Upload pretrained weights file
Browse files- trainer_state.json +553 -0
- video_chatgpt-7B.bin +3 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,553 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 8823,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 7.5471698113207555e-06,
|
| 13 |
+
"loss": 6.6425,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.07,
|
| 18 |
+
"learning_rate": 1.5094339622641511e-05,
|
| 19 |
+
"loss": 4.8785,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.1,
|
| 24 |
+
"learning_rate": 1.9999174617418052e-05,
|
| 25 |
+
"loss": 2.0576,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.14,
|
| 30 |
+
"learning_rate": 1.9987722672633802e-05,
|
| 31 |
+
"loss": 1.7927,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.17,
|
| 36 |
+
"learning_rate": 1.9962812964571567e-05,
|
| 37 |
+
"loss": 1.7317,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.2,
|
| 42 |
+
"learning_rate": 1.9924479057334537e-05,
|
| 43 |
+
"loss": 1.7022,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.24,
|
| 48 |
+
"learning_rate": 1.9872772603202818e-05,
|
| 49 |
+
"loss": 1.6741,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.27,
|
| 54 |
+
"learning_rate": 1.9807763273035574e-05,
|
| 55 |
+
"loss": 1.6608,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.31,
|
| 60 |
+
"learning_rate": 1.9729538662394363e-05,
|
| 61 |
+
"loss": 1.6597,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.34,
|
| 66 |
+
"learning_rate": 1.9638204173514217e-05,
|
| 67 |
+
"loss": 1.6598,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.37,
|
| 72 |
+
"learning_rate": 1.953388287328142e-05,
|
| 73 |
+
"loss": 1.6501,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.41,
|
| 78 |
+
"learning_rate": 1.9416715327409453e-05,
|
| 79 |
+
"loss": 1.637,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.44,
|
| 84 |
+
"learning_rate": 1.9286859411036396e-05,
|
| 85 |
+
"loss": 1.6317,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.48,
|
| 90 |
+
"learning_rate": 1.914449009599919e-05,
|
| 91 |
+
"loss": 1.6292,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.51,
|
| 96 |
+
"learning_rate": 1.898979921507119e-05,
|
| 97 |
+
"loss": 1.6279,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.54,
|
| 102 |
+
"learning_rate": 1.8822995203480823e-05,
|
| 103 |
+
"loss": 1.6225,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.58,
|
| 108 |
+
"learning_rate": 1.86443028180596e-05,
|
| 109 |
+
"loss": 1.6234,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.61,
|
| 114 |
+
"learning_rate": 1.8453962834397847e-05,
|
| 115 |
+
"loss": 1.6073,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.65,
|
| 120 |
+
"learning_rate": 1.8252231722416328e-05,
|
| 121 |
+
"loss": 1.6119,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.68,
|
| 126 |
+
"learning_rate": 1.8039381300790812e-05,
|
| 127 |
+
"loss": 1.5936,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.71,
|
| 132 |
+
"learning_rate": 1.781569837069528e-05,
|
| 133 |
+
"loss": 1.6175,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.75,
|
| 138 |
+
"learning_rate": 1.758148432935723e-05,
|
| 139 |
+
"loss": 1.6093,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.78,
|
| 144 |
+
"learning_rate": 1.7337054763945823e-05,
|
| 145 |
+
"loss": 1.6081,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.82,
|
| 150 |
+
"learning_rate": 1.7082739026340097e-05,
|
| 151 |
+
"loss": 1.6064,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.85,
|
| 156 |
+
"learning_rate": 1.6818879789350134e-05,
|
| 157 |
+
"loss": 1.6016,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.88,
|
| 162 |
+
"learning_rate": 1.6545832584989235e-05,
|
| 163 |
+
"loss": 1.6062,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.92,
|
| 168 |
+
"learning_rate": 1.6263965325419206e-05,
|
| 169 |
+
"loss": 1.6042,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.95,
|
| 174 |
+
"learning_rate": 1.5973657807214245e-05,
|
| 175 |
+
"loss": 1.5967,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.99,
|
| 180 |
+
"learning_rate": 1.5675301199611413e-05,
|
| 181 |
+
"loss": 1.6095,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.02,
|
| 186 |
+
"learning_rate": 1.536929751743723e-05,
|
| 187 |
+
"loss": 1.5958,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 1.05,
