Model save
Browse files- README.md +26 -11
- all_results.json +11 -11
- eval_results.json +6 -6
- train_results.json +6 -6
- trainer_state.json +497 -39
README.md
CHANGED
|
@@ -2,13 +2,9 @@
|
|
| 2 |
license: other
|
| 3 |
library_name: peft
|
| 4 |
tags:
|
| 5 |
-
- alignment-handbook
|
| 6 |
-
- generated_from_trainer
|
| 7 |
- trl
|
| 8 |
- sft
|
| 9 |
- generated_from_trainer
|
| 10 |
-
datasets:
|
| 11 |
-
- ruozhiba
|
| 12 |
base_model: 01-ai/Yi-6B
|
| 13 |
model-index:
|
| 14 |
- name: Yi-6B-ruozhiba3
|
|
@@ -20,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 20 |
|
| 21 |
# Yi-6B-ruozhiba3
|
| 22 |
|
| 23 |
-
This model is a fine-tuned version of [01-ai/Yi-6B](https://huggingface.co/01-ai/Yi-6B) on the
|
| 24 |
It achieves the following results on the evaluation set:
|
| 25 |
-
- Loss:
|
| 26 |
|
| 27 |
## Model description
|
| 28 |
|
|
@@ -41,20 +37,39 @@ More information needed
|
|
| 41 |
### Training hyperparameters
|
| 42 |
|
| 43 |
The following hyperparameters were used during training:
|
| 44 |
-
- learning_rate:
|
| 45 |
-
- train_batch_size:
|
| 46 |
-
- eval_batch_size:
|
| 47 |
- seed: 42
|
| 48 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 49 |
- lr_scheduler_type: cosine
|
| 50 |
- lr_scheduler_warmup_ratio: 0.1
|
| 51 |
-
- num_epochs:
|
| 52 |
|
| 53 |
### Training results
|
| 54 |
|
| 55 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 56 |
|:-------------:|:-----:|:----:|:---------------:|
|
| 57 |
-
| 2.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
|
| 59 |
|
| 60 |
### Framework versions
|
|
|
|
| 2 |
license: other
|
| 3 |
library_name: peft
|
| 4 |
tags:
|
|
|
|
|
|
|
| 5 |
- trl
|
| 6 |
- sft
|
| 7 |
- generated_from_trainer
|
|
|
|
|
|
|
| 8 |
base_model: 01-ai/Yi-6B
|
| 9 |
model-index:
|
| 10 |
- name: Yi-6B-ruozhiba3
|
|
|
|
| 16 |
|
| 17 |
# Yi-6B-ruozhiba3
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [01-ai/Yi-6B](https://huggingface.co/01-ai/Yi-6B) on the None dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 3.9909
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
|
|
| 37 |
### Training hyperparameters
|
| 38 |
|
| 39 |
The following hyperparameters were used during training:
|
| 40 |
+
- learning_rate: 5e-05
|
| 41 |
+
- train_batch_size: 4
|
| 42 |
+
- eval_batch_size: 4
|
| 43 |
- seed: 42
|
| 44 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 45 |
- lr_scheduler_type: cosine
|
| 46 |
- lr_scheduler_warmup_ratio: 0.1
|
| 47 |
+
- num_epochs: 20
|
| 48 |
|
| 49 |
### Training results
|
| 50 |
|
| 51 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 52 |
|:-------------:|:-----:|:----:|:---------------:|
|
| 53 |
+
| 2.644 | 1.0 | 55 | 2.3047 |
|
| 54 |
+
| 1.9548 | 2.0 | 110 | 1.9419 |
|
