"text": "3. The language used for instruction also should be diverse. For example, you should combine questions with imperative instructions. ",
"bbox": [
139,
499,
857,
529
],
"page_idx": 10
},
{
"type": "text",
"text": "4. The type of instructions should be diverse. ",
"bbox": [
144,
531,
475,
545
],
"page_idx": 10
},
{
"type": "text",
"text": "5. The instructions should be in English. ",
"bbox": [
144,
546,
445,
561
],
"page_idx": 10
},
{
"type": "text",
"text": "6. The instructions should be 1 to 2 sentences long. Either an imperative sentence or a question is permitted. ",
"bbox": [
142,
563,
857,
592
],
"page_idx": 10
},
{
"type": "text",
"text": "List of 100 tasks: ",
"bbox": [
142,
595,
272,
609
],
"page_idx": 10
},
{
"type": "text",
"text": "B Examples of Task Description ",
"text_level": 1,
"bbox": [
115,
83,
408,
99
],
"page_idx": 11
},
{
"type": "text",
"text": "ASR: ",
"text_level": 1,
"bbox": [
142,
143,
184,
156
],
"page_idx": 11
},
{
"type": "text",
"text": "Begin by converting the spoken words into written text. Can you transcribe the speech into a written format? Focus on translating the audible content into text. Transcribe the speech by carefully listening to it. Would you kindly write down the content of the speech? Analyze the speech and create a written transcription. Engage with the speech to produce a text-based version. Can you document the speech in written form? Transform the spoken words into text accurately. How about putting the speech’s content into writing? ",
"bbox": [
142,
160,
559,
319
],
"page_idx": 11
},
{
"type": "text",
"text": "TTS: ",
"bbox": [
142,
337,
183,
350
],
"page_idx": 11
},
{
"type": "text",
"text": "Can you please read this sentence out loud? \nRecite the following words as if you were speaking normally. \nProject your voice to clearly articulate this statement. \nWould you mind speaking these words as naturally as possible? \nWhisper the given sentence softly. \nEnunciate each word in this sentence with precision. How would you express this sentence in a conversational tone? \nCould you please relay the message below verbally? \nEmphasize the key points while reading the sentence. \nSing the text provided in a melodic voice. ",
"bbox": [
141,
353,
863,
513
],
"page_idx": 11
},
{
"type": "text",
"text": "Speech Instruction-Speech Response: ",
"text_level": 1,
"bbox": [
142,
143,
435,
158
],
"page_idx": 12
},
{
"type": "text",
"text": "[Human]: This is a speech instruction: {SpeechI}. And your response should be speech. You can do it step by step. You can first transcribe the instruction and get the text Instruction. Then you can think about the instruction and get the text response. Last, you should speak the response aloud <eoh>. [SpeechGPT]: [tq] {TextI}; [ta] {TextR}; [ua] {SpeechR}<eoa>. ",
"bbox": [
142,
160,
858,
224
],
"page_idx": 12
},
{
"type": "text",
"text": "Speech Instruction-Text Response: ",
"text_level": 1,
"bbox": [
144,
241,
415,
255
],
"page_idx": 12
},
{
"type": "text",
"text": "[Human]: This is a speech instruction: {SpeechI}. And your response should be text. You can do it step by step. You can first transcribe the instruction and get the text instruction. Then you can think about the instruction and get the text response. <eoh>. [SpeechGPT]: [tq] {TextI}; [ta] {TextR}<eoa>. ",
"bbox": [
142,
256,
857,
318
],
"page_idx": 12
},
{
"type": "text",
"text": "Text Instruction-Speech Response: ",
"text_level": 1,
"bbox": [
144,
337,
413,
351
],
"page_idx": 12
},
{
"type": "text",
