End of training
Browse files- all_results.json +14 -0
- eval_results.json +9 -0
- train_results.json +8 -0
- trainer_state.json +97 -0
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.0,
|
| 3 |
+
"eval_accuracy": 0.4946000366099213,
|
| 4 |
+
"eval_loss": 0.6931692361831665,
|
| 5 |
+
"eval_runtime": 30.42,
|
| 6 |
+
"eval_samples": 5463,
|
| 7 |
+
"eval_samples_per_second": 179.586,
|
| 8 |
+
"eval_steps_per_second": 5.621,
|
| 9 |
+
"train_loss": 0.7034443619781718,
|
| 10 |
+
"train_runtime": 7284.2364,
|
| 11 |
+
"train_samples": 104743,
|
| 12 |
+
"train_samples_per_second": 57.518,
|
| 13 |
+
"train_steps_per_second": 0.449
|
| 14 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.0,
|
| 3 |
+
"eval_accuracy": 0.4946000366099213,
|
| 4 |
+
"eval_loss": 0.6931692361831665,
|
| 5 |
+
"eval_runtime": 30.42,
|
| 6 |
+
"eval_samples": 5463,
|
| 7 |
+
"eval_samples_per_second": 179.586,
|
| 8 |
+
"eval_steps_per_second": 5.621
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.0,
|
| 3 |
+
"train_loss": 0.7034443619781718,
|
| 4 |
+
"train_runtime": 7284.2364,
|
| 5 |
+
"train_samples": 104743,
|
| 6 |
+
"train_samples_per_second": 57.518,
|
| 7 |
+
"train_steps_per_second": 0.449
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,97 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.9981670994348555,
|
| 5 |
+
"global_step": 3272,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.61,
|
| 12 |
+
"learning_rate": 4.246638141809291e-05,
|
| 13 |
+
"loss": 0.706,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 1.0,
|
| 18 |
+
"eval_accuracy": 0.4946000366099213,
|
| 19 |
+
"eval_loss": 0.7039228677749634,
|
| 20 |
+
"eval_runtime": 30.7626,
|
| 21 |
+
"eval_samples_per_second": 177.586,
|
| 22 |
+
"eval_steps_per_second": 5.559,
|
| 23 |
+
"step": 818
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 1.22,
|
| 27 |
+
"learning_rate": 3.4825794621026896e-05,
|
| 28 |
+
"loss": 0.7042,
|
| 29 |
+
"step": 1000
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 1.83,
|
| 33 |
+
"learning_rate": 2.718520782396088e-05,
|
| 34 |
+
"loss": 0.7032,
|
| 35 |
+
"step": 1500
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"epoch": 2.0,
|
| 39 |
+
"eval_accuracy": 0.4946000366099213,
|
| 40 |
+
"eval_loss": 0.6934332847595215,
|
| 41 |
+
"eval_runtime": 30.7656,
|
| 42 |
+
"eval_samples_per_second": 177.569,
|
| 43 |
+
"eval_steps_per_second": 5.558,
|
| 44 |
+
"step": 1636
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 2.44,
|
| 48 |
+
"learning_rate": 1.9544621026894865e-05,
|
| 49 |
+
"loss": 0.7031,
|
| 50 |
+
"step": 2000
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 3.0,
|
| 54 |
+
"eval_accuracy": 0.5053999633900788,
|
| 55 |
+
"eval_loss": 0.6931120753288269,
|
| 56 |
+
"eval_runtime": 30.7565,
|
| 57 |
+
"eval_samples_per_second": 177.621,
|
| 58 |
+
"eval_steps_per_second": 5.56,
|
| 59 |
+
"step": 2455
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 3.05,
|
| 63 |
+
"learning_rate": 1.1904034229828852e-05,
|
| 64 |
+
"loss": 0.7027,
|
| 65 |
+
"step": 2500
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 3.67,
|
| 69 |
+
"learning_rate": 4.263447432762836e-06,
|
| 70 |
+
"loss": 0.7021,
|
| 71 |
+
"step": 3000
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 4.0,
|
| 75 |
+
"eval_accuracy": 0.4946000366099213,
|
| 76 |
+
"eval_loss": 0.6931692361831665,
|
| 77 |
+
"eval_runtime": 30.7781,
|
| 78 |
+
"eval_samples_per_second": 177.497,
|
| 79 |
+
"eval_steps_per_second": 5.556,
|
| 80 |
+
"step": 3272
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 4.0,
|
| 84 |
+
"step": 3272,
|
| 85 |
+
"total_flos": 3.902826037167698e+17,
|
| 86 |
+
"train_loss": 0.7034443619781718,
|
| 87 |
+
"train_runtime": 7284.2364,
|
| 88 |
+
"train_samples_per_second": 57.518,
|
| 89 |
+
"train_steps_per_second": 0.449
|
| 90 |
+
}
|
| 91 |
+
],
|
| 92 |
+
"max_steps": 3272,
|
| 93 |
+
"num_train_epochs": 4,
|
| 94 |
+
"total_flos": 3.902826037167698e+17,
|
| 95 |
+
"trial_name": null,
|
| 96 |
+
"trial_params": null
|
| 97 |
+
}
|