Qin Liu
commited on
Commit
•
acc2564
1
Parent(s):
6dca8e1
Model save
Browse files- README.md +13 -9
- all_results.json +5 -5
- runs/Aug21_02-51-44_ip-172-31-10-237/events.out.tfevents.1724208715.ip-172-31-10-237.960579.0 +2 -2
- train_results.json +5 -5
- trainer_state.json +654 -115
README.md
CHANGED
|
@@ -1,11 +1,8 @@
|
|
| 1 |
---
|
| 2 |
base_model: meta-llama/Meta-Llama-3-8B
|
| 3 |
-
datasets:
|
| 4 |
-
- HuggingFaceH4/ultrachat_200k
|
| 5 |
library_name: peft
|
| 6 |
license: llama3
|
| 7 |
tags:
|
| 8 |
-
- alignment-handbook
|
| 9 |
- trl
|
| 10 |
- sft
|
| 11 |
- generated_from_trainer
|
|
@@ -19,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 19 |
|
| 20 |
# llama3-sudo-sanity
|
| 21 |
|
| 22 |
-
This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
-
- Loss: 1.
|
| 25 |
|
| 26 |
## Model description
|
| 27 |
|
|
@@ -52,15 +49,22 @@ The following hyperparameters were used during training:
|
|
| 52 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 53 |
- lr_scheduler_type: cosine
|
| 54 |
- lr_scheduler_warmup_ratio: 0.1
|
| 55 |
-
- num_epochs:
|
| 56 |
|
| 57 |
### Training results
|
| 58 |
|
| 59 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 60 |
|:-------------:|:------:|:----:|:---------------:|
|
| 61 |
-
| 1.
|
| 62 |
-
| 1.
|
| 63 |
-
| 1.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 64 |
|
| 65 |
|
| 66 |
### Framework versions
|
|
|
|
| 1 |
---
|
| 2 |
base_model: meta-llama/Meta-Llama-3-8B
|
|
|
|
|
|
|
| 3 |
library_name: peft
|
| 4 |
license: llama3
|
| 5 |
tags:
|
|
|
|
| 6 |
- trl
|
| 7 |
- sft
|
| 8 |
- generated_from_trainer
|
|
|
|
| 16 |
|
| 17 |
# llama3-sudo-sanity
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on an unknown dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 1.1030
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
|
|
| 49 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 50 |
- lr_scheduler_type: cosine
|
| 51 |
- lr_scheduler_warmup_ratio: 0.1
|
| 52 |
+
- num_epochs: 10
|
| 53 |
|
| 54 |
### Training results
|
| 55 |
|
| 56 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 57 |
|:-------------:|:------:|:----:|:---------------:|
|
| 58 |
+
| 1.8735 | 0.9899 | 49 | 1.8325 |
|
| 59 |
+
| 1.8231 | 2.0 | 99 | 1.7239 |
|
| 60 |
+
| 1.7516 | 2.9899 | 148 | 1.6330 |
|
| 61 |
+
| 1.6586 | 4.0 | 198 | 1.5280 |
|
| 62 |
+
| 1.5571 | 4.9899 | 247 | 1.4166 |
|
| 63 |
+
| 1.4677 | 6.0 | 297 | 1.3068 |
|
| 64 |
+
| 1.3422 | 6.9899 | 346 | 1.2082 |
|
| 65 |
+
| 1.2609 | 8.0 | 396 | 1.1378 |
|
| 66 |
+
| 1.1647 | 8.9899 | 445 | 1.1074 |
|
| 67 |
+
| 1.1571 | 9.8990 | 490 | 1.1030 |
|
| 68 |
|
| 69 |
|
| 70 |
### Framework versions
|
all_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 1.
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 6321,
|
| 7 |
-
"train_samples_per_second": 11.
|
| 8 |
"train_steps_per_second": 0.091
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 9.8989898989899,
|
| 3 |
+
"total_flos": 2344635780825088.0,
|
| 4 |
+
"train_loss": 1.5290005391957808,
|
| 5 |
+
"train_runtime": 5413.3076,
|
| 6 |
"train_samples": 6321,
|
| 7 |
+
"train_samples_per_second": 11.677,
|
| 8 |
"train_steps_per_second": 0.091
|
| 9 |
}
|
runs/Aug21_02-51-44_ip-172-31-10-237/events.out.tfevents.1724208715.ip-172-31-10-237.960579.0
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 sha256:658ebebff6d9c7b4b26c19b8f53502f8bcef7731700fb8153bb4498d3dfd5fb4
|
| 3 |
+
size 29418
|
train_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 1.
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 6321,
|
| 7 |
-
"train_samples_per_second": 11.
|
| 8 |
"train_steps_per_second": 0.091
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 9.8989898989899,
|
| 3 |
+
"total_flos": 2344635780825088.0,
|
| 4 |
+
"train_loss": 1.5290005391957808,
|
| 5 |
+
"train_runtime": 5413.3076,
|
| 6 |
"train_samples": 6321,
|
| 7 |
+
"train_samples_per_second": 11.677,
|
| 8 |
"train_steps_per_second": 0.091
|
| 9 |
}
|
trainer_state.json
CHANGED
|
@@ -1,261 +1,800 @@
|
|
| 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": 0.020202020202020204,
|
| 13 |
-
"grad_norm": 1.
|
| 14 |
-
"learning_rate":
|
| 15 |
-
"loss": 2.
|
| 16 |
"step": 1
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.10101010101010101,
|
| 20 |
-
"grad_norm": 1.
|
| 21 |
-
"learning_rate":
|
| 22 |
-
"loss": 2.
|
| 23 |
"step": 5
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.20202020202020202,
|
| 27 |
-
"grad_norm":
|
| 28 |
-
"learning_rate":
|
| 29 |
-
"loss": 2.
