Model save
Browse files- README.md +26 -8
- adapter_model.safetensors +1 -1
- all_results.json +8 -8
- eval_results.json +5 -5
- runs/Jan21_13-28-46_98f107f1aa39/events.out.tfevents.1705843831.98f107f1aa39.8746.0 +2 -2
- runs/Jan21_13-28-46_98f107f1aa39/events.out.tfevents.1705859995.98f107f1aa39.8746.1 +3 -0
- train_results.json +4 -4
- trainer_state.json +385 -45
README.md
CHANGED
|
@@ -1,4 +1,5 @@
|
|
| 1 |
---
|
|
|
|
| 2 |
library_name: peft
|
| 3 |
tags:
|
| 4 |
- trl
|
|
@@ -6,7 +7,7 @@ tags:
|
|
| 6 |
- generated_from_trainer
|
| 7 |
datasets:
|
| 8 |
- generator
|
| 9 |
-
base_model:
|
| 10 |
model-index:
|
| 11 |
- name: Tukan-1.1B-Chat-reasoning-sft
|
| 12 |
results: []
|
|
@@ -17,9 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 17 |
|
| 18 |
# Tukan-1.1B-Chat-reasoning-sft
|
| 19 |
|
| 20 |
-
This model
|
| 21 |
It achieves the following results on the evaluation set:
|
| 22 |
-
- Loss: 1.
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
@@ -47,16 +48,33 @@ The following hyperparameters were used during training:
|
|
| 47 |
- total_train_batch_size: 120
|
| 48 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 49 |
- lr_scheduler_type: cosine
|
| 50 |
-
- num_epochs:
|
| 51 |
|
| 52 |
### Training results
|
| 53 |
|
| 54 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 55 |
|:-------------:|:-----:|:----:|:---------------:|
|
| 56 |
-
| 1.
|
| 57 |
-
| 1.
|
| 58 |
-
| 1.
|
| 59 |
-
| 1.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 60 |
|
| 61 |
|
| 62 |
### Framework versions
|
|
|
|
| 1 |
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
library_name: peft
|
| 4 |
tags:
|
| 5 |
- trl
|
|
|
|
| 7 |
- generated_from_trainer
|
| 8 |
datasets:
|
| 9 |
- generator
|
| 10 |
+
base_model: TinyLlama/TinyLlama-1.1B-Chat-v1.0
|
| 11 |
model-index:
|
| 12 |
- name: Tukan-1.1B-Chat-reasoning-sft
|
| 13 |
results: []
|
|
|
|
| 18 |
|
| 19 |
# Tukan-1.1B-Chat-reasoning-sft
|
| 20 |
|
| 21 |
+
This model is a fine-tuned version of [TinyLlama/TinyLlama-1.1B-Chat-v1.0](https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0) on the generator dataset.
|
| 22 |
It achieves the following results on the evaluation set:
|
| 23 |
+
- Loss: 1.0196
|
| 24 |
|
| 25 |
## Model description
|
| 26 |
|
|
|
|
| 48 |
- total_train_batch_size: 120
|
| 49 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 50 |
- lr_scheduler_type: cosine
|
| 51 |
+
- num_epochs: 5
|
| 52 |
|
| 53 |
### Training results
|
| 54 |
|
| 55 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 56 |
|:-------------:|:-----:|:----:|:---------------:|
|
| 57 |
+
| 1.3366 | 0.24 | 10 | 1.2783 |
|
| 58 |
+
| 1.2563 | 0.47 | 20 | 1.2321 |
|
| 59 |
+
| 1.2289 | 0.71 | 30 | 1.2012 |
|
| 60 |
+
| 1.1837 | 0.94 | 40 | 1.1688 |
|
| 61 |
+
| 1.1534 | 1.18 | 50 | 1.1306 |
