Training in progress, step 46500, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 140637408
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f13565f10bec04a32788b51955784b6c8ad6af8c5a72ea32087ed5a6c173e53b
|
| 3 |
size 140637408
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 281362507
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e82549eff5ebe3d9ce68c5e9124cb8ff4ece4c4b568e8a1e3c4d5d4f502233d0
|
| 3 |
size 281362507
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14645
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1da75bced15d0c0ad08ac6f92e20da81e67fff47141ca76354909981931d2e09
|
| 3 |
size 14645
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ad90a43f66fabf6fc4d5d455d59e7d1cf3e1317fdb4e5836c7efe0c48a8cada7
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -33036,6 +33036,365 @@
|
|
| 33036 |
"eval_samples_per_second": 33.858,
|
| 33037 |
"eval_steps_per_second": 0.542,
|
| 33038 |
"step": 46000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33039 |
}
|
| 33040 |
],
|
| 33041 |
"logging_steps": 10,
|
|
@@ -33055,7 +33414,7 @@
|
|
| 33055 |
"attributes": {}
|
| 33056 |
}
|
| 33057 |
},
|
| 33058 |
-
"total_flos": 8.
|
| 33059 |
"train_batch_size": 128,
|
| 33060 |
"trial_name": null,
|
| 33061 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.6094843631216085,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 46500,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 33036 |
"eval_samples_per_second": 33.858,
|
| 33037 |
"eval_steps_per_second": 0.542,
|
| 33038 |
"step": 46000
|
| 33039 |
+
},
|
| 33040 |
+
{
|
| 33041 |
+
"epoch": 0.6030618397252733,
|
| 33042 |
+
"grad_norm": 0.6219264268875122,
|
| 33043 |
+
"learning_rate": 0.00010345472731011084,
|
| 33044 |
+
"loss": 3.2245,
|
| 33045 |
+
"step": 46010
|
| 33046 |
+
},
|
| 33047 |
+
{
|
| 33048 |
+
"epoch": 0.6031929116313209,
|
| 33049 |
+
"grad_norm": 0.614022433757782,
|
| 33050 |
+
"learning_rate": 0.0001033956266782268,
|
| 33051 |
+
"loss": 3.2619,
|
| 33052 |
+
"step": 46020
|
| 33053 |
+
},
|
| 33054 |
+
{
|
| 33055 |
+
"epoch": 0.6033239835373686,
|
| 33056 |
+
"grad_norm": 0.6726503968238831,
|
| 33057 |
+
"learning_rate": 0.0001033365340531018,
|
| 33058 |
+
"loss": 3.2604,
|
| 33059 |
+
"step": 46030
|
| 33060 |
+
},
|
| 33061 |
+
{
|
| 33062 |
+
"epoch": 0.6034550554434163,
|
| 33063 |
+
"grad_norm": 0.7042705416679382,
|
| 33064 |
+
"learning_rate": 0.00010327744944488807,
|
| 33065 |
+
"loss": 3.2621,
|
| 33066 |
+
"step": 46040
|
| 33067 |
+
},
|
| 33068 |
+
{
|
| 33069 |
+
"epoch": 0.6035861273494639,
|
| 33070 |
+
"grad_norm": 0.5855175852775574,
|
| 33071 |
+
"learning_rate": 0.00010321837286373647,
|
| 33072 |
+
"loss": 3.2132,
|
| 33073 |
+
"step": 46050
|
| 33074 |
+
},
|
| 33075 |
+
{
|
| 33076 |
+