|
| 30 |
"step": 10
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.30303030303030304,
|
| 34 |
-
"grad_norm": 0.
|
| 35 |
-
"learning_rate":
|
| 36 |
-
"loss": 2.
|
| 37 |
"step": 15
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.40404040404040403,
|
| 41 |
-
"grad_norm": 0.
|
| 42 |
-
"learning_rate":
|
| 43 |
-
"loss": 2.
|
| 44 |
"step": 20
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.5050505050505051,
|
| 48 |
-
"grad_norm": 0.
|
| 49 |
-
"learning_rate": 0.
|
| 50 |
-
"loss":
|
| 51 |
"step": 25
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.6060606060606061,
|
| 55 |
-
"grad_norm": 0.
|
| 56 |
-
"learning_rate": 0.
|
| 57 |
-
"loss":
|
| 58 |
"step": 30
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 0.7070707070707071,
|
| 62 |
-
"grad_norm": 0.
|
| 63 |
-
"learning_rate": 0.
|
| 64 |
-
"loss":
|
| 65 |
"step": 35
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 0.8080808080808081,
|
| 69 |
-
"grad_norm": 0.
|
| 70 |
-
"learning_rate": 0.
|
| 71 |
-
"loss": 1.
|
| 72 |
"step": 40
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"epoch": 0.9090909090909091,
|
| 76 |
-
"grad_norm": 0.
|
| 77 |
-
"learning_rate": 0.
|
| 78 |
-
"loss": 1.
|
| 79 |
"step": 45
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"epoch": 0.98989898989899,
|
| 83 |
-
"eval_loss": 1.
|
| 84 |
-
"eval_runtime":
|
| 85 |
-
"eval_samples_per_second":
|
| 86 |
-
"eval_steps_per_second": 2.
|
| 87 |
"step": 49
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 1.0101010101010102,
|
| 91 |
-
"grad_norm": 0.
|
| 92 |
-
"learning_rate": 0.
|
| 93 |
-
"loss": 1.
|
| 94 |
"step": 50
|
| 95 |
},
|
| 96 |
{
|
| 97 |
"epoch": 1.1111111111111112,
|
| 98 |
-
"grad_norm": 0.
|
| 99 |
-
"learning_rate": 0.
|
| 100 |
-
"loss": 1.
|
| 101 |
"step": 55
|
| 102 |
},
|
| 103 |
{
|
| 104 |
"epoch": 1.2121212121212122,
|
| 105 |
-
"grad_norm": 0.
|
| 106 |
-
"learning_rate": 0.
|
| 107 |
-
"loss": 1.
|
| 108 |
"step": 60
|
| 109 |
},
|
| 110 |
{
|
| 111 |
"epoch": 1.3131313131313131,
|
| 112 |
-
"grad_norm": 0.
|
| 113 |
-
"learning_rate": 0.
|
| 114 |
-
"loss": 1.
|
| 115 |
"step": 65
|
| 116 |
},
|
| 117 |
{
|
| 118 |
"epoch": 1.4141414141414141,
|
| 119 |
-
"grad_norm": 0.
|
| 120 |
-
"learning_rate": 0.
|
| 121 |
-
"loss": 1.
|
| 122 |
"step": 70
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 1.5151515151515151,
|
| 126 |
-
"grad_norm": 0.
|
| 127 |
-
"learning_rate": 0.
|
| 128 |
-
"loss": 1.
|
| 129 |
"step": 75
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 1.6161616161616161,
|
| 133 |
-
"grad_norm": 0.
|
| 134 |
-
"learning_rate": 0.
|
| 135 |
-
"loss": 1.
|
| 136 |
"step": 80
|
| 137 |
},
|
| 138 |
{
|
| 139 |
"epoch": 1.7171717171717171,
|
| 140 |
-
"grad_norm": 0.
|
| 141 |
-
"learning_rate":
|
| 142 |
-
"loss": 1.
|
| 143 |
"step": 85
|
| 144 |
},
|
| 145 |
{
|
| 146 |
"epoch": 1.8181818181818183,
|
| 147 |
-
"grad_norm": 0.
|
| 148 |
-
"learning_rate":
|
| 149 |
-
"loss": 1.
|
| 150 |
"step": 90
|
| 151 |
},
|
| 152 |
{
|
| 153 |
"epoch": 1.9191919191919191,
|
| 154 |
-
"grad_norm": 0.
|
| 155 |
-
"learning_rate":
|
| 156 |
-
"loss": 1.
|
| 157 |
"step": 95
|
| 158 |
},
|
| 159 |
{
|
| 160 |
"epoch": 2.0,
|
| 161 |
-
"eval_loss": 1.
|
| 162 |
-
"eval_runtime":
|
| 163 |
-
"eval_samples_per_second":
|
| 164 |
-
"eval_steps_per_second": 2.
|
| 165 |
"step": 99
|
| 166 |
},
|
| 167 |
{
|
| 168 |
"epoch": 2.0202020202020203,
|
| 169 |
-
"grad_norm": 0.
|
| 170 |
-
"learning_rate":
|
| 171 |
-
"loss": 1.
|
| 172 |
"step": 100
|
| 173 |
},
|
| 174 |
{
|
| 175 |
"epoch": 2.121212121212121,
|
| 176 |
-
"grad_norm": 0.
|
| 177 |
-
"learning_rate":
|
| 178 |
-
"loss": 1.
|
| 179 |
"step": 105
|
| 180 |
},
|
| 181 |
{
|
| 182 |
"epoch": 2.2222222222222223,
|
| 183 |
-
"grad_norm": 0.
|
| 184 |
-
"learning_rate":
|
| 185 |
-
"loss": 1.
|
| 186 |
"step": 110
|
| 187 |
},
|
| 188 |
{
|
| 189 |
"epoch": 2.323232323232323,
|
| 190 |
-
"grad_norm": 0.