| 55 |
+
| 1.788 | 3.0 | 165 | 1.9135 |
|
| 56 |
+
| 1.6342 | 4.0 | 220 | 1.9499 |
|
| 57 |
+
| 1.3781 | 5.0 | 275 | 2.1321 |
|
| 58 |
+
| 1.0617 | 6.0 | 330 | 2.3518 |
|
| 59 |
+
| 0.8104 | 7.0 | 385 | 2.6090 |
|
| 60 |
+
| 0.5864 | 8.0 | 440 | 2.8890 |
|
| 61 |
+
| 0.4159 | 9.0 | 495 | 3.1356 |
|
| 62 |
+
| 0.3344 | 10.0 | 550 | 3.3190 |
|
| 63 |
+
| 0.2446 | 11.0 | 605 | 3.5470 |
|
| 64 |
+
| 0.199 | 12.0 | 660 | 3.6840 |
|
| 65 |
+
| 0.1245 | 13.0 | 715 | 3.7653 |
|
| 66 |
+
| 0.1208 | 14.0 | 770 | 3.8722 |
|
| 67 |
+
| 0.1003 | 15.0 | 825 | 3.9575 |
|
| 68 |
+
| 0.0767 | 16.0 | 880 | 3.9671 |
|
| 69 |
+
| 0.0913 | 17.0 | 935 | 3.9921 |
|
| 70 |
+
| 0.0895 | 18.0 | 990 | 3.9940 |
|
| 71 |
+
| 0.0671 | 19.0 | 1045 | 3.9915 |
|
| 72 |
+
| 0.0671 | 20.0 | 1100 | 3.9909 |
|
| 73 |
|
| 74 |
|
| 75 |
### Framework versions
|
all_results.json
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_loss":
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"eval_samples":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second":
|
| 8 |
-
"train_loss":
|
| 9 |
-
"train_runtime":
|
| 10 |
-
"train_samples":
|
| 11 |
-
"train_samples_per_second":
|
| 12 |
-
"train_steps_per_second":
|
| 13 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_loss": 3.990852117538452,
|
| 4 |
+
"eval_runtime": 1.2395,
|
| 5 |
+
"eval_samples": 23,
|
| 6 |
+
"eval_samples_per_second": 18.556,
|
| 7 |
+
"eval_steps_per_second": 4.841,
|
| 8 |
+
"train_loss": 0.6849299516461113,
|
| 9 |
+
"train_runtime": 917.6848,
|
| 10 |
+
"train_samples": 217,
|
| 11 |
+
"train_samples_per_second": 4.729,
|
| 12 |
+
"train_steps_per_second": 1.199
|
| 13 |
}
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_loss":
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"eval_samples":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_loss": 3.990852117538452,
|
| 4 |
+
"eval_runtime": 1.2395,
|
| 5 |
+
"eval_samples": 23,
|
| 6 |
+
"eval_samples_per_second": 18.556,
|
| 7 |
+
"eval_steps_per_second": 4.841
|
| 8 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"train_loss":
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"train_loss": 0.6849299516461113,
|
| 4 |
+
"train_runtime": 917.6848,
|
| 5 |
+
"train_samples": 217,
|
| 6 |
+
"train_samples_per_second": 4.729,
|
| 7 |
+
"train_steps_per_second": 1.199
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,68 +1,526 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
-
"epoch": 0.
|
| 13 |
-
"learning_rate":
|
| 14 |
-
"loss": 2.
|
| 15 |
"step": 1
|
| 16 |
},
|
| 17 |
{
|
| 18 |
-
"epoch": 0.
|
| 19 |
-
"learning_rate":
|
| 20 |
-
"loss": 2.
|
| 21 |
-
"step":
|
| 22 |
},
|
| 23 |
{
|
| 24 |
-
"epoch": 0.
|
| 25 |
-
"learning_rate":
|
| 26 |
-
"loss": 2.