|
| 62 |
+
| 1.1254 | 1.42 | 60 | 1.1037 |
|
| 63 |
+
| 1.1011 | 1.65 | 70 | 1.0882 |
|
| 64 |
+
| 1.0825 | 1.89 | 80 | 1.0748 |
|
| 65 |
+
| 1.0876 | 2.12 | 90 | 1.0635 |
|
| 66 |
+
| 1.0716 | 2.36 | 100 | 1.0540 |
|
| 67 |
+
| 1.0517 | 2.59 | 110 | 1.0459 |
|
| 68 |
+
| 1.0289 | 2.83 | 120 | 1.0389 |
|
| 69 |
+
| 1.0564 | 3.07 | 130 | 1.0332 |
|
| 70 |
+
| 1.034 | 3.3 | 140 | 1.0288 |
|
| 71 |
+
| 1.0337 | 3.54 | 150 | 1.0253 |
|
| 72 |
+
| 1.033 | 3.77 | 160 | 1.0231 |
|
| 73 |
+
| 1.0312 | 4.01 | 170 | 1.0213 |
|
| 74 |
+
| 1.0207 | 4.25 | 180 | 1.0204 |
|
| 75 |
+
| 1.0271 | 4.48 | 190 | 1.0198 |
|
| 76 |
+
| 1.0351 | 4.72 | 200 | 1.0197 |
|
| 77 |
+
| 1.0339 | 4.95 | 210 | 1.0196 |
|
| 78 |
|
| 79 |
|
| 80 |
### Framework versions
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 26361536
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7f38d00bd2666a4a83e6ea634036fafdc3c28558f76c503bb2063e64b38e98ab
|
| 3 |
size 26361536
|
all_results.json
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_loss": 1.
|
| 4 |
-
"eval_runtime": 24.
|
| 5 |
"eval_samples": 662,
|
| 6 |
-
"eval_samples_per_second": 5.
|
| 7 |
-
"eval_steps_per_second": 1.
|
| 8 |
-
"train_loss": 1.
|
| 9 |
-
"train_runtime":
|
| 10 |
"train_samples": 25778,
|
| 11 |
-
"train_samples_per_second": 1.
|
| 12 |
"train_steps_per_second": 0.013
|
| 13 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 4.95,
|
| 3 |
+
"eval_loss": 1.0196079015731812,
|
| 4 |
+
"eval_runtime": 24.8904,
|
| 5 |
"eval_samples": 662,
|
| 6 |
+
"eval_samples_per_second": 5.183,
|
| 7 |
+
"eval_steps_per_second": 1.728,
|
| 8 |
+
"train_loss": 1.1020877304531278,
|
| 9 |
+
"train_runtime": 16139.1818,
|
| 10 |
"train_samples": 25778,
|
| 11 |
+
"train_samples_per_second": 1.575,
|
| 12 |
"train_steps_per_second": 0.013
|
| 13 |
}
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_loss": 1.
|
| 4 |
-
"eval_runtime": 24.
|
| 5 |
"eval_samples": 662,
|
| 6 |
-
"eval_samples_per_second": 5.
|
| 7 |
-
"eval_steps_per_second": 1.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 4.95,
|
| 3 |
+
"eval_loss": 1.0196079015731812,
|
| 4 |
+
"eval_runtime": 24.8904,
|
| 5 |
"eval_samples": 662,
|
| 6 |
+
"eval_samples_per_second": 5.183,
|
| 7 |
+
"eval_steps_per_second": 1.728
|
| 8 |
}
|
runs/Jan21_13-28-46_98f107f1aa39/events.out.tfevents.1705843831.98f107f1aa39.8746.0
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:edad9ea16b95d898cd98bc9c211c98ed2bc169748bef8396cbcca7678f3ba014
|
| 3 |
+
size 17178
|
runs/Jan21_13-28-46_98f107f1aa39/events.out.tfevents.1705859995.98f107f1aa39.8746.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:025886936cdb06c2776f8b164452cdb3ac51b8b50d3e63898f8672b8acfaca15
|
| 3 |
+
size 359
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"train_loss": 1.