|
| 191 |
-
"learning_rate":
|
| 192 |
-
"loss": 1.
|
| 193 |
"step": 115
|
| 194 |
},
|
| 195 |
{
|
| 196 |
"epoch": 2.4242424242424243,
|
| 197 |
-
"grad_norm": 0.
|
| 198 |
-
"learning_rate":
|
| 199 |
-
"loss": 1.
|
| 200 |
"step": 120
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 2.525252525252525,
|
| 204 |
-
"grad_norm": 0.
|
| 205 |
-
"learning_rate":
|
| 206 |
-
"loss": 1.
|
| 207 |
"step": 125
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"epoch": 2.6262626262626263,
|
| 211 |
-
"grad_norm": 0.
|
| 212 |
-
"learning_rate":
|
| 213 |
-
"loss": 1.
|
| 214 |
"step": 130
|
| 215 |
},
|
| 216 |
{
|
| 217 |
"epoch": 2.7272727272727275,
|
| 218 |
-
"grad_norm": 0.
|
| 219 |
-
"learning_rate":
|
| 220 |
-
"loss": 1.
|
| 221 |
"step": 135
|
| 222 |
},
|
| 223 |
{
|
| 224 |
"epoch": 2.8282828282828283,
|
| 225 |
-
"grad_norm": 0.
|
| 226 |
-
"learning_rate":
|
| 227 |
-
"loss": 1.
|
| 228 |
"step": 140
|
| 229 |
},
|
| 230 |
{
|
| 231 |
"epoch": 2.929292929292929,
|
| 232 |
-
"grad_norm": 0.
|
| 233 |
-
"learning_rate":
|
| 234 |
-
"loss": 1.
|
| 235 |
"step": 145
|
| 236 |
},
|
| 237 |
{
|
| 238 |
-
"epoch": 2.
|
| 239 |
-
"eval_loss": 1.
|
| 240 |
-
"eval_runtime":
|
| 241 |
-
"eval_samples_per_second":
|
| 242 |
-
"eval_steps_per_second": 2.
|
| 243 |
-
"step":
|
| 244 |
},
|
| 245 |
{
|
| 246 |
-
"epoch":
|
| 247 |
-
"
|
| 248 |
-
"
|
| 249 |
-
"
|
| 250 |
-
"
|
| 251 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 252 |
"train_steps_per_second": 0.091
|
| 253 |
}
|
| 254 |
],
|
| 255 |
"logging_steps": 5,
|
| 256 |
-
"max_steps":
|
| 257 |
"num_input_tokens_seen": 0,
|
| 258 |
-
"num_train_epochs":
|
| 259 |
"save_steps": 25,
|
| 260 |
"stateful_callbacks": {
|
| 261 |
"TrainerControl": {
|
|
@@ -269,7 +808,7 @@
|
|
| 269 |
"attributes": {}
|
| 270 |
}
|
| 271 |
},
|
| 272 |
-
"total_flos":
|
| 273 |
"train_batch_size": 8,
|
| 274 |
"trial_name": null,
|
| 275 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 9.8989898989899,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 490,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.020202020202020204,
|
| 13 |
+
"grad_norm": 1.1391378054420909,
|
| 14 |
+
"learning_rate": 4.081632653061224e-06,
|
| 15 |
+
"loss": 2.5995,
|
| 16 |
"step": 1
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.10101010101010101,
|
| 20 |
+
"grad_norm": 1.006731604503432,
|
| 21 |
+
"learning_rate": 2.0408163265306123e-05,
|
| 22 |
+
"loss": 2.5925,
|
| 23 |
"step": 5
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.20202020202020202,
|
| 27 |
+
"grad_norm": 1.3965084950072466,
|
| 28 |
+
"learning_rate": 4.0816326530612245e-05,
|
| 29 |
+
"loss": 2.546,
|
| 30 |
"step": 10
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.30303030303030304,
|
| 34 |
+
"grad_norm": 0.59095362852847,
|
| 35 |
+
"learning_rate": 6.122448979591838e-05,
|
| 36 |
+
"loss": 2.396,
|
| 37 |
"step": 15
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.40404040404040403,
|
| 41 |
+
"grad_norm": 0.33359551466461584,
|
| 42 |
+
"learning_rate": 8.163265306122449e-05,
|
| 43 |
+
"loss": 2.2744,
|
| 44 |
"step": 20
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.5050505050505051,
|
| 48 |
+
"grad_norm": 0.3767673243983956,
|
| 49 |
+
"learning_rate": 0.00010204081632653062,
|
| 50 |
+
"loss": 2.1608,
|
| 51 |
"step": 25
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.6060606060606061,
|
| 55 |
+
"grad_norm": 0.3530777092096336,
|
| 56 |
+
"learning_rate": 0.00012244897959183676,
|
| 57 |
+
"loss": 2.0261,
|
| 58 |
"step": 30
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 0.7070707070707071,
|
| 62 |
+
"grad_norm": 0.36168305575388426,
|
| 63 |
+
"learning_rate": 0.00014285714285714287,
|
| 64 |
+
"loss": 2.0091,
|
| 65 |
"step": 35
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 0.8080808080808081,
|
| 69 |
+
"grad_norm": 0.2764545304686734,
|
| 70 |
+
"learning_rate": 0.00016326530612244898,
|
| 71 |
+
"loss": 1.9434,
|
| 72 |
"step": 40
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"epoch": 0.9090909090909091,
|
| 76 |
+
"grad_norm": 0.2653033039849191,
|
| 77 |
+
"learning_rate": 0.00018367346938775512,
|
| 78 |
+
"loss": 1.8735,
|
| 79 |
"step": 45
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"epoch": 0.98989898989899,
|
| 83 |
+
"eval_loss": 1.8325327634811401,
|
| 84 |
+
"eval_runtime": 177.2293,
|
| 85 |
+
"eval_samples_per_second": 35.666,
|
| 86 |
+
"eval_steps_per_second": 2.234,
|
| 87 |
"step": 49
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 1.0101010101010102,