"epoch": 0.6037171992555116,
|
| 33077 |
+
"grad_norm": 0.6653004884719849,
|
| 33078 |
+
"learning_rate": 0.00010315930431979658,
|
| 33079 |
+
"loss": 3.2278,
|
| 33080 |
+
"step": 46060
|
| 33081 |
+
},
|
| 33082 |
+
{
|
| 33083 |
+
"epoch": 0.6038482711615593,
|
| 33084 |
+
"grad_norm": 0.6612849235534668,
|
| 33085 |
+
"learning_rate": 0.00010310024382321658,
|
| 33086 |
+
"loss": 3.2559,
|
| 33087 |
+
"step": 46070
|
| 33088 |
+
},
|
| 33089 |
+
{
|
| 33090 |
+
"epoch": 0.6039793430676069,
|
| 33091 |
+
"grad_norm": 0.6311576962471008,
|
| 33092 |
+
"learning_rate": 0.0001030411913841431,
|
| 33093 |
+
"loss": 3.2318,
|
| 33094 |
+
"step": 46080
|
| 33095 |
+
},
|
| 33096 |
+
{
|
| 33097 |
+
"epoch": 0.6041104149736546,
|
| 33098 |
+
"grad_norm": 0.6075425744056702,
|
| 33099 |
+
"learning_rate": 0.00010298214701272159,
|
| 33100 |
+
"loss": 3.2248,
|
| 33101 |
+
"step": 46090
|
| 33102 |
+
},
|
| 33103 |
+
{
|
| 33104 |
+
"epoch": 0.6042414868797022,
|
| 33105 |
+
"grad_norm": 0.6190316677093506,
|
| 33106 |
+
"learning_rate": 0.00010292311071909602,
|
| 33107 |
+
"loss": 3.2141,
|
| 33108 |
+
"step": 46100
|
| 33109 |
+
},
|
| 33110 |
+
{
|
| 33111 |
+
"epoch": 0.6043725587857499,
|
| 33112 |
+
"grad_norm": 0.6562777161598206,
|
| 33113 |
+
"learning_rate": 0.00010286408251340905,
|
| 33114 |
+
"loss": 3.2349,
|
| 33115 |
+
"step": 46110
|
| 33116 |
+
},
|
| 33117 |
+
{
|
| 33118 |
+
"epoch": 0.6045036306917975,
|
| 33119 |
+
"grad_norm": 0.608486533164978,
|
| 33120 |
+
"learning_rate": 0.00010280506240580179,
|
| 33121 |
+
"loss": 3.2307,
|
| 33122 |
+
"step": 46120
|
| 33123 |
+
},
|
| 33124 |
+
{
|
| 33125 |
+
"epoch": 0.6046347025978451,
|
| 33126 |
+
"grad_norm": 0.6395454406738281,
|
| 33127 |
+
"learning_rate": 0.00010274605040641408,
|
| 33128 |
+
"loss": 3.2475,
|
| 33129 |
+
"step": 46130
|
| 33130 |
+
},
|
| 33131 |
+
{
|
| 33132 |
+
"epoch": 0.6047657745038928,
|
| 33133 |
+
"grad_norm": 0.6653308868408203,
|
| 33134 |
+
"learning_rate": 0.0001026870465253844,
|
| 33135 |
+
"loss": 3.2326,
|
| 33136 |
+
"step": 46140
|
| 33137 |
+
},
|
| 33138 |
+
{
|
| 33139 |
+
"epoch": 0.6048968464099405,
|
| 33140 |
+
"grad_norm": 0.7694252133369446,
|
| 33141 |
+
"learning_rate": 0.00010262805077284969,
|
| 33142 |
+
"loss": 3.2535,
|
| 33143 |
+
"step": 46150
|
| 33144 |
+
},
|
| 33145 |
+
{
|
| 33146 |
+
"epoch": 0.6050279183159881,
|
| 33147 |
+
"grad_norm": 0.6712005734443665,
|
| 33148 |
+
"learning_rate": 0.00010256906315894562,
|
| 33149 |
+
"loss": 3.242,
|
| 33150 |
+
"step": 46160
|
| 33151 |
+
},
|
| 33152 |
+
{
|
| 33153 |
+
"epoch": 0.6051589902220358,
|
| 33154 |
+
"grad_norm": 0.6094689965248108,
|
| 33155 |
+
"learning_rate": 0.00010251008369380645,
|
| 33156 |
+
"loss": 3.2578,
|
| 33157 |
+
"step": 46170
|
| 33158 |
+