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 25778,
|
| 6 |
-
"train_samples_per_second": 1.
|
| 7 |
"train_steps_per_second": 0.013
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 4.95,
|
| 3 |
+
"train_loss": 1.1020877304531278,
|
| 4 |
+
"train_runtime": 16139.1818,
|
| 5 |
"train_samples": 25778,
|
| 6 |
+
"train_samples_per_second": 1.575,
|
| 7 |
"train_steps_per_second": 0.013
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,115 +1,455 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 10,
|
| 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": 0.02,
|
| 13 |
-
"learning_rate": 4.
|
| 14 |
-
"loss": 1.
|
| 15 |
"step": 1
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.12,
|
| 19 |
-
"learning_rate": 4.
|
| 20 |
-
"loss": 1.
|
| 21 |
"step": 5
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.24,
|
| 25 |
-
"learning_rate": 4.
|
| 26 |
-
"loss": 1.
|
| 27 |
"step": 10
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.24,
|
| 31 |
-
"eval_loss": 1.
|
| 32 |
-
"eval_runtime":
|
| 33 |
-
"eval_samples_per_second": 5.
|
| 34 |
-
"eval_steps_per_second": 1.
|
| 35 |
"step": 10
|
| 36 |
},
|
| 37 |
{
|
| 38 |
"epoch": 0.35,
|
| 39 |
-
"learning_rate":
|
| 40 |
-
"loss": 1.
|
| 41 |
"step": 15
|
| 42 |
},
|
| 43 |
{
|
| 44 |
"epoch": 0.47,
|
| 45 |
-
"learning_rate":
|
| 46 |
-
"loss": 1.
|
| 47 |
"step": 20
|
| 48 |
},
|
| 49 |
{
|
| 50 |
"epoch": 0.47,
|
| 51 |
-
"eval_loss": 1.
|
| 52 |
-
"eval_runtime":
|
| 53 |
-
"eval_samples_per_second": 5.
|
| 54 |
-
"eval_steps_per_second": 1.
|
| 55 |
"step": 20
|
| 56 |
},
|
| 57 |
{
|
| 58 |
"epoch": 0.59,
|
| 59 |
-
"learning_rate":
|
| 60 |
-
"loss": 1.
|
| 61 |
"step": 25
|
| 62 |
},
|
| 63 |
{
|
| 64 |
"epoch": 0.71,
|
| 65 |
-
"learning_rate":
|
| 66 |
-
"loss": 1.
|
| 67 |
"step": 30
|
| 68 |
},
|
| 69 |
{
|
| 70 |
"epoch": 0.71,
|
| 71 |
-
"eval_loss": 1.
|
| 72 |
-
"eval_runtime":
|
| 73 |
-
"eval_samples_per_second": 5.
|
| 74 |
-
"eval_steps_per_second": 1.
|
| 75 |
"step": 30
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.83,
|
| 79 |
-
"learning_rate":
|
| 80 |
-
"loss": 1.
|
| 81 |
"step": 35
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.94,
|
| 85 |
-
"learning_rate":
|
| 86 |
-
"loss": 1.
|
| 87 |
"step": 40
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.94,
|
| 91 |
-
"eval_loss": 1.
|
| 92 |
-
"eval_runtime":
|
| 93 |
-
"eval_samples_per_second": 5.
|
| 94 |
-
"eval_steps_per_second": 1.