|
| 91 |
+
"grad_norm": 0.2170406034599688,
|
| 92 |
+
"learning_rate": 0.00019999746258949147,
|
| 93 |
+
"loss": 1.8679,
|
| 94 |
"step": 50
|
| 95 |
},
|
| 96 |
{
|
| 97 |
"epoch": 1.1111111111111112,
|
| 98 |
+
"grad_norm": 0.2414932638214735,
|
| 99 |
+
"learning_rate": 0.00019990866674170983,
|
| 100 |
+
"loss": 1.8705,
|
| 101 |
"step": 55
|
| 102 |
},
|
| 103 |
{
|
| 104 |
"epoch": 1.2121212121212122,
|
| 105 |
+
"grad_norm": 0.280526285390802,
|
| 106 |
+
"learning_rate": 0.00019969312910817183,
|
| 107 |
+
"loss": 1.8428,
|
| 108 |
"step": 60
|
| 109 |
},
|
| 110 |
{
|
| 111 |
"epoch": 1.3131313131313131,
|
| 112 |
+
"grad_norm": 0.5716516166032067,
|
| 113 |
+
"learning_rate": 0.000199351123114852,
|
| 114 |
+
"loss": 1.8267,
|
| 115 |
"step": 65
|
| 116 |
},
|
| 117 |
{
|
| 118 |
"epoch": 1.4141414141414141,
|
| 119 |
+
"grad_norm": 0.22548657696350605,
|
| 120 |
+
"learning_rate": 0.00019888308262251285,
|
| 121 |
+
"loss": 1.7959,
|
| 122 |
"step": 70
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 1.5151515151515151,
|
| 126 |
+
"grad_norm": 0.2400943754848952,
|
| 127 |
+
"learning_rate": 0.00019828960137631928,
|
| 128 |
+
"loss": 1.8328,
|
| 129 |
"step": 75
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 1.6161616161616161,
|
| 133 |
+
"grad_norm": 0.24338845667140263,
|
| 134 |
+
"learning_rate": 0.00019757143225262728,
|
| 135 |
+
"loss": 1.8287,
|
| 136 |
"step": 80
|
| 137 |
},
|
| 138 |
{
|
| 139 |
"epoch": 1.7171717171717171,
|
| 140 |
+
"grad_norm": 0.23391106368008455,
|
| 141 |
+
"learning_rate": 0.00019672948630390294,
|
| 142 |
+
"loss": 1.8345,
|
| 143 |
"step": 85
|
| 144 |
},
|
| 145 |
{
|
| 146 |
"epoch": 1.8181818181818183,
|
| 147 |
+
"grad_norm": 0.27223367547235244,
|
| 148 |
+
"learning_rate": 0.00019576483160298246,
|
| 149 |
+
"loss": 1.7731,
|
| 150 |
"step": 90
|
| 151 |
},
|
| 152 |
{
|
| 153 |
"epoch": 1.9191919191919191,
|
| 154 |
+
"grad_norm": 0.25522514087403636,
|
| 155 |
+
"learning_rate": 0.00019467869188814023,
|
| 156 |
+
"loss": 1.8231,
|
| 157 |
"step": 95
|
| 158 |
},
|
| 159 |
{
|
| 160 |
"epoch": 2.0,
|
| 161 |
+
"eval_loss": 1.7238675355911255,
|
| 162 |
+
"eval_runtime": 175.3209,
|
| 163 |
+
"eval_samples_per_second": 36.054,
|
| 164 |
+
"eval_steps_per_second": 2.259,
|
| 165 |
"step": 99
|
| 166 |
},
|
| 167 |
{
|
| 168 |
"epoch": 2.0202020202020203,
|
| 169 |
+
"grad_norm": 0.24156163537353126,
|
| 170 |
+
"learning_rate": 0.00019347244501068312,
|
| 171 |
+
"loss": 1.8199,
|
| 172 |
"step": 100
|
| 173 |
},
|
| 174 |
{
|
| 175 |
"epoch": 2.121212121212121,
|
| 176 |
+
"grad_norm": 0.26738911600323706,
|
| 177 |
+
"learning_rate": 0.00019214762118704076,
|
| 178 |
+
"loss": 1.7554,
|
| 179 |
"step": 105
|
| 180 |
},
|
| 181 |
{
|
| 182 |
"epoch": 2.2222222222222223,
|
| 183 |
+
"grad_norm": 0.2974142479913874,
|
| 184 |
+
"learning_rate": 0.000190705901057569,
|
| 185 |
+
"loss": 1.7693,
|
| 186 |
"step": 110
|
| 187 |
},
|
| 188 |
{
|
| 189 |
"epoch": 2.323232323232323,
|
| 190 |
+
"grad_norm": 0.32489025521491743,
|
| 191 |
+
"learning_rate": 0.00018914911355452895,
|
| 192 |
+
"loss": 1.7036,
|
| 193 |
"step": 115
|
| 194 |
},
|
| 195 |
{
|
| 196 |
"epoch": 2.4242424242424243,
|
| 197 |
+
"grad_norm": 0.34564309759210704,
|
| 198 |
+
"learning_rate": 0.00018747923358194662,
|
| 199 |
+
"loss": 1.7449,
|
| 200 |
"step": 120
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 2.525252525252525,
|
| 204 |
+
"grad_norm": 0.33251849318291843,
|
| 205 |
+
"learning_rate": 0.00018569837951029595,
|
| 206 |
+
"loss": 1.7556,
|
| 207 |
"step": 125
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"epoch": 2.6262626262626263,
|
| 211 |
+
"grad_norm": 0.33565250510381917,
|
| 212 |
+
"learning_rate": 0.00018380881048918405,
|
| 213 |
+
"loss": 1.744,
|
| 214 |
"step": 130
|
| 215 |
},
|
| 216 |
{
|
| 217 |
"epoch": 2.7272727272727275,
|
| 218 |
+
"grad_norm": 0.3427001935365706,
|
| 219 |
+
"learning_rate": 0.00018181292358144703,
|
| 220 |
+
"loss": 1.7234,
|
| 221 |
"step": 135
|
| 222 |
},
|
| 223 |
{
|
| 224 |
"epoch": 2.8282828282828283,
|
| 225 |
+
"grad_norm": 0.32846296076937864,
|
| 226 |
+
"learning_rate": 0.00017971325072229226,
|
| 227 |
+
"loss": 1.7274,
|
| 228 |
"step": 140
|
| 229 |
},
|
| 230 |
{
|
| 231 |
"epoch": 2.929292929292929,
|
| 232 |
+
"grad_norm": 0.3446715050022125,
|
| 233 |
+
"learning_rate": 0.0001775124555073452,
|
| 234 |
+
"loss": 1.7516,
|
| 235 |
"step": 145
|
| 236 |
},
|
| 237 |
{
|
| 238 |
+
"epoch": 2.98989898989899,