|
| 192 |
+
"learning_rate": 1.5056059079420575e-05,
|
| 193 |
+
"loss": 1.6012,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.09,
|
| 198 |
+
"learning_rate": 1.4736007952621852e-05,
|
| 199 |
+
"loss": 1.5872,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.12,
|
| 204 |
+
"learning_rate": 1.4409575383726852e-05,
|
| 205 |
+
"loss": 1.5967,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.16,
|
| 210 |
+
"learning_rate": 1.4077201217971817e-05,
|
| 211 |
+
"loss": 1.5913,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.19,
|
| 216 |
+
"learning_rate": 1.3739333306482481e-05,
|
| 217 |
+
"loss": 1.5902,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.22,
|
| 222 |
+
"learning_rate": 1.3396426902825753e-05,
|
| 223 |
+
"loss": 1.5908,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.26,
|
| 228 |
+
"learning_rate": 1.3048944049587138e-05,
|
| 229 |
+
"loss": 1.5883,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.29,
|
| 234 |
+
"learning_rate": 1.2697352955800396e-05,
|
| 235 |
+
"loss": 1.5907,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.33,
|
| 240 |
+
"learning_rate": 1.2342127366068364e-05,
|
| 241 |
+
"loss": 1.5864,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.36,
|
| 246 |
+
"learning_rate": 1.1983745922224985e-05,
|
| 247 |
+
"loss": 1.5912,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.39,
|
| 252 |
+
"learning_rate": 1.1622691518398636e-05,
|
| 253 |
+
"loss": 1.59,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.43,
|
| 258 |
+
"learning_rate": 1.1259450650345798e-05,
|
| 259 |
+
"loss": 1.5837,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.46,
|
| 264 |
+
"learning_rate": 1.0894512759931785e-05,
|
| 265 |
+
"loss": 1.5805,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.5,
|
| 270 |
+
"learning_rate": 1.0528369575641793e-05,
|
| 271 |
+
"loss": 1.5916,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.53,
|
| 276 |
+
"learning_rate": 1.0161514450010882e-05,
|
| 277 |
+
"loss": 1.5975,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.56,
|
| 282 |
+
"learning_rate": 9.794441694865673e-06,
|
| 283 |
+
"loss": 1.5885,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.6,
|
| 288 |
+
"learning_rate": 9.427645915273446e-06,
|
| 289 |
+
"loss": 1.5849,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.63,
|
| 294 |
+
"learning_rate": 9.061621343096156e-06,
|
| 295 |
+
"loss": 1.5754,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.67,
|
| 300 |
+
"learning_rate": 8.696861171047268e-06,
|
| 301 |
+
"loss": 1.5906,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.7,
|
| 306 |
+
"learning_rate": 8.33385688814881e-06,
|
| 307 |
+
"loss": 1.5837,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.73,
|
| 312 |
+
"learning_rate": 7.97309761748402e-06,
|
| 313 |
+
"loss": 1.5807,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.77,
|
| 318 |
+
"learning_rate": 7.615069457137927e-06,
|
| 319 |
+
"loss": 1.5907,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.8,
|
| 324 |
+
"learning_rate": 7.260254825213902e-06,
|
| 325 |
+
"loss": 1.5826,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.84,
|
| 330 |
+
"learning_rate": 6.909131809808755e-06,
|
| 331 |
+
"loss": 1.5796,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.87,
|
| 336 |
+
"learning_rate": 6.562173524822188e-06,
|
| 337 |
+
"loss": 1.5814,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.9,
|
| 342 |
+
"learning_rate": 6.219847472468641e-06,
|
| 343 |
+
"loss": 1.5861,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.94,
|
| 348 |
+
"learning_rate": 5.882614913350499e-06,
|
| 349 |
+
"loss": 1.5744,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.97,
|
| 354 |
+
"learning_rate": 5.550930244941448e-06,
|
| 355 |
+
"loss": 1.5797,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 2.01,
|
| 360 |
+
"learning_rate": 5.2252403893173835e-06,
|
| 361 |
+
"loss": 1.5857,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 2.04,
|
| 366 |
+
"learning_rate": 4.9059841909599456e-06,
|
| 367 |
+
"loss": 1.5728,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 2.07,
|
| 372 |
+
"learning_rate": 4.593591825444028e-06,
|
| 373 |
+
"loss": 1.5701,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 2.11,
|
| 378 |
+
"learning_rate": 4.288484219806016e-06,