},
|
| 33159 |
+
{
|
| 33160 |
+
"epoch": 0.6052900621280835,
|
| 33161 |
+
"grad_norm": 0.6993538737297058,
|
| 33162 |
+
"learning_rate": 0.00010245111238756503,
|
| 33163 |
+
"loss": 3.2303,
|
| 33164 |
+
"step": 46180
|
| 33165 |
+
},
|
| 33166 |
+
{
|
| 33167 |
+
"epoch": 0.6054211340341311,
|
| 33168 |
+
"grad_norm": 0.6670473217964172,
|
| 33169 |
+
"learning_rate": 0.0001023921492503527,
|
| 33170 |
+
"loss": 3.218,
|
| 33171 |
+
"step": 46190
|
| 33172 |
+
},
|
| 33173 |
+
{
|
| 33174 |
+
"epoch": 0.6055522059401788,
|
| 33175 |
+
"grad_norm": 0.6516269445419312,
|
| 33176 |
+
"learning_rate": 0.00010233319429229955,
|
| 33177 |
+
"loss": 3.2395,
|
| 33178 |
+
"step": 46200
|
| 33179 |
+
},
|
| 33180 |
+
{
|
| 33181 |
+
"epoch": 0.6056832778462264,
|
| 33182 |
+
"grad_norm": 0.6479690074920654,
|
| 33183 |
+
"learning_rate": 0.00010227424752353422,
|
| 33184 |
+
"loss": 3.2595,
|
| 33185 |
+
"step": 46210
|
| 33186 |
+
},
|
| 33187 |
+
{
|
| 33188 |
+
"epoch": 0.6058143497522741,
|
| 33189 |
+
"grad_norm": 0.606497585773468,
|
| 33190 |
+
"learning_rate": 0.00010221530895418396,
|
| 33191 |
+
"loss": 3.2286,
|
| 33192 |
+
"step": 46220
|
| 33193 |
+
},
|
| 33194 |
+
{
|
| 33195 |
+
"epoch": 0.6059454216583218,
|
| 33196 |
+
"grad_norm": 0.6444209218025208,
|
| 33197 |
+
"learning_rate": 0.00010215637859437445,
|
| 33198 |
+
"loss": 3.2202,
|
| 33199 |
+
"step": 46230
|
| 33200 |
+
},
|
| 33201 |
+
{
|
| 33202 |
+
"epoch": 0.6060764935643694,
|
| 33203 |
+
"grad_norm": 0.6833163499832153,
|
| 33204 |
+
"learning_rate": 0.00010209745645423016,
|
| 33205 |
+
"loss": 3.287,
|
| 33206 |
+
"step": 46240
|
| 33207 |
+
},
|
| 33208 |
+
{
|
| 33209 |
+
"epoch": 0.6062075654704171,
|
| 33210 |
+
"grad_norm": 0.6393874287605286,
|
| 33211 |
+
"learning_rate": 0.00010203854254387413,
|
| 33212 |
+
"loss": 3.2517,
|
| 33213 |
+
"step": 46250
|
| 33214 |
+
},
|
| 33215 |
+
{
|
| 33216 |
+
"epoch": 0.6063386373764648,
|
| 33217 |
+
"grad_norm": 0.7110936045646667,
|
| 33218 |
+
"learning_rate": 0.00010197963687342781,
|
| 33219 |
+
"loss": 3.2121,
|
| 33220 |
+
"step": 46260
|
| 33221 |
+
},
|
| 33222 |
+
{
|
| 33223 |
+
"epoch": 0.6064697092825124,
|
| 33224 |
+
"grad_norm": 0.6357977986335754,
|
| 33225 |
+
"learning_rate": 0.00010192073945301145,
|
| 33226 |
+
"loss": 3.2764,
|
| 33227 |
+
"step": 46270
|
| 33228 |
+
},
|
| 33229 |
+
{
|
| 33230 |
+
"epoch": 0.6066007811885601,
|
| 33231 |
+
"grad_norm": 0.6365695595741272,
|
| 33232 |
+
"learning_rate": 0.00010186185029274372,
|
| 33233 |
+
"loss": 3.2677,
|
| 33234 |
+
"step": 46280
|
| 33235 |
+
},
|
| 33236 |
+
{
|
| 33237 |
+
"epoch": 0.6067318530946078,
|
| 33238 |
+
"grad_norm": 0.649030864238739,
|
| 33239 |
+
"learning_rate": 0.00010180296940274207,
|
| 33240 |
+
"loss": 3.2464,
|