|
| 239 |
+
"eval_loss": 1.6329888105392456,
|
| 240 |
+
"eval_runtime": 174.9738,
|
| 241 |
+
"eval_samples_per_second": 36.125,
|
| 242 |
+
"eval_steps_per_second": 2.263,
|
| 243 |
+
"step": 148
|
| 244 |
},
|
| 245 |
{
|
| 246 |
+
"epoch": 3.0303030303030303,
|
| 247 |
+
"grad_norm": 0.34753378836184123,
|
| 248 |
+
"learning_rate": 0.0001752133298136744,
|
| 249 |
+
"loss": 1.7442,
|
| 250 |
+
"step": 150
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 3.1313131313131315,
|
| 254 |
+
"grad_norm": 0.3899145091665638,
|
| 255 |
+
"learning_rate": 0.0001728187902580819,
|
| 256 |
+
"loss": 1.6414,
|
| 257 |
+
"step": 155
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 3.2323232323232323,
|
| 261 |
+
"grad_norm": 0.3969944429695798,
|
| 262 |
+
"learning_rate": 0.00017033187449715196,
|
| 263 |
+
"loss": 1.6411,
|
| 264 |
+
"step": 160
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 3.3333333333333335,
|
| 268 |
+
"grad_norm": 0.4463802224093316,
|
| 269 |
+
"learning_rate": 0.00016775573737375096,
|
| 270 |
+
"loss": 1.6955,
|
| 271 |
+
"step": 165
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 3.4343434343434343,
|
| 275 |
+
"grad_norm": 0.4873799041554826,
|
| 276 |
+
"learning_rate": 0.0001650936469148681,
|
| 277 |
+
"loss": 1.6812,
|
| 278 |
+
"step": 170
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 3.5353535353535355,
|
| 282 |
+
"grad_norm": 0.5138644486787001,
|
| 283 |
+
"learning_rate": 0.00016234898018587337,
|
| 284 |
+
"loss": 1.6455,
|
| 285 |
+
"step": 175
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 3.6363636363636362,
|
| 289 |
+
"grad_norm": 0.4441989179255284,
|
| 290 |
+
"learning_rate": 0.00015952521900645144,
|
| 291 |
+
"loss": 1.6537,
|
| 292 |
+
"step": 180
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 3.7373737373737375,
|
| 296 |
+
"grad_norm": 0.45397642246696135,
|
| 297 |
+
"learning_rate": 0.0001566259455336474,
|
| 298 |
+
"loss": 1.6384,
|
| 299 |
+
"step": 185
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 3.8383838383838382,
|
| 303 |
+
"grad_norm": 0.48522658034874977,
|
| 304 |
+
"learning_rate": 0.0001536548377176263,
|
| 305 |
+
"loss": 1.6292,
|
| 306 |
+
"step": 190
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 3.9393939393939394,
|
| 310 |
+
"grad_norm": 0.43244857762556244,
|
| 311 |
+
"learning_rate": 0.0001506156646359123,
|
| 312 |
+
"loss": 1.6586,
|
| 313 |
+
"step": 195
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.0,
|
| 317 |
+
"eval_loss": 1.5279655456542969,
|
| 318 |
+
"eval_runtime": 175.1053,
|
| 319 |
+
"eval_samples_per_second": 36.098,
|
| 320 |
+
"eval_steps_per_second": 2.261,
|
| 321 |
+
"step": 198
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 4.040404040404041,
|
| 325 |
+
"grad_norm": 0.48231588190167396,
|
| 326 |
+
"learning_rate": 0.0001475122817120253,
|
| 327 |
+
"loss": 1.6137,
|
| 328 |
+
"step": 200
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 4.141414141414141,
|
| 332 |
+
"grad_norm": 0.5842989060014684,
|
| 333 |
+
"learning_rate": 0.00014434862582458135,
|
| 334 |
+
"loss": 1.5082,
|
| 335 |
+
"step": 205
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 4.242424242424242,
|
| 339 |
+
"grad_norm": 0.5870996767325578,
|
| 340 |
+
"learning_rate": 0.00014112871031306119,
|
| 341 |
+
"loss": 1.5382,
|
| 342 |
+
"step": 210
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 4.343434343434343,
|
| 346 |
+
"grad_norm": 0.6294490520638103,
|
| 347 |
+
"learning_rate": 0.0001378566198865818,
|
| 348 |
+
"loss": 1.5738,
|
| 349 |
+
"step": 215
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 4.444444444444445,
|
| 353 |
+
"grad_norm": 0.6361554344671604,
|
| 354 |
+
"learning_rate": 0.00013453650544213076,
|
| 355 |
+
"loss": 1.5609,
|
| 356 |
+
"step": 220
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 4.545454545454545,
|
| 360 |
+
"grad_norm": 0.5845910737228225,
|
| 361 |
+
"learning_rate": 0.00013117257879883583,
|
| 362 |
+
"loss": 1.5832,
|
| 363 |
+
"step": 225
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 4.646464646464646,
|
| 367 |
+
"grad_norm": 0.6362570401491278,
|
| 368 |
+
"learning_rate": 0.00012776910735495003,
|
| 369 |
+
"loss": 1.5386,
|
| 370 |
+
"step": 230
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 4.747474747474747,
|
| 374 |
+
"grad_norm": 0.6079381787055775,
|
| 375 |
+
"learning_rate": 0.0001243304086743309,
|
| 376 |
+
"loss": 1.5408,
|
| 377 |
+
"step": 235
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 4.848484848484849,
