Training in progress, step 7500, checkpoint
Browse files
last-checkpoint/model-00001-of-00002.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4978139416
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:161c63173b11d278454f76e096130af0ea2e197a6c275cfa04ebf9e466be9c63
|
3 |
size 4978139416
|
last-checkpoint/model-00002-of-00002.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 3659223436
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:24a9ff8292338e7bef2a92a787972e8c312dc811b53b5c8f487afd672c88d3a2
|
3 |
size 3659223436
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 17241500333
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:effe03ee93685e827f87da010b75e451bf179e605eefd74137f038217c39e61d
|
3 |
size 17241500333
|
last-checkpoint/rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14567
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:070566d04626f1a1199d33a844fca7649870a2f1dc42c459cb3ffc8f9fae256c
|
3 |
size 14567
|
last-checkpoint/scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 623
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:baab263627df2d83ddbd983f1bff3bce2c66dd722c961b1e37caa6a98898e772
|
3 |
size 623
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch": 3.
|
5 |
"eval_steps": 100,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -1175,6 +1175,38 @@
|
|
1175 |
"eval_samples_per_second": 25.996,
|
1176 |
"eval_steps_per_second": 3.25,
|
1177 |
"step": 7300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1178 |
}
|
1179 |
],
|
1180 |
"logging_steps": 100,
|
@@ -1182,7 +1214,7 @@
|
|
1182 |
"num_input_tokens_seen": 0,
|
1183 |
"num_train_epochs": 30,
|
1184 |
"save_steps": 100,
|
1185 |
-
"total_flos": 8.
|
1186 |
"train_batch_size": 8,
|
1187 |
"trial_name": null,
|
1188 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 3.157230056830141,
|
5 |
"eval_steps": 100,
|
6 |
+
"global_step": 7500,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
1175 |
"eval_samples_per_second": 25.996,
|
1176 |
"eval_steps_per_second": 3.25,
|
1177 |
"step": 7300
|
1178 |
+
},
|
1179 |
+
{
|
1180 |
+
"epoch": 3.12,
|
1181 |
+
"grad_norm": 19.71111488342285,
|
1182 |
+
"learning_rate": 4.513003533568905e-05,
|
1183 |
+
"loss": 1.7744,
|
1184 |
+
"step": 7400
|
1185 |
+
},
|
1186 |
+
{
|
1187 |
+
"epoch": 3.12,
|
1188 |
+
"eval_cer": 0.5048052481326503,
|
1189 |
+
"eval_loss": 2.699909210205078,
|
1190 |
+
"eval_runtime": 377.4252,
|
1191 |
+
"eval_samples_per_second": 25.112,
|
1192 |
+
"eval_steps_per_second": 3.14,
|
1193 |
+
"step": 7400
|
1194 |
+
},
|
1195 |
+
{
|
1196 |
+
"epoch": 3.16,
|
1197 |
+
"grad_norm": 3.2357943058013916,
|
1198 |
+
"learning_rate": 4.505936395759717e-05,
|
1199 |
+
"loss": 1.8872,
|
1200 |
+
"step": 7500
|
1201 |
+
},
|
1202 |
+
{
|
1203 |
+
"epoch": 3.16,
|
1204 |
+
"eval_cer": 0.5123968558132259,
|
1205 |
+
"eval_loss": 2.0029726028442383,
|
1206 |
+
"eval_runtime": 372.4224,
|
1207 |
+
"eval_samples_per_second": 25.45,
|
1208 |
+
"eval_steps_per_second": 3.182,
|
1209 |
+
"step": 7500
|
1210 |
}
|
1211 |
],
|
1212 |
"logging_steps": 100,
|
|
|
1214 |
"num_input_tokens_seen": 0,
|
1215 |
"num_train_epochs": 30,
|
1216 |
"save_steps": 100,
|
1217 |
+
"total_flos": 8.232776432614433e+19,
|
1218 |
"train_batch_size": 8,
|
1219 |
"trial_name": null,
|
1220 |
"trial_params": null
|