Commit
•
f41837a
1
Parent(s):
5767298
Training in progress, step 176500
Browse files
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5544997664
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4829c8d97387ccb4c56f85fac2fec1891c54be6e06f2c40b8e1a50a7553e2513
|
| 3 |
size 5544997664
|
last-checkpoint/adapter_config.json
CHANGED
|
@@ -22,11 +22,11 @@
|
|
| 22 |
"target_modules": [
|
| 23 |
"gate_proj",
|
| 24 |
"up_proj",
|
|
|
|
|
|
|
| 25 |
"k_proj",
|
| 26 |
"o_proj",
|
| 27 |
-
"
|
| 28 |
-
"v_proj",
|
| 29 |
-
"down_proj"
|
| 30 |
],
|
| 31 |
"task_type": "CAUSAL_LM",
|
| 32 |
"use_dora": false,
|
|
|
|
| 22 |
"target_modules": [
|
| 23 |
"gate_proj",
|
| 24 |
"up_proj",
|
| 25 |
+
"down_proj",
|
| 26 |
+
"q_proj",
|
| 27 |
"k_proj",
|
| 28 |
"o_proj",
|
| 29 |
+
"v_proj"
|
|
|
|
|
|
|
| 30 |
],
|
| 31 |
"task_type": "CAUSAL_LM",
|
| 32 |
"use_dora": false,
|
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5544997664
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:827134f8bc4269f2b2e4de9a603020b794835578e2bbfc7f2c4377e360c08169
|
| 3 |
size 5544997664
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 674093138
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33445f1bf4eea592138ef02c9d3845698e8dbe97f6d897cf7c712fd7f8bc406b
|
| 3 |
size 674093138
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c3e1519b3525ee3b89552a8c064cc15f5a192f59a80e19fcabc9854e1e8b4732
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:244846527155197c1f1306848f61d350db80de94a16e52843db6a3c6063d2b04
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 200,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -12075,286 +12075,6 @@
|
|
| 12075 |
"learning_rate": 1.996596052275932e-05,
|
| 12076 |
"loss": 1.5559,
|
| 12077 |
"step": 172400
|
| 12078 |
-
},
|
| 12079 |
-
{
|
| 12080 |
-
"epoch": 0.13143588176713827,
|
| 12081 |
-
"grad_norm": 0.708541989326477,
|
| 12082 |
-
"learning_rate": 1.9965921043659414e-05,
|
| 12083 |
-
"loss": 1.5405,
|
| 12084 |
-
"step": 172500
|
| 12085 |
-
},
|
| 12086 |
-
{
|
| 12087 |
-
"epoch": 0.1315120764812062,
|
| 12088 |
-
"grad_norm": 0.5723733305931091,
|
| 12089 |
-
"learning_rate": 1.9965881541717857e-05,
|
| 12090 |
-
"loss": 1.5467,
|
| 12091 |
-
"step": 172600
|
| 12092 |
-
},
|
| 12093 |
-
{
|
| 12094 |
-
"epoch": 0.1315882711952741,
|
| 12095 |
-
"grad_norm": 0.5531366467475891,
|
| 12096 |
-
"learning_rate": 1.996584201693474e-05,
|
| 12097 |
-
"loss": 1.5528,
|
| 12098 |
-
"step": 172700
|
| 12099 |
-
},
|
| 12100 |
-
{
|
| 12101 |
-
"epoch": 0.131664465909342,
|
| 12102 |
-
"grad_norm": 0.5077796578407288,
|
| 12103 |
-
"learning_rate": 1.9965802469310148e-05,
|
| 12104 |
-
"loss": 1.5144,
|
| 12105 |
-
"step": 172800
|
| 12106 |
-
},
|
| 12107 |
-
{
|
| 12108 |
-
"epoch": 0.13174066062340992,
|
| 12109 |
-
"grad_norm": 0.6950669884681702,
|
| 12110 |
-
"learning_rate": 1.996576289884417e-05,
|
| 12111 |
-
"loss": 1.4381,
|
| 12112 |
-
"step": 172900
|
| 12113 |
-
},
|
| 12114 |
-
{
|
| 12115 |
-
"epoch": 0.1318168553374778,
