Training in progress, step 700
Browse files- adapter_model.safetensors +1 -1
- metrics.json +1 -0
- state.json +18 -3
adapter_model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 35668592
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f916fa238951c9e367e3db7fcbd24796cb3eb724be036ba458920b60714ae378
|
3 |
size 35668592
|
metrics.json
CHANGED
@@ -11,3 +11,4 @@
|
|
11 |
{"Step":550,"eval_loss":1.4466557503,"eval_runtime":27.6503,"eval_samples_per_second":3.617,"eval_steps_per_second":0.47,"epoch":0.88}
|
12 |
{"Step":600,"eval_loss":1.4460580349,"eval_runtime":27.6213,"eval_samples_per_second":3.62,"eval_steps_per_second":0.471,"epoch":0.96}
|
13 |
{"Step":650,"eval_loss":1.4470006227,"eval_runtime":27.6084,"eval_samples_per_second":3.622,"eval_steps_per_second":0.471,"epoch":1.04}
|
|
|
|
11 |
{"Step":550,"eval_loss":1.4466557503,"eval_runtime":27.6503,"eval_samples_per_second":3.617,"eval_steps_per_second":0.47,"epoch":0.88}
|
12 |
{"Step":600,"eval_loss":1.4460580349,"eval_runtime":27.6213,"eval_samples_per_second":3.62,"eval_steps_per_second":0.471,"epoch":0.96}
|
13 |
{"Step":650,"eval_loss":1.4470006227,"eval_runtime":27.6084,"eval_samples_per_second":3.622,"eval_steps_per_second":0.471,"epoch":1.04}
|
14 |
+
{"Step":700,"eval_loss":1.4475681782,"eval_runtime":27.6032,"eval_samples_per_second":3.623,"eval_steps_per_second":0.471,"epoch":1.12}
|
state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch": 1.
|
5 |
"eval_steps": 50,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -187,6 +187,21 @@
|
|
187 |
"eval_samples_per_second": 3.62,
|
188 |
"eval_steps_per_second": 0.471,
|
189 |
"step": 600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
190 |
}
|
191 |
],
|
192 |
"logging_steps": 50,
|
@@ -206,7 +221,7 @@
|
|
206 |
"attributes": {}
|
207 |
}
|
208 |
},
|
209 |
-
"total_flos":
|
210 |
"train_batch_size": 8,
|
211 |
"trial_name": null,
|
212 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.12,
|
5 |
"eval_steps": 50,
|
6 |
+
"global_step": 700,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
187 |
"eval_samples_per_second": 3.62,
|
188 |
"eval_steps_per_second": 0.471,
|
189 |
"step": 600
|
190 |
+
},
|
191 |
+
{
|
192 |
+
"epoch": 1.04,
|
193 |
+
"grad_norm": 0.039503153413534164,
|
194 |
+
"learning_rate": 0.00036336692353289865,
|
195 |
+
"loss": 1.4732,
|
196 |
+
"step": 650
|
197 |
+
},
|
198 |
+
{
|
199 |
+
"epoch": 1.04,
|
200 |
+
"eval_loss": 1.4470006227493286,
|
201 |
+
"eval_runtime": 27.6084,
|
202 |
+
"eval_samples_per_second": 3.622,
|
203 |
+
"eval_steps_per_second": 0.471,
|
204 |
+
"step": 650
|
205 |
}
|
206 |
],
|
207 |
"logging_steps": 50,
|
|
|
221 |
"attributes": {}
|
222 |
}
|
223 |
},
|
224 |
+
"total_flos": 5.235862179053568e+16,
|
225 |
"train_batch_size": 8,
|
226 |
"trial_name": null,
|
227 |
"trial_params": null
|