Upload svd_llm2/heal/llama2-7b-base/ranks.json with huggingface_hub

svd_llm2/heal/llama2-7b-base/ranks.json

@@ -0,0 +1,226 @@
+{
+ "model.layers.0.self_attn.q_proj": 2048,
+ "model.layers.0.self_attn.k_proj": 2048,
+ "model.layers.0.self_attn.v_proj": 2048,
+ "model.layers.0.self_attn.o_proj": 2048,
+ "model.layers.0.mlp.gate_proj": 2330,
+ "model.layers.0.mlp.up_proj": 2321,
+ "model.layers.0.mlp.down_proj": 2909,
+ "model.layers.1.self_attn.q_proj": 1908,
+ "model.layers.1.self_attn.k_proj": 1957,
+ "model.layers.1.self_attn.v_proj": 1823,
+ "model.layers.1.self_attn.o_proj": 2048,
+ "model.layers.1.mlp.gate_proj": 2104,
+ "model.layers.1.mlp.up_proj": 2083,
+ "model.layers.1.mlp.down_proj": 2275,
+ "model.layers.2.self_attn.q_proj": 1399,
+ "model.layers.2.self_attn.k_proj": 1400,
+ "model.layers.2.self_attn.v_proj": 1395,
+ "model.layers.2.self_attn.o_proj": 1693,
+ "model.layers.2.mlp.gate_proj": 1923,
+ "model.layers.2.mlp.up_proj": 1920,
+ "model.layers.2.mlp.down_proj": 2102,
+ "model.layers.3.self_attn.q_proj": 1249,
+ "model.layers.3.self_attn.k_proj": 1245,
+ "model.layers.3.self_attn.v_proj": 1283,
+ "model.layers.3.self_attn.o_proj": 1609,
+ "model.layers.3.mlp.gate_proj": 1818,
+ "model.layers.3.mlp.up_proj": 1823,
+ "model.layers.3.mlp.down_proj": 2007,
+ "model.layers.4.self_attn.q_proj": 1255,
+ "model.layers.4.self_attn.k_proj": 1252,
+ "model.layers.4.self_attn.v_proj": 1283,
+ "model.layers.4.self_attn.o_proj": 1493,
+ "model.layers.4.mlp.gate_proj": 1780,
+ "model.layers.4.mlp.up_proj": 1789,
+ "model.layers.4.mlp.down_proj": 1917,
+ "model.layers.5.self_attn.q_proj": 1239,
+ "model.layers.5.self_attn.k_proj": 1236,
+ "model.layers.5.self_attn.v_proj": 1266,
+ "model.layers.5.self_attn.o_proj": 1448,
+ "model.layers.5.mlp.gate_proj": 1758,
+ "model.layers.5.mlp.up_proj": 1766,
+ "model.layers.5.mlp.down_proj": 1860,
+ "model.layers.6.self_attn.q_proj": 1218,
+ "model.layers.6.self_attn.k_proj": 1214,
+ "model.layers.6.self_attn.v_proj": 1244,
+ "model.layers.6.self_attn.o_proj": 1398,
+ "model.layers.6.mlp.gate_proj": 1734,
+ "model.layers.6.mlp.up_proj": 1744,
+ "model.layers.6.mlp.down_proj": 1810,
+ "model.layers.7.self_attn.q_proj": 1209,
+ "model.layers.7.self_attn.k_proj": 1208,
+ "model.layers.7.self_attn.v_proj": 1237,
+ "model.layers.7.self_attn.o_proj": 1353,
+ "model.layers.7.mlp.gate_proj": 1720,
+ "model.layers.7.mlp.up_proj": 1728,
+ "model.layers.7.mlp.down_proj": 1772,
+ "model.layers.8.self_attn.q_proj": 1205,
+ "model.layers.8.self_attn.k_proj": 1202,
+ "model.layers.8.self_attn.v_proj": 1239,
+ "model.layers.8.self_attn.o_proj": 1310,
+ "model.layers.8.mlp.gate_proj": 1717,
+ "model.layers.8.mlp.up_proj": 1720,
+ "model.layers.8.mlp.down_proj": 1747,
+ "model.layers.9.self_attn.q_proj": 1193,
+ "model.layers.9.self_attn.k_proj": 1190,
+ "model.layers.9.self_attn.v_proj": 1229,
+ "model.layers.9.self_attn.o_proj": 1273,
+ "model.layers.9.mlp.gate_proj": 1711,
+ "model.layers.9.mlp.up_proj": 1711,
+ "model.layers.9.mlp.down_proj": 1727,
+ "model.layers.10.self_attn.q_proj": 1200,
+ "model.layers.10.self_attn.k_proj": 1194,
+ "model.layers.10.self_attn.v_proj": 1229,
+ "model.layers.10.self_attn.o_proj": 1248,