|
| 95 |
"step": 40
|
| 96 |
},
|
| 97 |
{
|
| 98 |
-
"epoch":
|
| 99 |
-
"
|
| 100 |
-
"
|
| 101 |
-
"
|
| 102 |
-
|
| 103 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 104 |
"train_steps_per_second": 0.013
|
| 105 |
}
|
| 106 |
],
|
| 107 |
"logging_steps": 5,
|
| 108 |
-
"max_steps":
|
| 109 |
"num_input_tokens_seen": 0,
|
| 110 |
-
"num_train_epochs":
|
| 111 |
"save_steps": 20,
|
| 112 |
-
"total_flos":
|
| 113 |
"train_batch_size": 6,
|
| 114 |
"trial_name": null,
|
| 115 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 4.952830188679245,
|
| 5 |
"eval_steps": 10,
|
| 6 |
+
"global_step": 210,
|
| 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": 4.999720254525684e-05,
|
| 14 |
+
"loss": 1.458,
|
| 15 |
"step": 1
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.12,
|
| 19 |
+
"learning_rate": 4.9930094929529506e-05,
|
| 20 |
+
"loss": 1.4236,
|
| 21 |
"step": 5
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.24,
|
| 25 |
+
"learning_rate": 4.972077065562821e-05,
|
| 26 |
+
"loss": 1.3366,
|
| 27 |
"step": 10
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.24,
|
| 31 |
+
"eval_loss": 1.278298020362854,
|
| 32 |
+
"eval_runtime": 24.9289,
|
| 33 |
+
"eval_samples_per_second": 5.175,
|
| 34 |
+
"eval_steps_per_second": 1.725,
|
| 35 |
"step": 10
|
| 36 |
},
|
| 37 |
{
|
| 38 |
"epoch": 0.35,
|
| 39 |
+
"learning_rate": 4.937319780454559e-05,
|
| 40 |
+
"loss": 1.2828,
|
| 41 |
"step": 15
|
| 42 |
},
|
| 43 |
{
|
| 44 |
"epoch": 0.47,
|
| 45 |
+
"learning_rate": 4.888932014465352e-05,
|
| 46 |
+
"loss": 1.2563,
|
| 47 |
"step": 20
|
| 48 |
},
|
| 49 |
{
|
| 50 |
"epoch": 0.47,
|
| 51 |
+
"eval_loss": 1.2320669889450073,
|
| 52 |
+
"eval_runtime": 24.9408,
|
| 53 |
+
"eval_samples_per_second": 5.172,
|
| 54 |
+
"eval_steps_per_second": 1.724,
|
| 55 |
"step": 20
|
| 56 |
},
|
| 57 |
{
|
| 58 |
"epoch": 0.59,
|
| 59 |
+
"learning_rate": 4.827184371610511e-05,
|
| 60 |
+
"loss": 1.2308,
|
| 61 |
"step": 25
|
| 62 |
},
|
| 63 |
{
|
| 64 |
"epoch": 0.71,
|
| 65 |
+
"learning_rate": 4.752422169756048e-05,
|
| 66 |
+
"loss": 1.2289,
|
| 67 |
"step": 30
|
| 68 |
},
|
| 69 |
{
|
| 70 |
"epoch": 0.71,
|
| 71 |
+
"eval_loss": 1.2012428045272827,
|
| 72 |
+
"eval_runtime": 24.9502,
|
| 73 |
+
"eval_samples_per_second": 5.17,
|
| 74 |
+
"eval_steps_per_second": 1.723,
|
| 75 |
"step": 30
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.83,
|
| 79 |
+
"learning_rate": 4.665063509461097e-05,
|
| 80 |
+
"loss": 1.2051,
|
| 81 |
"step": 35
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.94,
|
| 85 |
+
"learning_rate": 4.5655969357899874e-05,
|
| 86 |
+
"loss": 1.1837,
|
| 87 |
"step": 40
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.94,
|
| 91 |
+
"eval_loss": 1.1688246726989746,
|
| 92 |
+
"eval_runtime": 24.9605,
|
| 93 |
+
"eval_samples_per_second": 5.168,
|
| 94 |
+
"eval_steps_per_second": 1.723,
|
| 95 |
"step": 40
|
| 96 |
},
|
| 97 |
{
|
| 98 |
+
"epoch": 1.06,
|
| 99 |
+
"learning_rate": 4.454578706170075e-05,
|
| 100 |
+
"loss": 1.1831,
|
| 101 |
+
"step": 45
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.18,
|
| 105 |
+
"learning_rate": 4.332629679574566e-05,
|
| 106 |
+
"loss": 1.1534,
|
| 107 |
+
"step": 50
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 1.18,
|
| 111 |
+
"eval_loss": 1.1306308507919312,
|
| 112 |
+
"eval_runtime": 24.9566,