|
| 381 |
+
"grad_norm": 0.5955494164961348,
|
| 382 |
+
"learning_rate": 0.0001208608450092801,
|
| 383 |
+
"loss": 1.5767,
|
| 384 |
+
"step": 240
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 4.94949494949495,
|
| 388 |
+
"grad_norm": 0.5941973746172844,
|
| 389 |
+
"learning_rate": 0.00011736481776669306,
|
| 390 |
+
"loss": 1.5571,
|
| 391 |
+
"step": 245
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 4.98989898989899,
|
| 395 |
+
"eval_loss": 1.4166467189788818,
|
| 396 |
+
"eval_runtime": 174.6193,
|
| 397 |
+
"eval_samples_per_second": 36.199,
|
| 398 |
+
"eval_steps_per_second": 2.268,
|
| 399 |
+
"step": 247
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 5.05050505050505,
|
| 403 |
+
"grad_norm": 0.6955112160268645,
|
| 404 |
+
"learning_rate": 0.0001138467619245374,
|
| 405 |
+
"loss": 1.5011,
|
| 406 |
+
"step": 250
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 5.151515151515151,
|
| 410 |
+
"grad_norm": 0.7116916227562953,
|
| 411 |
+
"learning_rate": 0.00011031114040574437,
|
| 412 |
+
"loss": 1.4537,
|
| 413 |
+
"step": 255
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 5.252525252525253,
|
| 417 |
+
"grad_norm": 0.8295579161716972,
|
| 418 |
+
"learning_rate": 0.0001067624384166495,
|
| 419 |
+
"loss": 1.398,
|
| 420 |
+
"step": 260
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 5.353535353535354,
|
| 424 |
+
"grad_norm": 0.7415551092257379,
|
| 425 |
+
"learning_rate": 0.00010320515775716555,
|
| 426 |
+
"loss": 1.4474,
|
| 427 |
+
"step": 265
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 5.454545454545454,
|
| 431 |
+
"grad_norm": 0.7957507416152227,
|
| 432 |
+
"learning_rate": 9.96438111099047e-05,
|
| 433 |
+
"loss": 1.4459,
|
| 434 |
+
"step": 270
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 5.555555555555555,
|
| 438 |
+
"grad_norm": 0.8098108632452509,
|
| 439 |
+
"learning_rate": 9.608291631549574e-05,
|
| 440 |
+
"loss": 1.4266,
|
| 441 |
+
"step": 275
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 5.656565656565657,
|
| 445 |
+
"grad_norm": 0.8498743613190896,
|
| 446 |
+
"learning_rate": 9.252699064135758e-05,
|
| 447 |
+
"loss": 1.3931,
|
| 448 |
+
"step": 280
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 5.757575757575758,
|
| 452 |
+
"grad_norm": 0.7968761297367668,
|
| 453 |
+
"learning_rate": 8.898054505119989e-05,
|
| 454 |
+
"loss": 1.4628,
|
| 455 |
+
"step": 285
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 5.858585858585858,
|
| 459 |
+
"grad_norm": 0.8166566096084199,
|
| 460 |
+
"learning_rate": 8.54480784825207e-05,
|
| 461 |
+
"loss": 1.4777,
|
| 462 |
+
"step": 290
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 5.959595959595959,
|
| 466 |
+
"grad_norm": 0.7564583944169918,
|
| 467 |
+
"learning_rate": 8.193407213936012e-05,
|
| 468 |
+
"loss": 1.4677,
|
| 469 |
+
"step": 295
|
| 470 |
+
},
|
| 471 |
+
{
|
| 472 |
+
"epoch": 6.0,
|
| 473 |
+
"eval_loss": 1.3067700862884521,
|
| 474 |
+
"eval_runtime": 175.1357,
|
| 475 |
+
"eval_samples_per_second": 36.092,
|
| 476 |
+
"eval_steps_per_second": 2.261,
|
| 477 |
+
"step": 297
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 6.0606060606060606,
|
| 481 |
+
"grad_norm": 0.8369827825158762,
|
| 482 |
+
"learning_rate": 7.844298380755003e-05,
|
| 483 |
+
"loss": 1.375,
|
| 484 |
+
"step": 300
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 6.161616161616162,
|
| 488 |
+
"grad_norm": 0.9204188282340791,
|
| 489 |
+
"learning_rate": 7.497924219967209e-05,
|
| 490 |
+
"loss": 1.2999,
|
| 491 |
+
"step": 305
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 6.262626262626263,
|
| 495 |
+
"grad_norm": 0.9559880600184892,
|
| 496 |
+
"learning_rate": 7.154724133689677e-05,
|
| 497 |
+
"loss": 1.3084,
|
| 498 |
+
"step": 310
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 6.363636363636363,
|
| 502 |
+
"grad_norm": 0.9272296702059781,
|
| 503 |
+
"learning_rate": 6.815133497483157e-05,
|
| 504 |
+
"loss": 1.3405,
|
| 505 |
+
"step": 315
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 6.4646464646464645,
|
| 509 |
+
"grad_norm": 1.0203421193696094,
|
| 510 |
+
"learning_rate": 6.479583108044899e-05,
|
| 511 |
+
"loss": 1.3165,
|
| 512 |
+
"step": 320
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 6.565656565656566,
|
| 516 |
+
"grad_norm": 0.8932381508297077,
|
| 517 |
+
"learning_rate": 6.148498636710092e-05,
|
| 518 |
+
"loss": 1.3641,
|
| 519 |
+
"step": 325
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 6.666666666666667,
|
| 523 |