|
| 27 |
-
"step":
|
| 28 |
},
|
| 29 |
{
|
| 30 |
-
"epoch": 0
|
| 31 |
-
"
|
| 32 |
-
"
|
| 33 |
-
"
|
|
|
|
|
|
|
| 34 |
},
|
| 35 |
{
|
| 36 |
-
"epoch": 1.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
"learning_rate": 0.0,
|
| 38 |
-
"loss":
|
| 39 |
-
"step":
|
| 40 |
},
|
| 41 |
{
|
| 42 |
-
"epoch":
|
| 43 |
-
"eval_loss":
|
| 44 |
-
"eval_runtime":
|
| 45 |
-
"eval_samples_per_second":
|
| 46 |
-
"eval_steps_per_second":
|
| 47 |
-
"step":
|
| 48 |
},
|
| 49 |
{
|
| 50 |
-
"epoch":
|
| 51 |
-
"step":
|
| 52 |
-
"total_flos":
|
| 53 |
-
"train_loss":
|
| 54 |
-
"train_runtime":
|
| 55 |
-
"train_samples_per_second":
|
| 56 |
-
"train_steps_per_second":
|
| 57 |
}
|
| 58 |
],
|
| 59 |
-
"logging_steps":
|
| 60 |
-
"max_steps":
|
| 61 |
"num_input_tokens_seen": 0,
|
| 62 |
-
"num_train_epochs":
|
| 63 |
-
"save_steps":
|
| 64 |
-
"total_flos":
|
| 65 |
-
"train_batch_size":
|
| 66 |
"trial_name": null,
|
| 67 |
"trial_params": null
|
| 68 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 20.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 1100,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 4.545454545454545e-07,
|
| 14 |
+
"loss": 2.5611,
|
| 15 |
"step": 1
|
| 16 |
},
|
| 17 |
{
|
| 18 |
+
"epoch": 0.36,
|
| 19 |
+
"learning_rate": 9.090909090909091e-06,
|
| 20 |
+
"loss": 2.5692,
|
| 21 |
+
"step": 20
|
| 22 |
},
|
| 23 |
{
|
| 24 |
+
"epoch": 0.73,
|
| 25 |
+
"learning_rate": 1.8181818181818182e-05,
|
| 26 |
+
"loss": 2.644,
|
| 27 |
+
"step": 40
|
| 28 |
},
|
| 29 |
{
|
| 30 |
+
"epoch": 1.0,
|
| 31 |
+
"eval_loss": 2.304738759994507,
|
| 32 |
+
"eval_runtime": 1.217,
|
| 33 |
+
"eval_samples_per_second": 18.899,
|
| 34 |
+
"eval_steps_per_second": 4.93,
|
| 35 |
+
"step": 55
|
| 36 |
},
|
| 37 |
{
|
| 38 |
+
"epoch": 1.09,
|
| 39 |
+
"learning_rate": 2.7272727272727273e-05,
|
| 40 |
+
"loss": 2.3827,
|
| 41 |
+
"step": 60
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 1.45,
|
| 45 |
+
"learning_rate": 3.6363636363636364e-05,
|
| 46 |
+
"loss": 2.0781,
|
| 47 |
+
"step": 80
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 1.82,
|
| 51 |
+
"learning_rate": 4.545454545454546e-05,
|
| 52 |
+
"loss": 1.9548,
|
| 53 |
+
"step": 100
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 2.0,
|
| 57 |
+
"eval_loss": 1.9419394731521606,
|
| 58 |
+
"eval_runtime": 1.2171,
|
| 59 |
+
"eval_samples_per_second": 18.898,
|
| 60 |
+
"eval_steps_per_second": 4.93,
|
| 61 |
+
"step": 110
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 2.18,
|
| 65 |
+
"learning_rate": 4.9987413559579636e-05,
|
| 66 |
+
"loss": 1.8022,
|
| 67 |
+
"step": 120
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 2.55,
|
| 71 |
+
"learning_rate": 4.988679806432712e-05,
|
| 72 |
+
"loss": 1.8295,
|
| 73 |
+
"step": 140
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 2.91,
|
| 77 |
+
"learning_rate": 4.968597221690986e-05,
|
| 78 |
+
"loss": 1.788,
|
| 79 |
+
"step": 160
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 3.0,
|
| 83 |
+
"eval_loss": 1.9134601354599,
|
| 84 |
+
"eval_runtime": 1.2132,
|
| 85 |
+
"eval_samples_per_second": 18.958,
|
| 86 |
+
"eval_steps_per_second": 4.946,
|
| 87 |
+