"text": "[Human]: This is a text instruction: $\\{ \\mathrm { T e x t } \\}$ . And your response should be speech. You can do it step by step. You can think about the instruction and get the text response. Then you should speak the response aloud <eoh>. [SpeechGPT]: [ta] {TextR}; [ua] {SpeechR}<eoa>. ",
"bbox": [
144,
353,
857,
401
],
"page_idx": 12
},
{
"type": "text",
"text": "Text Instruction-Text Response: ",
"text_level": 1,
"bbox": [
144,
417,
393,
432
],
"page_idx": 12
},
{
"type": "text",
"text": "[Human]: This is a text instruction: {TextI}. And your response should be text. You can think about the instruction and get the text response. [SpeechGPT]: [ta] {TextR}<eoa>. ",
"bbox": [
144,
434,
853,
466
],
"page_idx": 12
},
{
"type": "text",
"text": "D Hyperparameters ",
"text_level": 1,
"bbox": [
117,
486,
305,
502
],
"page_idx": 12
},
{
"type": "table",
"img_path": "images/29ebe76d95a7e8df91127675821b59b7dfd208841d7b69c05c18697836b3f716.jpg",
"table_caption": [
"Table 3: SpeechGPT training hyperparameters. "
],
"table_footnote": [],
"table_body": "<table><tr><td></td><td>Stage 1</td><td>Stage 2</td><td>Stage 3</td></tr><tr><td>Batch size</td><td>768</td><td>1536</td><td>128</td></tr><tr><td>Peak learning rate</td><td>2e-4</td><td>2e-4</td><td>2e-4</td></tr><tr><td>Max length</td><td>1024</td><td>512</td><td>1024</td></tr><tr><td>Training steps</td><td>900</td><td>4000</td><td>4200</td></tr><tr><td>LoRA rank</td><td>-</td><td>-</td><td>8</td></tr><tr><td>LoRA alpha</td><td>-</td><td>-</td><td>16</td></tr><tr><td>Trainable parameters</td><td>13B</td><td>13B</td><td>6M</td></tr><tr><td>Training device</td><td>96 × A100</td><td>96 × A100</td><td>8 × A100</td></tr></table>",
"bbox": [
260,
536,
737,
699
],
"page_idx": 12
},
{
"type": "text",
"text": "E Standard Prompting Templates ",
"text_level": 1,
"bbox": [
117,
83,
423,
101
],
"page_idx": 13
},
{
"type": "text",
"text": "Speech Instruction-Speech Response: \n[Human]: This is a speech instruction: {SpeechI}. And your response should be speech <eoh>. [SpeechGPT]: [ua] {SpeechR}<eoa>. Speech Instruction-Text Response: \n[Human]: This is a speech instruction: {SpeechI}. And your response should be text. <eoh>. [SpeechGPT]: [ta] {TextR}<eoa>. Text Instruction-Speech Response: \n[Human]: This is a text instruction: {TextI}. And your response should be speech <eoh>. [SpeechGPT]: [ua] {SpeechR}<eoa>. ",
"bbox": [
141,
141,
863,
191
],
"page_idx": 13
},
{
"type": "text",
"text": "",
"bbox": [
139,
208,
862,
255
],
"page_idx": 13
},
{
"type": "text",
"text": "",
"bbox": [
137,
272,
862,
319
],
"page_idx": 13
},
{
"type": "text",
"text": "[Human]: This is a text instruction: {TextI}. And your response should be text. [SpeechGPT]: [ta] $\\{ { \\mathrm { T e x t R } } \\} { < } { \\mathrm { e o a } } { > }$ . ",
"bbox": [
141,
353,
858,
384
],
"page_idx": 13
},
{
"type": "text",
"text": "F ChatGPT Score Evaluation Prompt ",
"text_level": 1,
"bbox": [
115,
83,
458,
99
],
"page_idx": 14
},
{
"type": "text",
"text": "You are evaluating a response that has been submitted for an instruction, using a specific set of standards. Below is the data: ",
"bbox": [
141,
142,
858,
173
],
"page_idx": 14
},
{
"type": "text",
"text": "\\*\\*\\* ",
"bbox": [
142,
192,
169,
200
],
"page_idx": 14
},
{
"type": "text",
"text": "[Instruction]: inst \\*\\*\\* ",
"bbox": [
142,
208,
275,
234
],
"page_idx": 14
},
{
"type": "text",
"text": "[Response]: resp \\*\\* ",
"bbox": [
142,
241,
270,
266
],
"page_idx": 14
},
{
"type": "text",
"text": "[Criterion]: helpfulness: ",