|
| 379 |
+
"loss": 1.5846,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.14,
|
| 384 |
+
"learning_rate": 3.991072485373858e-06,
|
| 385 |
+
"loss": 1.5716,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.18,
|
| 390 |
+
"learning_rate": 3.7017573638230296e-06,
|
| 391 |
+
"loss": 1.5798,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.21,
|
| 396 |
+
"learning_rate": 3.420928687204965e-06,
|
| 397 |
+
"loss": 1.5739,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.24,
|
| 402 |
+
"learning_rate": 3.1489648526753913e-06,
|
| 403 |
+
"loss": 1.5845,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.28,
|
| 408 |
+
"learning_rate": 2.8862323126304427e-06,
|
| 409 |
+
"loss": 1.5808,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.31,
|
| 414 |
+
"learning_rate": 2.6330850809374685e-06,
|
| 415 |
+
"loss": 1.5704,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.35,
|
| 420 |
+
"learning_rate": 2.389864255925913e-06,
|
| 421 |
+
"loss": 1.5769,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.38,
|
| 426 |
+
"learning_rate": 2.1568975607809895e-06,
|
| 427 |
+
"loss": 1.5765,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.41,
|
| 432 |
+
"learning_rate": 1.934498901959424e-06,
|
| 433 |
+
"loss": 1.5665,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.45,
|
| 438 |
+
"learning_rate": 1.722967946222277e-06,
|
| 439 |
+
"loss": 1.5822,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.48,
|
| 444 |
+
"learning_rate": 1.5225897168548032e-06,
|
| 445 |
+
"loss": 1.5931,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.52,
|
| 450 |
+
"learning_rate": 1.3336342096173239e-06,
|
| 451 |
+
"loss": 1.5855,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.55,
|
| 456 |
+
"learning_rate": 1.1563560289446819e-06,
|
| 457 |
+
"loss": 1.5736,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.58,
|
| 462 |
+
"learning_rate": 9.909940448844412e-07,
|
| 463 |
+
"loss": 1.5942,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.62,
|
| 468 |
+
"learning_rate": 8.377710712360631e-07,
|
| 469 |
+
"loss": 1.5792,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.65,
|
| 474 |
+
"learning_rate": 6.968935653247766e-07,
|
| 475 |
+
"loss": 1.5715,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.69,
|
| 480 |
+
"learning_rate": 5.685513498146533e-07,
|
| 481 |
+
"loss": 1.5777,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.72,
|
| 486 |
+
"learning_rate": 4.529173569357459e-07,
|
| 487 |
+
"loss": 1.5795,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.75,
|
| 492 |
+
"learning_rate": 3.5014739546990087e-07,
|
| 493 |
+
"loss": 1.5799,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.79,
|
| 498 |
+
"learning_rate": 2.603799408092389e-07,
|
| 499 |
+
"loss": 1.5743,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.82,
|
| 504 |
+
"learning_rate": 1.8373594837017505e-07,
|
| 505 |
+
"loss": 1.5779,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.86,
|
| 510 |
+
"learning_rate": 1.2031869061438494e-07,
|
| 511 |
+
"loss": 1.5689,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.89,
|
| 516 |
+
"learning_rate": 7.02136178963242e-08,
|
| 517 |
+
"loss": 1.5854,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.92,
|
| 522 |
+
"learning_rate": 3.3488243324814044e-08,
|
| 523 |
+
"loss": 1.5654,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.96,
|
| 528 |
+
"learning_rate": 1.0192051793809221e-08,
|
| 529 |
+
"loss": 1.5818,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.99,
|
| 534 |
+
"learning_rate": 3.5643330492995953e-10,
|
| 535 |
+
"loss": 1.5853,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 3.0,
|
| 540 |
+
"step": 8823,
|
| 541 |
+
"total_flos": 6.175868485067866e+18,
|
| 542 |
+
"train_loss": 1.6989395520347432,
|
| 543 |
+
"train_runtime": 13309.6743,
|
| 544 |
+
"train_samples_per_second": 21.208,
|
| 545 |
+
"train_steps_per_second": 0.663
|
| 546 |
+
}
|
| 547 |
+
],
|
| 548 |
+
"max_steps": 8823,
|
| 549 |
+
"num_train_epochs": 3,
|
| 550 |
+
"total_flos": 6.175868485067866e+18,
|
| 551 |
+
"trial_name": null,
|
| 552 |
+
"trial_params": null
|
| 553 |
+
}
|
video_chatgpt-7B.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c5d78b70ca98dd8ae06cfc0f49b3fd7a7057dcfdddb33a19736fea9e7eabaff8
|
| 3 |
+
size 541181308
|