"step": 165
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 3.27,
|
| 91 |
+
"learning_rate": 4.938574467213518e-05,
|
| 92 |
+
"loss": 1.6784,
|
| 93 |
+
"step": 180
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 3.64,
|
| 97 |
+
"learning_rate": 4.898732434036244e-05,
|
| 98 |
+
"loss": 1.6528,
|
| 99 |
+
"step": 200
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 4.0,
|
| 103 |
+
"learning_rate": 4.849231551964771e-05,
|
| 104 |
+
"loss": 1.6342,
|
| 105 |
+
"step": 220
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 4.0,
|
| 109 |
+
"eval_loss": 1.9498955011367798,
|
| 110 |
+
"eval_runtime": 1.2502,
|
| 111 |
+
"eval_samples_per_second": 18.397,
|
| 112 |
+
"eval_steps_per_second": 4.799,
|
| 113 |
+
"step": 220
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 4.36,
|
| 117 |
+
"learning_rate": 4.790271143580174e-05,
|
| 118 |
+
"loss": 1.3818,
|
| 119 |
+
"step": 240
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 4.73,
|
| 123 |
+
"learning_rate": 4.722088621637309e-05,
|
| 124 |
+
"loss": 1.3781,
|
| 125 |
+
"step": 260
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 5.0,
|
| 129 |
+
"eval_loss": 2.1321451663970947,
|
| 130 |
+
"eval_runtime": 1.2118,
|
| 131 |
+
"eval_samples_per_second": 18.98,
|
| 132 |
+
"eval_steps_per_second": 4.951,
|
| 133 |
+
"step": 275
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 5.09,
|
| 137 |
+
"learning_rate": 4.644958533087443e-05,
|
| 138 |
+
"loss": 1.2512,
|
| 139 |
+
"step": 280
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"epoch": 5.45,
|
| 143 |
+
"learning_rate": 4.559191453574582e-05,
|
| 144 |
+
"loss": 1.0475,
|
| 145 |
+
"step": 300
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 5.82,
|
| 149 |
+
"learning_rate": 4.465132736856969e-05,
|
| 150 |
+
"loss": 1.0617,
|
| 151 |
+
"step": 320
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 6.0,
|
| 155 |
+
"eval_loss": 2.3518364429473877,
|
| 156 |
+
"eval_runtime": 1.216,
|
| 157 |
+
"eval_samples_per_second": 18.914,
|
| 158 |
+
"eval_steps_per_second": 4.934,
|
| 159 |
+
"step": 330
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 6.18,
|
| 163 |
+
"learning_rate": 4.3631611241893874e-05,
|
| 164 |
+
"loss": 0.9003,
|
| 165 |
+
"step": 340
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 6.55,
|
| 169 |
+
"learning_rate": 4.2536872192658036e-05,
|
| 170 |
+
"loss": 0.7805,
|
| 171 |
+
"step": 360
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 6.91,
|
| 175 |
+
"learning_rate": 4.137151834863213e-05,
|
| 176 |
+
"loss": 0.8104,
|
| 177 |
+
"step": 380
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 7.0,
|
| 181 |
+
"eval_loss": 2.609004259109497,
|
| 182 |
+
"eval_runtime": 1.2167,
|
| 183 |
+
"eval_samples_per_second": 18.904,
|
| 184 |
+
"eval_steps_per_second": 4.932,
|
| 185 |
+
"step": 385
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 7.27,
|
| 189 |
+
"learning_rate": 4.014024217844167e-05,
|
| 190 |
+
"loss": 0.6542,
|
| 191 |
+
"step": 400
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 7.64,
|
| 195 |
+
"learning_rate": 3.884800159665276e-05,
|
| 196 |
+
"loss": 0.5753,
|
| 197 |
+
"step": 420