+
"grad_norm": 0.9527012454845684,
|
| 524 |
+
"learning_rate": 5.822300089455211e-05,
|
| 525 |
+
"loss": 1.3179,
|
| 526 |
+
"step": 330
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 6.767676767676767,
|
| 530 |
+
"grad_norm": 0.9644270167383292,
|
| 531 |
+
"learning_rate": 5.5014012740883115e-05,
|
| 532 |
+
"loss": 1.3295,
|
| 533 |
+
"step": 335
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 6.8686868686868685,
|
| 537 |
+
"grad_norm": 0.9489303492473159,
|
| 538 |
+
"learning_rate": 5.1862092753021754e-05,
|
| 539 |
+
"loss": 1.3482,
|
| 540 |
+
"step": 340
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 6.96969696969697,
|
| 544 |
+
"grad_norm": 0.9417366559193787,
|
| 545 |
+
"learning_rate": 4.8771239382562287e-05,
|
| 546 |
+
"loss": 1.3422,
|
| 547 |
+
"step": 345
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 6.98989898989899,
|
| 551 |
+
"eval_loss": 1.2082042694091797,
|
| 552 |
+
"eval_runtime": 174.4886,
|
| 553 |
+
"eval_samples_per_second": 36.226,
|
| 554 |
+
"eval_steps_per_second": 2.269,
|
| 555 |
+
"step": 346
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 7.070707070707071,
|
| 559 |
+
"grad_norm": 1.0189885797060951,
|
| 560 |
+
"learning_rate": 4.574537361342407e-05,
|
| 561 |
+
"loss": 1.2447,
|
| 562 |
+
"step": 350
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 7.171717171717171,
|
| 566 |
+
"grad_norm": 1.0456878645941505,
|
| 567 |
+
"learning_rate": 4.278833398778306e-05,
|
| 568 |
+
"loss": 1.2438,
|
| 569 |
+
"step": 355
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 7.2727272727272725,
|
| 573 |
+
"grad_norm": 1.0906515200546398,
|
| 574 |
+
"learning_rate": 3.990387173658774e-05,
|
| 575 |
+
"loss": 1.2135,
|
| 576 |
+
"step": 360
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 7.373737373737374,
|
| 580 |
+
"grad_norm": 1.1138045736907602,
|
| 581 |
+
"learning_rate": 3.7095646020835754e-05,
|
| 582 |
+
"loss": 1.2152,
|
| 583 |
+
"step": 365
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 7.474747474747475,
|
| 587 |
+
"grad_norm": 1.1333935442018617,
|
| 588 |
+
"learning_rate": 3.436721928964819e-05,
|
| 589 |
+
"loss": 1.2004,
|
| 590 |
+
"step": 370
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 7.575757575757576,
|
| 594 |
+
"grad_norm": 1.0135992096066218,
|
| 595 |
+
"learning_rate": 3.172205276103033e-05,
|
| 596 |
+
"loss": 1.1904,
|
| 597 |
+
"step": 375
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 7.6767676767676765,
|
| 601 |
+
"grad_norm": 1.0811091792166911,
|
| 602 |
+
"learning_rate": 2.916350203105207e-05,
|
| 603 |
+
"loss": 1.2475,
|
| 604 |
+
"step": 380
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 7.777777777777778,
|
| 608 |
+
"grad_norm": 1.1722915377984628,
|
| 609 |
+
"learning_rate": 2.669481281701739e-05,
|
| 610 |
+
"loss": 1.2273,
|
| 611 |
+
"step": 385
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 7.878787878787879,
|
| 615 |
+
"grad_norm": 1.0151820296117031,
|
| 616 |
+
"learning_rate": 2.4319116840023813e-05,
|
| 617 |
+
"loss": 1.2462,
|
| 618 |
+
"step": 390
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 7.97979797979798,
|
| 622 |
+
"grad_norm": 1.0359433658999384,
|
| 623 |
+
"learning_rate": 2.2039427852134788e-05,
|
| 624 |
+
"loss": 1.2609,
|
| 625 |
+
"step": 395
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 8.0,
|
| 629 |
+
"eval_loss": 1.137781023979187,
|
| 630 |
+
"eval_runtime": 163.6859,
|
| 631 |
+
"eval_samples_per_second": 38.617,
|
| 632 |
+
"eval_steps_per_second": 2.419,
|
| 633 |
+
"step": 396
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 8.080808080808081,
|
| 637 |
+
"grad_norm": 1.0914414519615843,
|
| 638 |
+
"learning_rate": 1.985863781320435e-05,
|
| 639 |
+
"loss": 1.1457,
|
| 640 |
+
"step": 400
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 8.181818181818182,
|
| 644 |
+
"grad_norm": 1.2849174669504693,
|
| 645 |
+
"learning_rate": 1.777951322220508e-05,
|
| 646 |
+
"loss": 1.1925,
|
| 647 |
+
"step": 405
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 8.282828282828282,
|
| 651 |
+
"grad_norm": 1.0562037397284274,
|
| 652 |
+
"learning_rate": 1.580469160771253e-05,
|
| 653 |
+
"loss": 1.1653,
|
| 654 |
+
"step": 410
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 8.383838383838384,
|
| 658 |
+
"grad_norm": 1.1942325172166053,
|
| 659 |
+
"learning_rate": 1.3936678181998374e-05,
|
| 660 |
+
"loss": 1.1451,
|
| 661 |
+
"step": 415
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 8.484848484848484,
|
| 665 |