|
| 12116 |
-
"grad_norm": 0.6486429572105408,
|
| 12117 |
-
"learning_rate": 1.996572330553691e-05,
|
| 12118 |
-
"loss": 1.5391,
|
| 12119 |
-
"step": 173000
|
| 12120 |
-
},
|
| 12121 |
-
{
|
| 12122 |
-
"epoch": 0.13189305005154572,
|
| 12123 |
-
"grad_norm": 0.707037627696991,
|
| 12124 |
-
"learning_rate": 1.9965683689388448e-05,
|
| 12125 |
-
"loss": 1.5175,
|
| 12126 |
-
"step": 173100
|
| 12127 |
-
},
|
| 12128 |
-
{
|
| 12129 |
-
"epoch": 0.13196924476561364,
|
| 12130 |
-
"grad_norm": 0.8228338956832886,
|
| 12131 |
-
"learning_rate": 1.9965644050398875e-05,
|
| 12132 |
-
"loss": 1.5695,
|
| 12133 |
-
"step": 173200
|
| 12134 |
-
},
|
| 12135 |
-
{
|
| 12136 |
-
"epoch": 0.13204543947968153,
|
| 12137 |
-
"grad_norm": 0.8611860871315002,
|
| 12138 |
-
"learning_rate": 1.9965604388568286e-05,
|
| 12139 |
-
"loss": 1.5804,
|
| 12140 |
-
"step": 173300
|
| 12141 |
-
},
|
| 12142 |
-
{
|
| 12143 |
-
"epoch": 0.13212163419374945,
|
| 12144 |
-
"grad_norm": 0.593332827091217,
|
| 12145 |
-
"learning_rate": 1.996556470389677e-05,
|
| 12146 |
-
"loss": 1.5022,
|
| 12147 |
-
"step": 173400
|
| 12148 |
-
},
|
| 12149 |
-
{
|
| 12150 |
-
"epoch": 0.13219782890781734,
|
| 12151 |
-
"grad_norm": 0.4934659004211426,
|
| 12152 |
-
"learning_rate": 1.9965524996384415e-05,
|
| 12153 |
-
"loss": 1.5131,
|
| 12154 |
-
"step": 173500
|
| 12155 |
-
},
|
| 12156 |
-
{
|
| 12157 |
-
"epoch": 0.13227402362188526,
|
| 12158 |
-
"grad_norm": 0.461875319480896,
|
| 12159 |
-
"learning_rate": 1.9965485266031317e-05,
|
| 12160 |
-
"loss": 1.4976,
|
| 12161 |
-
"step": 173600
|
| 12162 |
-
},
|
| 12163 |
-
{
|
| 12164 |
-
"epoch": 0.13235021833595317,
|
| 12165 |
-
"grad_norm": 0.7695358991622925,
|
| 12166 |
-
"learning_rate": 1.996544551283757e-05,
|
| 12167 |
-
"loss": 1.4759,
|
| 12168 |
-
"step": 173700
|
| 12169 |
-
},
|
| 12170 |
-
{
|
| 12171 |
-
"epoch": 0.13242641305002106,
|
| 12172 |
-
"grad_norm": 0.23193073272705078,
|
| 12173 |
-
"learning_rate": 1.9965405736803248e-05,
|
| 12174 |
-
"loss": 1.5739,
|
| 12175 |
-
"step": 173800
|
| 12176 |
-
},
|
| 12177 |
-
{
|
| 12178 |
-
"epoch": 0.13250260776408898,
|
| 12179 |
-
"grad_norm": 0.5889491438865662,
|
| 12180 |
-
"learning_rate": 1.9965365937928464e-05,
|
| 12181 |
-
"loss": 1.6115,
|
| 12182 |
-
"step": 173900
|
| 12183 |
-
},
|
| 12184 |
-
{
|
| 12185 |
-
"epoch": 0.13257880247815687,
|
| 12186 |
-
"grad_norm": 0.17042429745197296,
|
| 12187 |
-
"learning_rate": 1.9965326116213294e-05,
|
| 12188 |
-
"loss": 1.6021,
|
| 12189 |
-
"step": 174000
|
| 12190 |
-
},
|
| 12191 |
-
{
|
| 12192 |
-
"epoch": 0.1326549971922248,
|
| 12193 |
-
"grad_norm": 0.7887442708015442,
|
| 12194 |
-
"learning_rate": 1.9965286271657837e-05,
|
| 12195 |
-
"loss": 1.5276,
|
| 12196 |
-
"step": 174100
|
| 12197 |
-
},
|
| 12198 |
-
{
|
| 12199 |
-
"epoch": 0.1327311919062927,
|
| 12200 |
-
"grad_norm": 0.522075355052948,
|
| 12201 |
-
"learning_rate": 1.996524640426218e-05,
|
| 12202 |
-
"loss": 1.6293,
|
| 12203 |
-
"step": 174200
|
| 12204 |
-
},
|
| 12205 |