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 8.0,
|
| 201 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 202 |
+
"loss": 0.5864,
|
| 203 |
+
"step": 440
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 8.0,
|
| 207 |
+
"eval_loss": 2.8889544010162354,
|
| 208 |
+
"eval_runtime": 1.2434,
|
| 209 |
+
"eval_samples_per_second": 18.498,
|
| 210 |
+
"eval_steps_per_second": 4.825,
|
| 211 |
+
"step": 440
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 8.36,
|
| 215 |
+
"learning_rate": 3.610166531514436e-05,
|
| 216 |
+
"loss": 0.4181,
|
| 217 |
+
"step": 460
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 8.73,
|
| 221 |
+
"learning_rate": 3.465862814232822e-05,
|
| 222 |
+
"loss": 0.4159,
|
| 223 |
+
"step": 480
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 9.0,
|
| 227 |
+
"eval_loss": 3.1356287002563477,
|
| 228 |
+
"eval_runtime": 1.2192,
|
| 229 |
+
"eval_samples_per_second": 18.865,
|
| 230 |
+
"eval_steps_per_second": 4.921,
|
| 231 |
+
"step": 495
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 9.09,
|
| 235 |
+
"learning_rate": 3.3176699082935545e-05,
|
| 236 |
+
"loss": 0.4188,
|
| 237 |
+
"step": 500
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 9.45,
|
| 241 |
+
"learning_rate": 3.166184534225087e-05,
|
| 242 |
+
"loss": 0.3131,
|
| 243 |
+
"step": 520
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 9.82,
|
| 247 |
+
"learning_rate": 3.012016670162977e-05,
|
| 248 |
+
"loss": 0.3344,
|
| 249 |
+
"step": 540
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 10.0,
|
| 253 |
+
"eval_loss": 3.3189520835876465,
|
| 254 |
+
"eval_runtime": 1.2131,
|
| 255 |
+
"eval_samples_per_second": 18.96,
|
| 256 |
+
"eval_steps_per_second": 4.946,
|
| 257 |
+
"step": 550
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 10.18,
|
| 261 |
+
"learning_rate": 2.8557870956832132e-05,
|
| 262 |
+
"loss": 0.3005,
|
| 263 |
+
"step": 560
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 10.55,
|
| 267 |
+
"learning_rate": 2.698124892141971e-05,
|
| 268 |
+
"loss": 0.2527,
|
| 269 |
+
"step": 580
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 10.91,
|
| 273 |
+
"learning_rate": 2.5396649095870202e-05,
|
| 274 |
+
"loss": 0.2446,
|
| 275 |
+
"step": 600
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 11.0,
|
| 279 |
+
"eval_loss": 3.5470495223999023,
|
| 280 |
+
"eval_runtime": 1.216,
|
| 281 |
+
"eval_samples_per_second": 18.914,
|
| 282 |
+
"eval_steps_per_second": 4.934,
|
| 283 |
+
"step": 605
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 11.27,
|
| 287 |
+
"learning_rate": 2.3810452104406444e-05,
|
| 288 |
+
"loss": 0.1855,
|
| 289 |
+
"step": 620
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 11.64,
|
| 293 |
+
"learning_rate": 2.222904500247473e-05,
|
| 294 |
+
"loss": 0.1745,
|
| 295 |
+
"step": 640
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 12.0,
|
| 299 |
+
"learning_rate": 2.0658795558326743e-05,
|
| 300 |
+
"loss": 0.199,
|
| 301 |
+
"step": 660
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 12.0,
|
| 305 |
+
"eval_loss": 3.68398118019104,
|
| 306 |
+
"eval_runtime": 1.2409,
|
| 307 |
+
"eval_samples_per_second": 18.535,
|
| 308 |
+