|
| 113 |
+
"eval_samples_per_second": 5.169,
|
| 114 |
+
"eval_steps_per_second": 1.723,
|
| 115 |
+
"step": 50
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 1.3,
|
| 119 |
+
"learning_rate": 4.2004318444272985e-05,
|
| 120 |
+
"loss": 1.1413,
|
| 121 |
+
"step": 55
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 1.42,
|
| 125 |
+
"learning_rate": 4.058724504646834e-05,
|
| 126 |
+
"loss": 1.1254,
|
| 127 |
+
"step": 60
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 1.42,
|
| 131 |
+
"eval_loss": 1.1036521196365356,
|
| 132 |
+
"eval_runtime": 24.9587,
|
| 133 |
+
"eval_samples_per_second": 5.169,
|
| 134 |
+
"eval_steps_per_second": 1.723,
|
| 135 |
+
"step": 60
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 1.53,
|
| 139 |
+
"learning_rate": 3.908300145159055e-05,
|
| 140 |
+
"loss": 1.1167,
|
| 141 |
+
"step": 65
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 1.65,
|
| 145 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 146 |
+
"loss": 1.1011,
|
| 147 |
+
"step": 70
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 1.65,
|
| 151 |
+
"eval_loss": 1.0882474184036255,
|
| 152 |
+
"eval_runtime": 24.8955,
|
| 153 |
+
"eval_samples_per_second": 5.182,
|
| 154 |
+
"eval_steps_per_second": 1.727,
|
| 155 |
+
"step": 70
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 1.77,
|
| 159 |
+
"learning_rate": 3.5847093477938956e-05,
|
| 160 |
+
"loss": 1.0891,
|
| 161 |
+
"step": 75
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.89,
|
| 165 |
+
"learning_rate": 3.413352560915988e-05,
|
| 166 |
+
"loss": 1.0825,
|
| 167 |
+
"step": 80
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 1.89,
|
| 171 |
+
"eval_loss": 1.074791669845581,
|
| 172 |
+
"eval_runtime": 24.8917,
|
| 173 |
+
"eval_samples_per_second": 5.182,
|
| 174 |
+
"eval_steps_per_second": 1.727,
|
| 175 |
+
"step": 80
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 2.0,
|
| 179 |
+
"learning_rate": 3.2368879360272606e-05,
|
| 180 |
+
"loss": 1.0778,
|
| 181 |
+
"step": 85
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 2.12,
|
| 185 |
+
"learning_rate": 3.056302334890786e-05,
|
| 186 |
+
"loss": 1.0876,
|
| 187 |
+
"step": 90
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 2.12,
|
| 191 |
+
"eval_loss": 1.063535451889038,
|
| 192 |
+
"eval_runtime": 24.8907,
|
| 193 |
+
"eval_samples_per_second": 5.183,
|
| 194 |
+
"eval_steps_per_second": 1.728,
|
| 195 |
+
"step": 90
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 2.24,
|
| 199 |
+
"learning_rate": 2.872605665440436e-05,
|
| 200 |
+
"loss": 1.0835,
|
| 201 |
+
"step": 95
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 2.36,
|
| 205 |
+
"learning_rate": 2.686825233966061e-05,
|
| 206 |
+
"loss": 1.0716,
|
| 207 |
+
"step": 100
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 2.36,
|
| 211 |
+
"eval_loss": 1.0540255308151245,
|
| 212 |
+
"eval_runtime": 24.8912,
|
| 213 |
+
"eval_samples_per_second": 5.183,
|
| 214 |
+
"eval_steps_per_second": 1.728,
|
| 215 |
+
"step": 100
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 2.48,
|
| 219 |
+
"learning_rate": 2.5e-05,
|
| 220 |
+
"loss": 1.0767,
|
| 221 |
+
"step": 105
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 2.59,
|
| 225 |
+
"learning_rate": 2.3131747660339394e-05,
|
| 226 |
+
"loss": 1.0517,
|
| 227 |
+
"step": 110
