Training in progress, step 200
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 sha256:6f939bd09b4e1ef84fc2a775173b99eed6cca9209aeed0b9233d8de29bc13c86
|
3 |
size 35668592
|
metrics.json
CHANGED
@@ -1,3 +1,4 @@
|
|
1 |
{"Step":50,"eval_loss":0.9260319471,"eval_runtime":60.9012,"eval_samples_per_second":1.642,"eval_steps_per_second":0.213,"epoch":0.0967117988}
|
2 |
{"Step":100,"eval_loss":0.8202204108,"eval_runtime":56.5241,"eval_samples_per_second":1.769,"eval_steps_per_second":0.23,"epoch":0.1934235977}
|
3 |
{"Step":150,"eval_loss":0.789476037,"eval_runtime":56.6018,"eval_samples_per_second":1.767,"eval_steps_per_second":0.23,"epoch":0.2901353965}
|
|
|
|
1 |
{"Step":50,"eval_loss":0.9260319471,"eval_runtime":60.9012,"eval_samples_per_second":1.642,"eval_steps_per_second":0.213,"epoch":0.0967117988}
|
2 |
{"Step":100,"eval_loss":0.8202204108,"eval_runtime":56.5241,"eval_samples_per_second":1.769,"eval_steps_per_second":0.23,"epoch":0.1934235977}
|
3 |
{"Step":150,"eval_loss":0.789476037,"eval_runtime":56.6018,"eval_samples_per_second":1.767,"eval_steps_per_second":0.23,"epoch":0.2901353965}
|
4 |
+
{"Step":200,"eval_loss":0.7783958316,"eval_runtime":56.5799,"eval_samples_per_second":1.767,"eval_steps_per_second":0.23,"epoch":0.3868471954}
|
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,
|
@@ -37,6 +37,21 @@
|
|
37 |
"eval_samples_per_second": 1.769,
|
38 |
"eval_steps_per_second": 0.23,
|
39 |
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
}
|
41 |
],
|
42 |
"logging_steps": 50,
|
@@ -56,7 +71,7 @@
|
|
56 |
"attributes": {}
|
57 |
}
|
58 |
},
|
59 |
-
"total_flos": 1.
|
60 |
"train_batch_size": 8,
|
61 |
"trial_name": null,
|
62 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.38684719535783363,
|
5 |
"eval_steps": 50,
|
6 |
+
"global_step": 200,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
37 |
"eval_samples_per_second": 1.769,
|
38 |
"eval_steps_per_second": 0.23,
|
39 |
"step": 100
|
40 |
+
},
|
41 |
+
{
|
42 |
+
"epoch": 0.2901353965183752,
|
43 |
+
"grad_norm": 0.02802114002406597,
|
44 |
+
"learning_rate": 0.0004807692307692308,
|
45 |
+
"loss": 0.7996,
|
46 |
+
"step": 150
|
47 |
+
},
|
48 |
+
{
|
49 |
+
"epoch": 0.2901353965183752,
|
50 |
+
"eval_loss": 0.7894760370254517,
|
51 |
+
"eval_runtime": 56.6018,
|
52 |
+
"eval_samples_per_second": 1.767,
|
53 |
+
"eval_steps_per_second": 0.23,
|
54 |
+
"step": 150
|
55 |
}
|
56 |
],
|
57 |
"logging_steps": 50,
|
|
|
71 |
"attributes": {}
|
72 |
}
|
73 |
},
|
74 |
+
"total_flos": 1.956722676252672e+16,
|
75 |
"train_batch_size": 8,
|
76 |
"trial_name": null,
|
77 |
"trial_params": null
|