Commit
·
a3eedbd
1
Parent(s):
351e4f3
End of training
Browse files- all_results.json +16 -0
- eval_results.json +11 -0
- runs/Sep18_18-21-44_patrick-general-gpu/events.out.tfevents.1631989416.patrick-general-gpu.389664.0 +2 -2
- runs/Sep18_18-21-44_patrick-general-gpu/events.out.tfevents.1632012588.patrick-general-gpu.389664.2 +3 -0
- train_results.json +8 -0
- trainer_state.json +76 -0
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_accuracy": 0.8847390551570616,
|
| 4 |
+
"eval_combined_score": 0.865679382097654,
|
| 5 |
+
"eval_f1": 0.8466197090382463,
|
| 6 |
+
"eval_loss": 0.36861559748649597,
|
| 7 |
+
"eval_runtime": 293.3009,
|
| 8 |
+
"eval_samples": 40430,
|
| 9 |
+
"eval_samples_per_second": 137.845,
|
| 10 |
+
"eval_steps_per_second": 17.231,
|
| 11 |
+
"train_loss": 0.2536681392615943,
|
| 12 |
+
"train_runtime": 22875.5112,
|
| 13 |
+
"train_samples": 363846,
|
| 14 |
+
"train_samples_per_second": 47.716,
|
| 15 |
+
"train_steps_per_second": 2.982
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_accuracy": 0.8847390551570616,
|
| 4 |
+
"eval_combined_score": 0.865679382097654,
|
| 5 |
+
"eval_f1": 0.8466197090382463,
|
| 6 |
+
"eval_loss": 0.36861559748649597,
|
| 7 |
+
"eval_runtime": 293.3009,
|
| 8 |
+
"eval_samples": 40430,
|
| 9 |
+
"eval_samples_per_second": 137.845,
|
| 10 |
+
"eval_steps_per_second": 17.231
|
| 11 |
+
}
|
runs/Sep18_18-21-44_patrick-general-gpu/events.out.tfevents.1631989416.patrick-general-gpu.389664.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9d57c7fd194109843ad062181c7c337dc7f1696f7be39d8afa21e93d2dc7fdd1
|
| 3 |
+
size 5321
|
runs/Sep18_18-21-44_patrick-general-gpu/events.out.tfevents.1632012588.patrick-general-gpu.389664.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:85f0121f6c5a68a455bf6309f15658c446de5f0185f734d70271f6ba846e2437
|
| 3 |
+
size 475
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"train_loss": 0.2536681392615943,
|
| 4 |
+
"train_runtime": 22875.5112,
|
| 5 |
+
"train_samples": 363846,
|
| 6 |
+
"train_samples_per_second": 47.716,
|
| 7 |
+
"train_steps_per_second": 2.982
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,76 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 68223,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 1.0,
|
| 12 |
+
"learning_rate": 1.3333333333333333e-05,
|
| 13 |
+
"loss": 0.3484,
|
| 14 |
+
"step": 22741
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 1.0,
|
| 18 |
+
"eval_accuracy": 0.867647786297304,
|
| 19 |
+
"eval_combined_score": 0.8486629929990515,
|
| 20 |
+
"eval_f1": 0.829678199700799,
|
| 21 |
+
"eval_loss": 0.3014103174209595,
|
| 22 |
+
"eval_runtime": 292.1565,
|
| 23 |
+
"eval_samples_per_second": 138.385,
|
| 24 |
+
"eval_steps_per_second": 17.299,
|
| 25 |
+
"step": 22741
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 2.0,
|
| 29 |
+
"learning_rate": 6.666666666666667e-06,
|
| 30 |
+
"loss": 0.2387,
|
| 31 |
+
"step": 45482
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 2.0,
|
| 35 |
+
"eval_accuracy": 0.8800890427900074,
|
| 36 |
+
"eval_combined_score": 0.8615115077230022,
|
| 37 |
+
"eval_f1": 0.8429339726559969,
|
| 38 |
+
"eval_loss": 0.3011280596256256,
|
| 39 |
+
"eval_runtime": 292.4225,
|
| 40 |
+
"eval_samples_per_second": 138.259,
|
| 41 |
+
"eval_steps_per_second": 17.283,
|
| 42 |
+
"step": 45482
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 3.0,
|
| 46 |
+
"learning_rate": 0.0,
|
| 47 |
+
"loss": 0.1739,
|
| 48 |
+
"step": 68223
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 3.0,
|
| 52 |
+
"eval_accuracy": 0.8847390551570616,
|
| 53 |
+
"eval_combined_score": 0.865679382097654,
|
| 54 |
+
"eval_f1": 0.8466197090382463,
|
| 55 |
+
"eval_loss": 0.36861559748649597,
|
| 56 |
+
"eval_runtime": 291.8123,
|
| 57 |
+
"eval_samples_per_second": 138.548,
|
| 58 |
+
"eval_steps_per_second": 17.319,
|
| 59 |
+
"step": 68223
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 3.0,
|
| 63 |
+
"step": 68223,
|
| 64 |
+
"total_flos": 1.941180482476892e+17,
|
| 65 |
+
"train_loss": 0.2536681392615943,
|
| 66 |
+
"train_runtime": 22875.5112,
|
| 67 |
+
"train_samples_per_second": 47.716,
|
| 68 |
+
"train_steps_per_second": 2.982
|
| 69 |
+
}
|
| 70 |
+
],
|
| 71 |
+
"max_steps": 68223,
|
| 72 |
+
"num_train_epochs": 3,
|
| 73 |
+
"total_flos": 1.941180482476892e+17,
|
| 74 |
+
"trial_name": null,
|
| 75 |
+
"trial_params": null
|
| 76 |
+
}
|