-
{
|
| 12206 |
-
"epoch": 0.1328073866203606,
|
| 12207 |
-
"grad_norm": 0.8914234042167664,
|
| 12208 |
-
"learning_rate": 1.996520651402642e-05,
|
| 12209 |
-
"loss": 1.5579,
|
| 12210 |
-
"step": 174300
|
| 12211 |
-
},
|
| 12212 |
-
{
|
| 12213 |
-
"epoch": 0.1328835813344285,
|
| 12214 |
-
"grad_norm": 0.5836375951766968,
|
| 12215 |
-
"learning_rate": 1.9965166600950642e-05,
|
| 12216 |
-
"loss": 1.5017,
|
| 12217 |
-
"step": 174400
|
| 12218 |
-
},
|
| 12219 |
-
{
|
| 12220 |
-
"epoch": 0.1329597760484964,
|
| 12221 |
-
"grad_norm": 0.8345896005630493,
|
| 12222 |
-
"learning_rate": 1.9965126665034943e-05,
|
| 12223 |
-
"loss": 1.5726,
|
| 12224 |
-
"step": 174500
|
| 12225 |
-
},
|
| 12226 |
-
{
|
| 12227 |
-
"epoch": 0.13303597076256432,
|
| 12228 |
-
"grad_norm": 0.5659860968589783,
|
| 12229 |
-
"learning_rate": 1.9965086706279407e-05,
|
| 12230 |
-
"loss": 1.5298,
|
| 12231 |
-
"step": 174600
|
| 12232 |
-
},
|
| 12233 |
-
{
|
| 12234 |
-
"epoch": 0.1331121654766322,
|
| 12235 |
-
"grad_norm": 0.4374712109565735,
|
| 12236 |
-
"learning_rate": 1.9965046724684136e-05,
|
| 12237 |
-
"loss": 1.5165,
|
| 12238 |
-
"step": 174700
|
| 12239 |
-
},
|
| 12240 |
-
{
|
| 12241 |
-
"epoch": 0.13318836019070013,
|
| 12242 |
-
"grad_norm": 0.5508346557617188,
|
| 12243 |
-
"learning_rate": 1.9965006720249214e-05,
|
| 12244 |
-
"loss": 1.5273,
|
| 12245 |
-
"step": 174800
|
| 12246 |
-
},
|
| 12247 |
-
{
|
| 12248 |
-
"epoch": 0.13326455490476805,
|
| 12249 |
-
"grad_norm": 1.0864572525024414,
|
| 12250 |
-
"learning_rate": 1.9964966692974733e-05,
|
| 12251 |
-
"loss": 1.5108,
|
| 12252 |
-
"step": 174900
|
| 12253 |
-
},
|
| 12254 |
-
{
|
| 12255 |
-
"epoch": 0.13334074961883594,
|
| 12256 |
-
"grad_norm": 0.3920418322086334,
|
| 12257 |
-
"learning_rate": 1.9964926642860785e-05,
|
| 12258 |
-
"loss": 1.5229,
|
| 12259 |
-
"step": 175000
|
| 12260 |
-
},
|
| 12261 |
-
{
|
| 12262 |
-
"epoch": 0.13341694433290385,
|
| 12263 |
-
"grad_norm": 0.845400869846344,
|
| 12264 |
-
"learning_rate": 1.9964886569907468e-05,
|
| 12265 |
-
"loss": 1.4647,
|
| 12266 |
-
"step": 175100
|
| 12267 |
-
},
|
| 12268 |
-
{
|
| 12269 |
-
"epoch": 0.13349313904697174,
|
| 12270 |
-
"grad_norm": 0.717110276222229,
|
| 12271 |
-
"learning_rate": 1.9964846474114866e-05,
|
| 12272 |
-
"loss": 1.5466,
|
| 12273 |
-
"step": 175200
|
| 12274 |
-
},
|
| 12275 |
-
{
|
| 12276 |
-
"epoch": 0.13356933376103966,
|
| 12277 |
-
"grad_norm": 0.5370656847953796,
|
| 12278 |
-
"learning_rate": 1.9964806355483074e-05,
|
| 12279 |
-
"loss": 1.5762,
|
| 12280 |
-
"step": 175300
|
| 12281 |
-
},
|
| 12282 |
-
{
|
| 12283 |
-
"epoch": 0.13364552847510758,
|
| 12284 |
-
"grad_norm": 1.161123514175415,
|
| 12285 |
-
"learning_rate": 1.9964766214012184e-05,
|
| 12286 |
-
"loss": 1.5829,
|
| 12287 |
-
"step": 175400
|
| 12288 |
-
},
|
| 12289 |
-
{
|
| 12290 |
-
"epoch": 0.13372172318917547,
|
| 12291 |
-
"grad_norm": 0.6493296027183533,
|
| 12292 |
-
"learning_rate": 1.996472604970229e-05,
|
| 12293 |
-
"loss": 1.5684,
|
