Training in progress, step 350
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 sha256:31a8d2429cebb283bbde4c281dd25c1ea90c2fb3b23caa9c63898b7d60443418
|
3 |
size 35668592
|
metrics.json
CHANGED
@@ -4,3 +4,4 @@
|
|
4 |
{"Step":200,"eval_loss":1.4602761269,"eval_runtime":27.615,"eval_samples_per_second":3.621,"eval_steps_per_second":0.471,"epoch":0.32}
|
5 |
{"Step":250,"eval_loss":1.4578460455,"eval_runtime":27.6799,"eval_samples_per_second":3.613,"eval_steps_per_second":0.47,"epoch":0.4}
|
6 |
{"Step":300,"eval_loss":1.4557801485,"eval_runtime":27.6578,"eval_samples_per_second":3.616,"eval_steps_per_second":0.47,"epoch":0.48}
|
|
|
|
4 |
{"Step":200,"eval_loss":1.4602761269,"eval_runtime":27.615,"eval_samples_per_second":3.621,"eval_steps_per_second":0.471,"epoch":0.32}
|
5 |
{"Step":250,"eval_loss":1.4578460455,"eval_runtime":27.6799,"eval_samples_per_second":3.613,"eval_steps_per_second":0.47,"epoch":0.4}
|
6 |
{"Step":300,"eval_loss":1.4557801485,"eval_runtime":27.6578,"eval_samples_per_second":3.616,"eval_steps_per_second":0.47,"epoch":0.48}
|
7 |
+
{"Step":350,"eval_loss":1.4533815384,"eval_runtime":27.6598,"eval_samples_per_second":3.615,"eval_steps_per_second":0.47,"epoch":0.56}
|
state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch": 0.
|
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,
|
@@ -82,6 +82,21 @@
|
|
82 |
"eval_samples_per_second": 3.613,
|
83 |
"eval_steps_per_second": 0.47,
|
84 |
"step": 250
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
85 |
}
|
86 |
],
|
87 |
"logging_steps": 50,
|
@@ -101,7 +116,7 @@
|
|
101 |
"attributes": {}
|
102 |
}
|
103 |
},
|
104 |
-
"total_flos":
|
105 |
"train_batch_size": 8,
|
106 |
"trial_name": null,
|
107 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.56,
|
5 |
"eval_steps": 50,
|
6 |
+
"global_step": 350,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
82 |
"eval_samples_per_second": 3.613,
|
83 |
"eval_steps_per_second": 0.47,
|
84 |
"step": 250
|
85 |
+
},
|
86 |
+
{
|
87 |
+
"epoch": 0.48,
|
88 |
+
"grad_norm": 0.040425803512334824,
|
89 |
+
"learning_rate": 0.00046710136336692356,
|
90 |
+
"loss": 1.4813,
|
91 |
+
"step": 300
|
92 |
+
},
|
93 |
+
{
|
94 |
+
"epoch": 0.48,
|
95 |
+
"eval_loss": 1.4557801485061646,
|
96 |
+
"eval_runtime": 27.6578,
|
97 |
+
"eval_samples_per_second": 3.616,
|
98 |
+
"eval_steps_per_second": 0.47,
|
99 |
+
"step": 300
|
100 |
}
|
101 |
],
|
102 |
"logging_steps": 50,
|
|
|
116 |
"attributes": {}
|
117 |
}
|
118 |
},
|
119 |
+
"total_flos": 2.4063644577792e+16,
|
120 |
"train_batch_size": 8,
|
121 |
"trial_name": null,
|
122 |
"trial_params": null
|