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 2.59,
|
| 231 |
+
"eval_loss": 1.0458558797836304,
|
| 232 |
+
"eval_runtime": 24.9528,
|
| 233 |
+
"eval_samples_per_second": 5.17,
|
| 234 |
+
"eval_steps_per_second": 1.723,
|
| 235 |
+
"step": 110
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 2.71,
|
| 239 |
+
"learning_rate": 2.1273943345595637e-05,
|
| 240 |
+
"loss": 1.0564,
|
| 241 |
+
"step": 115
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 2.83,
|
| 245 |
+
"learning_rate": 1.9436976651092144e-05,
|
| 246 |
+
"loss": 1.0289,
|
| 247 |
+
"step": 120
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 2.83,
|
| 251 |
+
"eval_loss": 1.0388729572296143,
|
| 252 |
+
"eval_runtime": 24.9226,
|
| 253 |
+
"eval_samples_per_second": 5.176,
|
| 254 |
+
"eval_steps_per_second": 1.725,
|
| 255 |
+
"step": 120
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 2.95,
|
| 259 |
+
"learning_rate": 1.7631120639727393e-05,
|
| 260 |
+
"loss": 1.0318,
|
| 261 |
+
"step": 125
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 3.07,
|
| 265 |
+
"learning_rate": 1.5866474390840125e-05,
|
| 266 |
+
"loss": 1.0564,
|
| 267 |
+
"step": 130
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 3.07,
|
| 271 |
+
"eval_loss": 1.0331891775131226,
|
| 272 |
+
"eval_runtime": 24.9512,
|
| 273 |
+
"eval_samples_per_second": 5.17,
|
| 274 |
+
"eval_steps_per_second": 1.723,
|
| 275 |
+
"step": 130
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 3.18,
|
| 279 |
+
"learning_rate": 1.4152906522061048e-05,
|
| 280 |
+
"loss": 1.0499,
|
| 281 |
+
"step": 135
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 3.3,
|
| 285 |
+
"learning_rate": 1.2500000000000006e-05,
|
| 286 |
+
"loss": 1.034,
|
| 287 |
+
"step": 140
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 3.3,
|
| 291 |
+
"eval_loss": 1.0287576913833618,
|
| 292 |
+
"eval_runtime": 24.9512,
|
| 293 |
+
"eval_samples_per_second": 5.17,
|
| 294 |
+
"eval_steps_per_second": 1.723,
|
| 295 |
+
"step": 140
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 3.42,
|
| 299 |
+
"learning_rate": 1.0916998548409449e-05,
|
| 300 |
+
"loss": 1.0387,
|
| 301 |
+
"step": 145
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 3.54,
|
| 305 |
+
"learning_rate": 9.412754953531663e-06,
|
| 306 |
+
"loss": 1.0337,
|
| 307 |
+
"step": 150
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 3.54,
|
| 311 |
+
"eval_loss": 1.0253461599349976,
|
| 312 |
+
"eval_runtime": 24.8842,
|
| 313 |
+
"eval_samples_per_second": 5.184,
|
| 314 |
+
"eval_steps_per_second": 1.728,
|
| 315 |
+
"step": 150
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 3.66,
|
| 319 |
+
"learning_rate": 7.99568155572701e-06,
|
| 320 |
+
"loss": 1.0375,
|
| 321 |
+
"step": 155
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 3.77,
|
| 325 |
+
"learning_rate": 6.673703204254347e-06,
|
| 326 |
+
"loss": 1.033,
|
| 327 |
+
"step": 160
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 3.77,
|
| 331 |
+
"eval_loss": 1.0230636596679688,
|
| 332 |
+
"eval_runtime": 24.9446,
|
| 333 |
+
"eval_samples_per_second": 5.171,
|
| 334 |
+
"eval_steps_per_second": 1.724,
|
| 335 |
+
"step": 160
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 3.89,
|
| 339 |
+
"learning_rate": 5.454212938299255e-06,
|
| 340 |
+
"loss": 1.0292,
|
| 341 |
+
"step": 165