+ "model.layers.10.mlp.gate_proj": 1708,
+ "model.layers.10.mlp.up_proj": 1705,
+ "model.layers.10.mlp.down_proj": 1713,
+ "model.layers.11.self_attn.q_proj": 1196,
+ "model.layers.11.self_attn.k_proj": 1194,
+ "model.layers.11.self_attn.v_proj": 1211,
+ "model.layers.11.self_attn.o_proj": 1250,
+ "model.layers.11.mlp.gate_proj": 1698,
+ "model.layers.11.mlp.up_proj": 1693,
+ "model.layers.11.mlp.down_proj": 1696,
+ "model.layers.12.self_attn.q_proj": 1178,
+ "model.layers.12.self_attn.k_proj": 1177,
+ "model.layers.12.self_attn.v_proj": 1205,
+ "model.layers.12.self_attn.o_proj": 1223,
+ "model.layers.12.mlp.gate_proj": 1691,
+ "model.layers.12.mlp.up_proj": 1685,
+ "model.layers.12.mlp.down_proj": 1678,
+ "model.layers.13.self_attn.q_proj": 1174,
+ "model.layers.13.self_attn.k_proj": 1169,
+ "model.layers.13.self_attn.v_proj": 1188,
+ "model.layers.13.self_attn.o_proj": 1222,
+ "model.layers.13.mlp.gate_proj": 1686,
+ "model.layers.13.mlp.up_proj": 1677,
+ "model.layers.13.mlp.down_proj": 1655,
+ "model.layers.14.self_attn.q_proj": 1179,
+ "model.layers.14.self_attn.k_proj": 1178,
+ "model.layers.14.self_attn.v_proj": 1191,
+ "model.layers.14.self_attn.o_proj": 1204,
+ "model.layers.14.mlp.gate_proj": 1677,
+ "model.layers.14.mlp.up_proj": 1668,
+ "model.layers.14.mlp.down_proj": 1638,
+ "model.layers.15.self_attn.q_proj": 1176,
+ "model.layers.15.self_attn.k_proj": 1173,
+ "model.layers.15.self_attn.v_proj": 1186,
+ "model.layers.15.self_attn.o_proj": 1190,
+ "model.layers.15.mlp.gate_proj": 1668,
+ "model.layers.15.mlp.up_proj": 1658,
+ "model.layers.15.mlp.down_proj": 1613,
+ "model.layers.16.self_attn.q_proj": 1174,
+ "model.layers.16.self_attn.k_proj": 1172,
+ "model.layers.16.self_attn.v_proj": 1171,
+ "model.layers.16.self_attn.o_proj": 1163,
+ "model.layers.16.mlp.gate_proj": 1659,
+ "model.layers.16.mlp.up_proj": 1652,
+ "model.layers.16.mlp.down_proj": 1584,
+ "model.layers.17.self_attn.q_proj": 1160,
+ "model.layers.17.self_attn.k_proj": 1161,
+ "model.layers.17.self_attn.v_proj": 1165,
+ "model.layers.17.self_attn.o_proj": 1179,
+ "model.layers.17.mlp.gate_proj": 1645,
+ "model.layers.17.mlp.up_proj": 1640,
+ "model.layers.17.mlp.down_proj": 1574,
+ "model.layers.18.self_attn.q_proj": 1149,
+ "model.layers.18.self_attn.k_proj": 1147,
+ "model.layers.18.self_attn.v_proj": 1149,
+ "model.layers.18.self_attn.o_proj": 1177,
+ "model.layers.18.mlp.gate_proj": 1631,
+ "model.layers.18.mlp.up_proj": 1630,
+ "model.layers.18.mlp.down_proj": 1558,
+ "model.layers.19.self_attn.q_proj": 1154,
+ "model.layers.19.self_attn.k_proj": 1155,
+ "model.layers.19.self_attn.v_proj": 1151,
+ "model.layers.19.self_attn.o_proj": 1174,
+ "model.layers.19.mlp.gate_proj": 1621,
+ "model.layers.19.mlp.up_proj": 1622,
+ "model.layers.19.mlp.down_proj": 1548,
+ "model.layers.20.self_attn.q_proj": 1154,
+ "model.layers.20.self_attn.k_proj": 1154,
+ "model.layers.20.self_attn.v_proj": 1145,
+ "model.layers.20.self_attn.o_proj": 1165,
+ "model.layers.20.mlp.gate_proj": 1617,
+ "model.layers.20.mlp.up_proj": 1619,
+ "model.layers.20.mlp.down_proj": 1531,
+ "model.layers.21.self_attn.q_proj": 1147,
+ "model.layers.21.self_attn.k_proj": 1147,