| 33241 |
+
"step": 46290
|
| 33242 |
+
},
|
| 33243 |
+
{
|
| 33244 |
+
"epoch": 0.6068629250006553,
|
| 33245 |
+
"grad_norm": 0.651207447052002,
|
| 33246 |
+
"learning_rate": 0.00010174409679312222,
|
| 33247 |
+
"loss": 3.2355,
|
| 33248 |
+
"step": 46300
|
| 33249 |
+
},
|
| 33250 |
+
{
|
| 33251 |
+
"epoch": 0.606993996906703,
|
| 33252 |
+
"grad_norm": 0.6729286313056946,
|
| 33253 |
+
"learning_rate": 0.0001016852324739988,
|
| 33254 |
+
"loss": 3.2191,
|
| 33255 |
+
"step": 46310
|
| 33256 |
+
},
|
| 33257 |
+
{
|
| 33258 |
+
"epoch": 0.6071250688127506,
|
| 33259 |
+
"grad_norm": 0.737335741519928,
|
| 33260 |
+
"learning_rate": 0.00010162637645548474,
|
| 33261 |
+
"loss": 3.1865,
|
| 33262 |
+
"step": 46320
|
| 33263 |
+
},
|
| 33264 |
+
{
|
| 33265 |
+
"epoch": 0.6072561407187983,
|
| 33266 |
+
"grad_norm": 0.660658597946167,
|
| 33267 |
+
"learning_rate": 0.00010156752874769177,
|
| 33268 |
+
"loss": 3.2212,
|
| 33269 |
+
"step": 46330
|
| 33270 |
+
},
|
| 33271 |
+
{
|
| 33272 |
+
"epoch": 0.607387212624846,
|
| 33273 |
+
"grad_norm": 0.6110531091690063,
|
| 33274 |
+
"learning_rate": 0.00010150868936072999,
|
| 33275 |
+
"loss": 3.2376,
|
| 33276 |
+
"step": 46340
|
| 33277 |
+
},
|
| 33278 |
+
{
|
| 33279 |
+
"epoch": 0.6075182845308936,
|
| 33280 |
+
"grad_norm": 0.5958556532859802,
|
| 33281 |
+
"learning_rate": 0.00010144985830470821,
|
| 33282 |
+
"loss": 3.2181,
|
| 33283 |
+
"step": 46350
|
| 33284 |
+
},
|
| 33285 |
+
{
|
| 33286 |
+
"epoch": 0.6076493564369413,
|
| 33287 |
+
"grad_norm": 0.7129707336425781,
|
| 33288 |
+
"learning_rate": 0.00010139103558973384,
|
| 33289 |
+
"loss": 3.2596,
|
| 33290 |
+
"step": 46360
|
| 33291 |
+
},
|
| 33292 |
+
{
|
| 33293 |
+
"epoch": 0.607780428342989,
|
| 33294 |
+
"grad_norm": 0.665770411491394,
|
| 33295 |
+
"learning_rate": 0.00010133222122591265,
|
| 33296 |
+
"loss": 3.2223,
|
| 33297 |
+
"step": 46370
|
| 33298 |
+
},
|
| 33299 |
+
{
|
| 33300 |
+
"epoch": 0.6079115002490366,
|
| 33301 |
+
"grad_norm": 0.6651104688644409,
|
| 33302 |
+
"learning_rate": 0.00010127341522334918,
|
| 33303 |
+
"loss": 3.2264,
|
| 33304 |
+
"step": 46380
|
| 33305 |
+
},
|
| 33306 |
+
{
|
| 33307 |
+
"epoch": 0.6080425721550843,
|
| 33308 |
+
"grad_norm": 1.0241354703903198,
|
| 33309 |
+
"learning_rate": 0.00010121461759214649,
|
| 33310 |
+
"loss": 3.2714,
|
| 33311 |
+
"step": 46390
|
| 33312 |
+
},
|
| 33313 |
+
{
|
| 33314 |
+
"epoch": 0.608173644061132,
|
| 33315 |
+
"grad_norm": 0.7431330680847168,
|
| 33316 |
+
"learning_rate": 0.0001011558283424061,
|
| 33317 |
+
"loss": 3.2469,
|
| 33318 |
+
"step": 46400
|
| 33319 |
+
},
|
| 33320 |
+
{
|
| 33321 |
+
"epoch": 0.6083047159671796,
|
| 33322 |
+
"grad_norm": 0.6433840990066528,
|
| 33323 |
+
"learning_rate": 0.00010109704748422824,
|