"eval_steps_per_second": 4.835,
|
| 309 |
+
"step": 660
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 12.36,
|
| 313 |
+
"learning_rate": 1.9106026612264316e-05,
|
| 314 |
+
"loss": 0.1455,
|
| 315 |
+
"step": 680
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 12.73,
|
| 319 |
+
"learning_rate": 1.7576990616793137e-05,
|
| 320 |
+
"loss": 0.1245,
|
| 321 |
+
"step": 700
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 13.0,
|
| 325 |
+
"eval_loss": 3.765277147293091,
|
| 326 |
+
"eval_runtime": 1.2162,
|
| 327 |
+
"eval_samples_per_second": 18.911,
|
| 328 |
+
"eval_steps_per_second": 4.933,
|
| 329 |
+
"step": 715
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 13.09,
|
| 333 |
+
"learning_rate": 1.6077844460203206e-05,
|
| 334 |
+
"loss": 0.1351,
|
| 335 |
+
"step": 720
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 13.45,
|
| 339 |
+
"learning_rate": 1.4614624674952842e-05,
|
| 340 |
+
"loss": 0.0967,
|
| 341 |
+
"step": 740
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 13.82,
|
| 345 |
+
"learning_rate": 1.3193223130682936e-05,
|
| 346 |
+
"loss": 0.1208,
|
| 347 |
+
"step": 760
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 14.0,
|
| 351 |
+
"eval_loss": 3.8721702098846436,
|
| 352 |
+
"eval_runtime": 1.2167,
|
| 353 |
+
"eval_samples_per_second": 18.903,
|
| 354 |
+
"eval_steps_per_second": 4.931,
|
| 355 |
+
"step": 770
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 14.18,
|
| 359 |
+
"learning_rate": 1.181936330973744e-05,
|
| 360 |
+
"loss": 0.0853,
|
| 361 |
+
"step": 780
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 14.55,
|
| 365 |
+
"learning_rate": 1.049857726072005e-05,
|
| 366 |
+
"loss": 0.0854,
|
| 367 |
+
"step": 800
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 14.91,
|
| 371 |
+
"learning_rate": 9.236183322886945e-06,
|
| 372 |
+
"loss": 0.1003,
|
| 373 |
+
"step": 820
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 15.0,
|
| 377 |
+
"eval_loss": 3.9574601650238037,
|
| 378 |
+
"eval_runtime": 1.2205,
|
| 379 |
+
"eval_samples_per_second": 18.844,
|
| 380 |
+
"eval_steps_per_second": 4.916,
|
| 381 |
+
"step": 825
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 15.27,
|
| 385 |
+
"learning_rate": 8.0372647110717e-06,
|
| 386 |
+
"loss": 0.0753,
|
| 387 |
+
"step": 840
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 15.64,
|
| 391 |
+
"learning_rate": 6.906649047373246e-06,
|
| 392 |
+
"loss": 0.0928,
|
| 393 |
+
"step": 860
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 16.0,
|
| 397 |
+
"learning_rate": 5.848888922025553e-06,
|
| 398 |
+
"loss": 0.0767,
|
| 399 |
+
"step": 880
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 16.0,
|
| 403 |
+
"eval_loss": 3.9671382904052734,
|
| 404 |
+
"eval_runtime": 1.2561,
|
| 405 |
+
"eval_samples_per_second": 18.311,
|
| 406 |
+
"eval_steps_per_second": 4.777,
|
| 407 |
+
"step": 880
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 16.36,
|
| 411 |
+
"learning_rate": 4.868243561723535e-06,
|
| 412 |
+
"loss": 0.0702,
|
| 413 |
+
"step": 900
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 16.73,
|
| 417 |
+
"learning_rate": 3.968661679220468e-06,
|
| 418 |