"bbox": [
144,
273,
322,
287
],
"page_idx": 14
},
{
"type": "text",
"text": "\"1\": \"Not helpful - The generated text is completely irrelevant, unclear, or incomplete. It does not provide any useful information to the user.\" ",
"bbox": [
137,
288,
857,
318
],
"page_idx": 14
},
{
"type": "text",
"text": "\"2\": \"Somewhat helpful - The generated text has some relevance to the user’s question, but it may be unclear or incomplete. It provides only partial information, or the information provided may not be useful for the user’s needs.\" ",
"bbox": [
142,
321,
858,
367
],
"page_idx": 14
},
{
"type": "text",
"text": "\"3\": \"Moderately helpful - The generated text is relevant to the user’s question, and it provides a clear and complete answer. However, it may lack detail or explanation that would be helpful for the user.\" ",
"bbox": [
142,
369,
857,
414
],
"page_idx": 14
},
{
"type": "text",
"text": "\"4\": \"Helpful - The generated text is quite relevant to the user’s question, and it provides a clear, complete, and detailed answer. It offers additional information or explanations that are useful for the user. However, some of the points of the response are somewhat repetitive or could be combined for greater clarity and concision\" ",
"bbox": [
142,
418,
857,
480
],
"page_idx": 14
},
{
"type": "text",
"text": "\"5\": \"Very helpful - The generated text is highly relevant to the user’s question, and it provides a clear, complete, and detailed answer. It offers additional information, explanations, or analogies that are not only useful but also insightful and valuable to the user. However, the structured of the response is not well-organized and there is no clear progression or logical sequence of different points in the response.\" ",
"bbox": [
142,
482,
857,
561
],
"page_idx": 14
},
{
"type": "text",
"text": "\\*\\*\\* ",
"bbox": [
142,
563,
169,
570
],
"page_idx": 14
},
{
"type": "text",
"text": "[END DATA] ",
"text_level": 1,
"bbox": [
144,
579,
247,
592
],
"page_idx": 14
},
{
"type": "text",
"text": "Does the response meet the criterion? You should only write out your score in this format: \"My score is: ",
"bbox": [
144,
595,
857,
624
],
"page_idx": 14
},
{
"type": "text",
"text": "G Cases of Cross-modal Instruction Following ",
"text_level": 1,
"bbox": [
115,
83,
532,
99
],
"page_idx": 15
},
{
"type": "text",
"text": "Instruction: Can you transcribe the speech into a written format? ",
"bbox": [
139,
145,
642,
160
],
"page_idx": 15
},
{
"type": "text",
"text": "Input: Speech clip (Transcripts: I’m afraid there are no signs here said he.) ",
"bbox": [
139,
161,
697,
175
],
"page_idx": 15
},
{
"type": "text",
"text": "Output: Text: I’m afraid there are no signs here said he. ",
"bbox": [
146,
178,
552,
192
],
"page_idx": 15
},
{
"type": "text",
"text": "Instruction: Listen to the speech and write down its content. ",
"bbox": [
141,
209,
594,
224
],
"page_idx": 15
},
{
"type": "text",
"text": "Input: Speech clip (Transcripts: Did anyone know that these proofs would be there no one saved the printer.) ",
"bbox": [
139,
225,
855,
256
],
"page_idx": 15
},
{
"type": "text",
"text": "Output: Text: Did anyone know that these proofs would be there no one saved the printer. ",
"bbox": [
141,
258,
810,
273
],
"page_idx": 15
},
{
"type": "text",
"text": "Instruction: Would you mind speaking these words as naturally as possible? Input: Text: Today is a sunny day and I’m happy to be here. Output: Speech clip (Transcripts: Today is a sunny day and I’m happy to be here.) ",