| 33324 |
+
"loss": 3.2413,
|
| 33325 |
+
"step": 46410
|
| 33326 |
+
},
|
| 33327 |
+
{
|
| 33328 |
+
"epoch": 0.6084357878732273,
|
| 33329 |
+
"grad_norm": 0.6310520768165588,
|
| 33330 |
+
"learning_rate": 0.00010103827502771156,
|
| 33331 |
+
"loss": 3.2671,
|
| 33332 |
+
"step": 46420
|
| 33333 |
+
},
|
| 33334 |
+
{
|
| 33335 |
+
"epoch": 0.6085668597792749,
|
| 33336 |
+
"grad_norm": 0.6456714868545532,
|
| 33337 |
+
"learning_rate": 0.00010097951098295333,
|
| 33338 |
+
"loss": 3.2625,
|
| 33339 |
+
"step": 46430
|
| 33340 |
+
},
|
| 33341 |
+
{
|
| 33342 |
+
"epoch": 0.6086979316853226,
|
| 33343 |
+
"grad_norm": 0.6332817673683167,
|
| 33344 |
+
"learning_rate": 0.0001009207553600495,
|
| 33345 |
+
"loss": 3.2386,
|
| 33346 |
+
"step": 46440
|
| 33347 |
+
},
|
| 33348 |
+
{
|
| 33349 |
+
"epoch": 0.6088290035913703,
|
| 33350 |
+
"grad_norm": 0.7312588095664978,
|
| 33351 |
+
"learning_rate": 0.00010086200816909425,
|
| 33352 |
+
"loss": 3.2637,
|
| 33353 |
+
"step": 46450
|
| 33354 |
+
},
|
| 33355 |
+
{
|
| 33356 |
+
"epoch": 0.6089600754974179,
|
| 33357 |
+
"grad_norm": 0.622722864151001,
|
| 33358 |
+
"learning_rate": 0.00010080326942018063,
|
| 33359 |
+
"loss": 3.2665,
|
| 33360 |
+
"step": 46460
|
| 33361 |
+
},
|
| 33362 |
+
{
|
| 33363 |
+
"epoch": 0.6090911474034656,
|
| 33364 |
+
"grad_norm": 0.6822323799133301,
|
| 33365 |
+
"learning_rate": 0.00010074453912340011,
|
| 33366 |
+
"loss": 3.2037,
|
| 33367 |
+
"step": 46470
|
| 33368 |
+
},
|
| 33369 |
+
{
|
| 33370 |
+
"epoch": 0.6092222193095131,
|
| 33371 |
+
"grad_norm": 0.7424899339675903,
|
| 33372 |
+
"learning_rate": 0.0001006858172888427,
|
| 33373 |
+
"loss": 3.2416,
|
| 33374 |
+
"step": 46480
|
| 33375 |
+
},
|
| 33376 |
+
{
|
| 33377 |
+
"epoch": 0.6093532912155608,
|
| 33378 |
+
"grad_norm": 0.6434851288795471,
|
| 33379 |
+
"learning_rate": 0.00010062710392659695,
|
| 33380 |
+
"loss": 3.2346,
|
| 33381 |
+
"step": 46490
|
| 33382 |
+
},
|
| 33383 |
+
{
|
| 33384 |
+
"epoch": 0.6094843631216085,
|
| 33385 |
+
"grad_norm": 0.7062909007072449,
|
| 33386 |
+
"learning_rate": 0.00010056839904675,
|
| 33387 |
+
"loss": 3.2432,
|
| 33388 |
+
"step": 46500
|
| 33389 |
+
},
|
| 33390 |
+
{
|
| 33391 |
+
"epoch": 0.6094843631216085,
|
| 33392 |
+
"eval_accuracy": 0.39903616813294235,
|
| 33393 |
+
"eval_loss": 3.2309646606445312,
|
| 33394 |
+
"eval_runtime": 60.7899,
|
| 33395 |
+
"eval_samples_per_second": 32.9,
|
| 33396 |
+
"eval_steps_per_second": 0.526,
|
| 33397 |
+
"step": 46500
|
| 33398 |
}
|
| 33399 |
],
|
| 33400 |
"logging_steps": 10,
|
|
|
|
| 33414 |
"attributes": {}
|
| 33415 |
}
|
| 33416 |
},
|
| 33417 |
+
"total_flos": 8.36246691643392e+17,
|
| 33418 |
"train_batch_size": 128,
|
| 33419 |
"trial_name": null,
|
| 33420 |
"trial_params": null
|