+
"grad_norm": 1.2292184186394104,
|
| 666 |
+
"learning_rate": 1.2177842662977135e-05,
|
| 667 |
+
"loss": 1.1432,
|
| 668 |
+
"step": 420
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 8.585858585858587,
|
| 672 |
+
"grad_norm": 1.1449254109310076,
|
| 673 |
+
"learning_rate": 1.0530416268037702e-05,
|
| 674 |
+
"loss": 1.1459,
|
| 675 |
+
"step": 425
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 8.686868686868687,
|
| 679 |
+
"grad_norm": 1.1159137674762092,
|
| 680 |
+
"learning_rate": 8.99648888357335e-06,
|
| 681 |
+
"loss": 1.1889,
|
| 682 |
+
"step": 430
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 8.787878787878787,
|
| 686 |
+
"grad_norm": 1.1893818134430183,
|
| 687 |
+
"learning_rate": 7.578006413801075e-06,
|
| 688 |
+
"loss": 1.1809,
|
| 689 |
+
"step": 435
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 8.88888888888889,
|
| 693 |
+
"grad_norm": 1.131890459862098,
|
| 694 |
+
"learning_rate": 6.276768312233228e-06,
|
| 695 |
+
"loss": 1.1806,
|
| 696 |
+
"step": 440
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 8.98989898989899,
|
| 700 |
+
"grad_norm": 1.0926795618787801,
|
| 701 |
+
"learning_rate": 5.094425298933136e-06,
|
| 702 |
+
"loss": 1.1647,
|
| 703 |
+
"step": 445
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 8.98989898989899,
|
| 707 |
+
"eval_loss": 1.107386827468872,
|
| 708 |
+
"eval_runtime": 163.223,
|
| 709 |
+
"eval_samples_per_second": 38.726,
|
| 710 |
+
"eval_steps_per_second": 2.426,
|
| 711 |
+
"step": 445
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 9.090909090909092,
|
| 715 |
+
"grad_norm": 1.1051037332814697,
|
| 716 |
+
"learning_rate": 4.0324772664503296e-06,
|
| 717 |
+
"loss": 1.1438,
|
| 718 |
+
"step": 450
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 9.191919191919192,
|
| 722 |
+
"grad_norm": 1.164376038164282,
|
| 723 |
+
"learning_rate": 3.092271377092215e-06,
|
| 724 |
+
"loss": 1.1481,
|
| 725 |
+
"step": 455
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 9.292929292929292,
|
| 729 |
+
"grad_norm": 1.2303081966513765,
|
| 730 |
+
"learning_rate": 2.2750003539455998e-06,
|
| 731 |
+
"loss": 1.1202,
|
| 732 |
+
"step": 460
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 9.393939393939394,
|
| 736 |
+
"grad_norm": 1.2010643624794166,
|
| 737 |
+
"learning_rate": 1.5817009678162685e-06,
|
| 738 |
+
"loss": 1.142,
|
| 739 |
+
"step": 465
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 9.494949494949495,
|
| 743 |
+
"grad_norm": 1.2423558300638782,
|
| 744 |
+
"learning_rate": 1.013252722005842e-06,
|
| 745 |
+
"loss": 1.1842,
|
| 746 |
+
"step": 470
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 9.595959595959595,
|
| 750 |
+
"grad_norm": 1.1980002179676799,
|
| 751 |
+
"learning_rate": 5.703767365946466e-07,
|
| 752 |
+
"loss": 1.1236,
|
| 753 |
+
"step": 475
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 9.696969696969697,
|
| 757 |
+
"grad_norm": 1.2034300162155251,
|
| 758 |
+
"learning_rate": 2.536348336456551e-07,
|
| 759 |
+
"loss": 1.1168,
|
| 760 |
+
"step": 480
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 9.797979797979798,
|
| 764 |
+
"grad_norm": 1.2022297984086214,
|
| 765 |
+
"learning_rate": 6.342882449029696e-08,
|
| 766 |
+
"loss": 1.1133,
|
| 767 |
+
"step": 485
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 9.8989898989899,
|
| 771 |
+
"grad_norm": 1.0699752057421905,
|
| 772 |
+
"learning_rate": 0.0,
|
| 773 |
+
"loss": 1.1571,
|
| 774 |
+
"step": 490
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 9.8989898989899,
|
| 778 |
+
"eval_loss": 1.1029597520828247,
|
| 779 |
+
"eval_runtime": 163.4283,
|
| 780 |
+
"eval_samples_per_second": 38.678,
|
| 781 |
+
"eval_steps_per_second": 2.423,
|
| 782 |
+
"step": 490
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 9.8989898989899,
|
| 786 |
+
"step": 490,
|
| 787 |
+
"total_flos": 2344635780825088.0,
|
| 788 |
+
"train_loss": 1.5290005391957808,
|
| 789 |
+
"train_runtime": 5413.3076,
|
| 790 |
+
"train_samples_per_second": 11.677,
|
| 791 |
"train_steps_per_second": 0.091
|
| 792 |
}
|
| 793 |
],
|
| 794 |
"logging_steps": 5,
|
| 795 |
+
"max_steps": 490,
|
| 796 |
"num_input_tokens_seen": 0,
|
| 797 |
+
"num_train_epochs": 10,
|
| 798 |
"save_steps": 25,
|
| 799 |
"stateful_callbacks": {
|
| 800 |
"TrainerControl": {
|
|
|
|
| 808 |
"attributes": {}
|
| 809 |
}
|
| 810 |
},
|
| 811 |
+
"total_flos": 2344635780825088.0,
|
| 812 |
"train_batch_size": 8,
|
| 813 |
"trial_name": null,
|
| 814 |
"trial_params": null
|