| 12294 |
-
"step": 175500
|
| 12295 |
-
},
|
| 12296 |
-
{
|
| 12297 |
-
"epoch": 0.13379791790324339,
|
| 12298 |
-
"grad_norm": 1.104617714881897,
|
| 12299 |
-
"learning_rate": 1.9964685862553482e-05,
|
| 12300 |
-
"loss": 1.5545,
|
| 12301 |
-
"step": 175600
|
| 12302 |
-
},
|
| 12303 |
-
{
|
| 12304 |
-
"epoch": 0.13387411261731127,
|
| 12305 |
-
"grad_norm": 1.4800457954406738,
|
| 12306 |
-
"learning_rate": 1.996464565256585e-05,
|
| 12307 |
-
"loss": 1.5038,
|
| 12308 |
-
"step": 175700
|
| 12309 |
-
},
|
| 12310 |
-
{
|
| 12311 |
-
"epoch": 0.1339503073313792,
|
| 12312 |
-
"grad_norm": 0.966915488243103,
|
| 12313 |
-
"learning_rate": 1.9964605419739488e-05,
|
| 12314 |
-
"loss": 1.4949,
|
| 12315 |
-
"step": 175800
|
| 12316 |
-
},
|
| 12317 |
-
{
|
| 12318 |
-
"epoch": 0.1340265020454471,
|
| 12319 |
-
"grad_norm": 0.9788563847541809,
|
| 12320 |
-
"learning_rate": 1.9964565164074493e-05,
|
| 12321 |
-
"loss": 1.5273,
|
| 12322 |
-
"step": 175900
|
| 12323 |
-
},
|
| 12324 |
-
{
|
| 12325 |
-
"epoch": 0.134102696759515,
|
| 12326 |
-
"grad_norm": 0.5662705898284912,
|
| 12327 |
-
"learning_rate": 1.9964524885570945e-05,
|
| 12328 |
-
"loss": 1.495,
|
| 12329 |
-
"step": 176000
|
| 12330 |
-
},
|
| 12331 |
-
{
|
| 12332 |
-
"epoch": 0.13417889147358292,
|
| 12333 |
-
"grad_norm": 1.1160712242126465,
|
| 12334 |
-
"learning_rate": 1.9964484584228952e-05,
|
| 12335 |
-
"loss": 1.5652,
|
| 12336 |
-
"step": 176100
|
| 12337 |
-
},
|
| 12338 |
-
{
|
| 12339 |
-
"epoch": 0.1342550861876508,
|
| 12340 |
-
"grad_norm": 0.8404836058616638,
|
| 12341 |
-
"learning_rate": 1.9964444260048593e-05,
|
| 12342 |
-
"loss": 1.533,
|
| 12343 |
-
"step": 176200
|
| 12344 |
-
},
|
| 12345 |
-
{
|
| 12346 |
-
"epoch": 0.13433128090171872,
|
| 12347 |
-
"grad_norm": 0.6309127807617188,
|
| 12348 |
-
"learning_rate": 1.996440391302997e-05,
|
| 12349 |
-
"loss": 1.5509,
|
| 12350 |
-
"step": 176300
|
| 12351 |
-
},
|
| 12352 |
-
{
|
| 12353 |
-
"epoch": 0.13440747561578664,
|
| 12354 |
-
"grad_norm": 0.8548604249954224,
|
| 12355 |
-
"learning_rate": 1.9964363543173166e-05,
|
| 12356 |
-
"loss": 1.4678,
|
| 12357 |
-
"step": 176400
|
| 12358 |
}
|
| 12359 |
],
|
| 12360 |
"logging_steps": 100,
|
|
@@ -12374,7 +12094,7 @@
|
|
| 12374 |
"attributes": {}
|
| 12375 |
}
|
| 12376 |
},
|
| 12377 |
-
"total_flos": 2.
|
| 12378 |
"train_batch_size": 1,
|
| 12379 |
"trial_name": null,
|
| 12380 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.13135968705307038,
|
| 5 |
"eval_steps": 200,
|
| 6 |
+
"global_step": 172400,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 12075 |
"learning_rate": 1.996596052275932e-05,
|
| 12076 |
"loss": 1.5559,
|
| 12077 |
"step": 172400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12078 |
}
|
| 12079 |
],
|
| 12080 |
"logging_steps": 100,
|
|
|
|
| 12094 |
"attributes": {}
|
| 12095 |
}
|
| 12096 |
},
|
| 12097 |
+
"total_flos": 2.3486437306994196e+18,
|
| 12098 |
"train_batch_size": 1,
|
| 12099 |
"trial_name": null,
|
| 12100 |
"trial_params": null
|