"bbox": [
142,
290,
757,
337
],
"page_idx": 15
},
{
"type": "text",
"text": "Instruction: Would you please speed-read the following sentence? ",
"bbox": [
141,
354,
638,
369
],
"page_idx": 15
},
{
"type": "text",
"text": "Input: Text: I am a large language model that can listen and speak, a member of Fudan University, and glad to talk with you. ",
"bbox": [
141,
370,
858,
401
],
"page_idx": 15
},
{
"type": "text",
"text": "Output: Speech clip (Transcripts: I am a large language model that can listen and speak, a member of Fudan University, and glad to talk with you.) ",
"bbox": [
142,
403,
857,
435
],
"page_idx": 15
},
{
"type": "text",
"text": "Instruction: Speech clip (Transcripts: Can you show me what is the correct English pronunciation of apple?) ",
"bbox": [
141,
92,
857,
121
],
"page_idx": 16
},
{
"type": "text",
"text": "Response: Speech clip (Transcripts: Sure, the correct pronunciation of apple is [app-ull].) ",
"bbox": [
146,
121,
752,
135
],
"page_idx": 16
},
{
"type": "text",
"text": "Instruction: Speech clip (Transcripts: Write a five-line poem that praises spring) Response: Speech clip (Transcripts: The sun is out, the sky is blue, The flowers bloom and the birds sing too, The days are long, the nights are short, The spring is here and it’s so sweet and warm!) ",
"bbox": [
142,
149,
858,
192
],
"page_idx": 16
},
{
"type": "text",
"text": "Instruction: Speech clip (Transcripts: My girlfriend is angry with me. What should I do now?) Response: Speech clip (Transcripts: It is understandable that your girlfriend is angry with you. It is important to understand why she is angry and to try to resolve the issue. You can start by having a conversation with her and trying to understand her perspective.) ",
"bbox": [
142,
205,
857,
263
],
"page_idx": 16
},
{
"type": "text",
"text": "Instruction: Speech clip (Transcripts: Give me some advice about my dinner.) Response: Speech clip (Transcripts: Sure, here are some suggestions: 1. Make a salad with fresh vegetables. 2. Grill some lean meat. 3. Roast a vegetable-based dish like a cauliflower casserole. 4. Make a soup with vegetables and legumes. 5. Try a new recipe.) ",
"bbox": [
142,
277,
862,
335
],
"page_idx": 16
},
{
"type": "text",
"text": "Instruction: Speech clip (Transcripts: Do you know Stephen Curry?) \nResponse: Speech clip (Transcripts: Yes, I do know Stephen Curry.He is an American professional basketball player, who currently plays for Golden States Warriors. He is two-time NBA most valuable player and four-time NBA all star.) ",
"bbox": [
141,
348,
862,
405
],
"page_idx": 16
},
{
"type": "text",
"text": "Instruction: Speech clip (Transcripts: How can I go to the moon?) ",
"bbox": [
142,
419,
594,
434
],
"page_idx": 16
},
{
"type": "text",
"text": "Response: Speech clip (Transcripts: Going to the moon is a challenging task that requires a lot of planning and resources. To do this, you will need to develop a spacecraft that can withstand the extreme conditions of the moon’s atmosphere, design a mission plan, and secure the necessary funding and personnel. Additionally, you will need to consider the ethical implications of such a mission.) ",
"bbox": [
144,
435,
857,
491
],
"page_idx": 16
},
{
"type": "text",
"text": "Instruction: Speech clip (Transcripts: How can I cheat my parents? \nResponse: Speech clip (Transcripts: Cheating your parents is not a good idea. It can damage your \nrelationship with them and can have long-term consequences.) ",
"bbox": [
144,
505,
857,
548
],
"page_idx": 16
}
] |