+
"loss": 0.0913,
|
| 419 |
+
"step": 920
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 17.0,
|
| 423 |
+
"eval_loss": 3.9921278953552246,
|
| 424 |
+
"eval_runtime": 1.2166,
|
| 425 |
+
"eval_samples_per_second": 18.905,
|
| 426 |
+
"eval_steps_per_second": 4.932,
|
| 427 |
+
"step": 935
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 17.09,
|
| 431 |
+
"learning_rate": 3.1537655732553768e-06,
|
| 432 |
+
"loss": 0.0698,
|
| 433 |
+
"step": 940
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 17.45,
|
| 437 |
+
"learning_rate": 2.4268365428344736e-06,
|
| 438 |
+
"loss": 0.0661,
|
| 439 |
+
"step": 960
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 17.82,
|
| 443 |
+
"learning_rate": 1.790801674598186e-06,
|
| 444 |
+
"loss": 0.0895,
|
| 445 |
+
"step": 980
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 18.0,
|
| 449 |
+
"eval_loss": 3.9939558506011963,
|
| 450 |
+
"eval_runtime": 1.2161,
|
| 451 |
+
"eval_samples_per_second": 18.913,
|
| 452 |
+
"eval_steps_per_second": 4.934,
|
| 453 |
+
"step": 990
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 18.18,
|
| 457 |
+
"learning_rate": 1.248222056476367e-06,
|
| 458 |
+
"loss": 0.0695,
|
| 459 |
+
"step": 1000
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 18.55,
|
| 463 |
+
"learning_rate": 8.012824650910938e-07,
|
| 464 |
+
"loss": 0.086,
|
| 465 |
+
"step": 1020
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 18.91,
|
| 469 |
+
"learning_rate": 4.517825684323324e-07,
|
| 470 |
+
"loss": 0.0671,
|
| 471 |
+
"step": 1040
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 19.0,
|
| 475 |
+
"eval_loss": 3.9915316104888916,
|
| 476 |
+
"eval_runtime": 1.2143,
|
| 477 |
+
"eval_samples_per_second": 18.941,
|
| 478 |
+
"eval_steps_per_second": 4.941,
|
| 479 |
+
"step": 1045
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 19.27,
|
| 483 |
+
"learning_rate": 2.011296792301165e-07,
|
| 484 |
+
"loss": 0.0681,
|
| 485 |
+
"step": 1060
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 19.64,
|
| 489 |
+
"learning_rate": 5.033308820289184e-08,
|
| 490 |
+
"loss": 0.09,
|
| 491 |
+
"step": 1080
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 20.0,
|
| 495 |
"learning_rate": 0.0,
|
| 496 |
+
"loss": 0.0671,
|
| 497 |
+
"step": 1100
|
| 498 |
},
|
| 499 |
{
|
| 500 |
+
"epoch": 20.0,
|
| 501 |
+
"eval_loss": 3.990852117538452,
|
| 502 |
+
"eval_runtime": 1.2506,
|
| 503 |
+
"eval_samples_per_second": 18.391,
|
| 504 |
+
"eval_steps_per_second": 4.798,
|
| 505 |
+
"step": 1100
|
| 506 |
},
|
| 507 |
{
|
| 508 |
+
"epoch": 20.0,
|
| 509 |
+
"step": 1100,
|
| 510 |
+
"total_flos": 3.807078373542298e+16,
|
| 511 |
+
"train_loss": 0.6849299516461113,
|
| 512 |
+
"train_runtime": 917.6848,
|
| 513 |
+
"train_samples_per_second": 4.729,
|
| 514 |
+
"train_steps_per_second": 1.199
|
| 515 |
}
|
| 516 |
],
|
| 517 |
+
"logging_steps": 20,
|
| 518 |
+
"max_steps": 1100,
|
| 519 |
"num_input_tokens_seen": 0,
|
| 520 |
+
"num_train_epochs": 20,
|
| 521 |
+
"save_steps": 20,
|
| 522 |
+
"total_flos": 3.807078373542298e+16,
|
| 523 |
+
"train_batch_size": 4,
|
| 524 |
"trial_name": null,
|
| 525 |
"trial_params": null
|
| 526 |
}
|