+ "model.layers.21.self_attn.v_proj": 1136,
+ "model.layers.21.self_attn.o_proj": 1166,
+ "model.layers.21.mlp.gate_proj": 1608,
+ "model.layers.21.mlp.up_proj": 1612,
+ "model.layers.21.mlp.down_proj": 1524,
+ "model.layers.22.self_attn.q_proj": 1138,
+ "model.layers.22.self_attn.k_proj": 1138,
+ "model.layers.22.self_attn.v_proj": 1130,
+ "model.layers.22.self_attn.o_proj": 1143,
+ "model.layers.22.mlp.gate_proj": 1601,
+ "model.layers.22.mlp.up_proj": 1606,
+ "model.layers.22.mlp.down_proj": 1513,
+ "model.layers.23.self_attn.q_proj": 1123,
+ "model.layers.23.self_attn.k_proj": 1125,
+ "model.layers.23.self_attn.v_proj": 1116,
+ "model.layers.23.self_attn.o_proj": 1139,
+ "model.layers.23.mlp.gate_proj": 1592,
+ "model.layers.23.mlp.up_proj": 1596,
+ "model.layers.23.mlp.down_proj": 1506,
+ "model.layers.24.self_attn.q_proj": 1140,
+ "model.layers.24.self_attn.k_proj": 1141,
+ "model.layers.24.self_attn.v_proj": 1120,
+ "model.layers.24.self_attn.o_proj": 1136,
+ "model.layers.24.mlp.gate_proj": 1584,
+ "model.layers.24.mlp.up_proj": 1588,
+ "model.layers.24.mlp.down_proj": 1501,
+ "model.layers.25.self_attn.q_proj": 1118,
+ "model.layers.25.self_attn.k_proj": 1118,
+ "model.layers.25.self_attn.v_proj": 1102,
+ "model.layers.25.self_attn.o_proj": 1150,
+ "model.layers.25.mlp.gate_proj": 1576,
+ "model.layers.25.mlp.up_proj": 1580,
+ "model.layers.25.mlp.down_proj": 1497,
+ "model.layers.26.self_attn.q_proj": 1129,
+ "model.layers.26.self_attn.k_proj": 1129,
+ "model.layers.26.self_attn.v_proj": 1105,
+ "model.layers.26.self_attn.o_proj": 1129,
+ "model.layers.26.mlp.gate_proj": 1569,
+ "model.layers.26.mlp.up_proj": 1573,
+ "model.layers.26.mlp.down_proj": 1490,
+ "model.layers.27.self_attn.q_proj": 1111,
+ "model.layers.27.self_attn.k_proj": 1110,
+ "model.layers.27.self_attn.v_proj": 1098,
+ "model.layers.27.self_attn.o_proj": 1119,
+ "model.layers.27.mlp.gate_proj": 1562,
+ "model.layers.27.mlp.up_proj": 1566,
+ "model.layers.27.mlp.down_proj": 1481,
+ "model.layers.28.self_attn.q_proj": 1116,
+ "model.layers.28.self_attn.k_proj": 1114,
+ "model.layers.28.self_attn.v_proj": 1093,
+ "model.layers.28.self_attn.o_proj": 1117,
+ "model.layers.28.mlp.gate_proj": 1559,
+ "model.layers.28.mlp.up_proj": 1561,
+ "model.layers.28.mlp.down_proj": 1469,
+ "model.layers.29.self_attn.q_proj": 1132,
+ "model.layers.29.self_attn.k_proj": 1131,
+ "model.layers.29.self_attn.v_proj": 1098,
+ "model.layers.29.self_attn.o_proj": 1110,
+ "model.layers.29.mlp.gate_proj": 1557,
+ "model.layers.29.mlp.up_proj": 1559,
+ "model.layers.29.mlp.down_proj": 1458,
+ "model.layers.30.self_attn.q_proj": 1120,
+ "model.layers.30.self_attn.k_proj": 1119,
+ "model.layers.30.self_attn.v_proj": 1090,
+ "model.layers.30.self_attn.o_proj": 1106,
+ "model.layers.30.mlp.gate_proj": 1557,
+ "model.layers.30.mlp.up_proj": 1562,
+ "model.layers.30.mlp.down_proj": 1447,
+ "model.layers.31.self_attn.q_proj": 1161,
+ "model.layers.31.self_attn.k_proj": 1156,
+ "model.layers.31.self_attn.v_proj": 1131,
+ "model.layers.31.self_attn.o_proj": 1121,
+ "model.layers.31.mlp.gate_proj": 1576,
+ "model.layers.31.mlp.up_proj": 1583,
+ "model.layers.31.mlp.down_proj": 1438
+}