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 4.01,
|
| 345 |
+
"learning_rate": 4.344030642100133e-06,
|
| 346 |
+
"loss": 1.0312,
|
| 347 |
+
"step": 170
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 4.01,
|
| 351 |
+
"eval_loss": 1.021319031715393,
|
| 352 |
+
"eval_runtime": 24.9503,
|
| 353 |
+
"eval_samples_per_second": 5.17,
|
| 354 |
+
"eval_steps_per_second": 1.723,
|
| 355 |
+
"step": 170
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 4.13,
|
| 359 |
+
"learning_rate": 3.3493649053890326e-06,
|
| 360 |
+
"loss": 1.0343,
|
| 361 |
+
"step": 175
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 4.25,
|
| 365 |
+
"learning_rate": 2.475778302439524e-06,
|
| 366 |
+
"loss": 1.0207,
|
| 367 |
+
"step": 180
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 4.25,
|
| 371 |
+
"eval_loss": 1.0203514099121094,
|
| 372 |
+
"eval_runtime": 24.9557,
|
| 373 |
+
"eval_samples_per_second": 5.169,
|
| 374 |
+
"eval_steps_per_second": 1.723,
|
| 375 |
+
"step": 180
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 4.36,
|
| 379 |
+
"learning_rate": 1.7281562838948966e-06,
|
| 380 |
+
"loss": 1.0413,
|
| 381 |
+
"step": 185
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 4.48,
|
| 385 |
+
"learning_rate": 1.1106798553464804e-06,
|
| 386 |
+
"loss": 1.0271,
|
| 387 |
+
"step": 190
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 4.48,
|
| 391 |
+
"eval_loss": 1.019814133644104,
|
| 392 |
+
"eval_runtime": 24.9781,
|
| 393 |
+
"eval_samples_per_second": 5.165,
|
| 394 |
+
"eval_steps_per_second": 1.722,
|
| 395 |
+
"step": 190
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 4.6,
|
| 399 |
+
"learning_rate": 6.268021954544096e-07,
|
| 400 |
+
"loss": 1.0233,
|
| 401 |
+
"step": 195
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 4.72,
|
| 405 |
+
"learning_rate": 2.7922934437178695e-07,
|
| 406 |
+
"loss": 1.0351,
|
| 407 |
+
"step": 200
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 4.72,
|
| 411 |
+
"eval_loss": 1.0197088718414307,
|
| 412 |
+
"eval_runtime": 24.9625,
|
| 413 |
+
"eval_samples_per_second": 5.168,
|
| 414 |
+
"eval_steps_per_second": 1.723,
|
| 415 |
+
"step": 200
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 4.83,
|
| 419 |
+
"learning_rate": 6.990507047049676e-08,
|
| 420 |
+
"loss": 1.015,
|
| 421 |
+
"step": 205
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"epoch": 4.95,
|
| 425 |
+
"learning_rate": 0.0,
|
| 426 |
+
"loss": 1.0339,
|
| 427 |
+
"step": 210
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 4.95,
|
| 431 |
+
"eval_loss": 1.0196079015731812,
|
| 432 |
+
"eval_runtime": 24.9619,
|
| 433 |
+
"eval_samples_per_second": 5.168,
|
| 434 |
+
"eval_steps_per_second": 1.723,
|
| 435 |
+
"step": 210
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 4.95,
|
| 439 |
+
"step": 210,
|
| 440 |
+
"total_flos": 3.221775454641848e+17,
|
| 441 |
+
"train_loss": 1.1020877304531278,
|
| 442 |
+
"train_runtime": 16139.1818,
|
| 443 |
+
"train_samples_per_second": 1.575,
|
| 444 |
"train_steps_per_second": 0.013
|
| 445 |
}
|
| 446 |
],
|
| 447 |
"logging_steps": 5,
|
| 448 |
+
"max_steps": 210,
|
| 449 |
"num_input_tokens_seen": 0,
|
| 450 |
+
"num_train_epochs": 5,
|
| 451 |
"save_steps": 20,
|
| 452 |
+
"total_flos": 3.221775454641848e+17,
|
| 453 |
"train_batch_size": 6,
|
| 454 |
"trial_name": null,
|
| 455 |
"trial_params": null
|