End of training
Browse files- README.md +2 -1
- all_results.json +12 -0
- eval_results.json +7 -0
- train_results.json +8 -0
- trainer_state.json +318 -0
- training_eval_loss.png +0 -0
- training_loss.png +0 -0
README.md
CHANGED
|
@@ -4,6 +4,7 @@ license: llama2
|
|
| 4 |
base_model: meta-llama/Llama-2-7b-hf
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
|
|
|
| 7 |
- generated_from_trainer
|
| 8 |
model-index:
|
| 9 |
- name: llama2_oh_teknium_scaling_down_random_0.3
|
|
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# llama2_oh_teknium_scaling_down_random_0.3
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
- Loss: 0.4943
|
| 21 |
|
|
|
|
| 4 |
base_model: meta-llama/Llama-2-7b-hf
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
| 7 |
+
- full
|
| 8 |
- generated_from_trainer
|
| 9 |
model-index:
|
| 10 |
- name: llama2_oh_teknium_scaling_down_random_0.3
|
|
|
|
| 16 |
|
| 17 |
# llama2_oh_teknium_scaling_down_random_0.3
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on the mlfoundations-dev/oh_teknium_scaling_down_random_0.3 dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
- Loss: 0.4943
|
| 22 |
|
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.987551867219917,
|
| 3 |
+
"eval_loss": 0.49428611993789673,
|
| 4 |
+
"eval_runtime": 122.5065,
|
| 5 |
+
"eval_samples_per_second": 26.488,
|
| 6 |
+
"eval_steps_per_second": 0.416,
|
| 7 |
+
"total_flos": 602804028702720.0,
|
| 8 |
+
"train_loss": 0.5124640332327949,
|
| 9 |
+
"train_runtime": 20041.158,
|
| 10 |
+
"train_samples_per_second": 9.227,
|
| 11 |
+
"train_steps_per_second": 0.018
|
| 12 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.987551867219917,
|
| 3 |
+
"eval_loss": 0.49428611993789673,
|
| 4 |
+
"eval_runtime": 122.5065,
|
| 5 |
+
"eval_samples_per_second": 26.488,
|
| 6 |
+
"eval_steps_per_second": 0.416
|
| 7 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.987551867219917,
|
| 3 |
+
"total_flos": 602804028702720.0,
|
| 4 |
+
"train_loss": 0.5124640332327949,
|
| 5 |
+
"train_runtime": 20041.158,
|
| 6 |
+
"train_samples_per_second": 9.227,
|
| 7 |
+
"train_steps_per_second": 0.018
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,318 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.987551867219917,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 360,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.08298755186721991,
|
| 13 |
+
"grad_norm": 0.4138890994781238,
|
| 14 |
+
"learning_rate": 5e-06,
|
| 15 |
+
"loss": 0.6819,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.16597510373443983,
|
| 20 |
+
"grad_norm": 0.29279799441362253,
|
| 21 |
+
"learning_rate": 5e-06,
|
| 22 |
+
"loss": 0.6152,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.24896265560165975,
|
| 27 |
+
"grad_norm": 0.22155490022279595,
|
| 28 |
+
"learning_rate": 5e-06,
|
| 29 |
+
"loss": 0.5911,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.33195020746887965,
|
| 34 |
+
"grad_norm": 0.21703792973144043,
|
| 35 |
+
"learning_rate": 5e-06,
|
| 36 |
+
"loss": 0.5758,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.4149377593360996,
|
| 41 |
+
"grad_norm": 0.19782125949582666,
|
| 42 |
+
"learning_rate": 5e-06,
|
| 43 |
+
"loss": 0.5658,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.4979253112033195,
|
| 48 |
+
"grad_norm": 0.20429620764864578,
|
| 49 |
+
"learning_rate": 5e-06,
|
| 50 |
+
"loss": 0.5621,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.5809128630705395,
|
| 55 |
+
"grad_norm": 0.1947179711144,
|
| 56 |
+
"learning_rate": 5e-06,
|
| 57 |
+
"loss": 0.5481,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.6639004149377593,
|
| 62 |
+
"grad_norm": 0.20952965041956714,
|
| 63 |
+
"learning_rate": 5e-06,
|
| 64 |
+
"loss": 0.5491,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.7468879668049793,
|
| 69 |
+
"grad_norm": 0.2092027679734135,
|
| 70 |
+
"learning_rate": 5e-06,
|
| 71 |
+
"loss": 0.5473,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.8298755186721992,
|
| 76 |
+
"grad_norm": 0.1973704614234666,
|
| 77 |
+
"learning_rate": 5e-06,
|
| 78 |
+
"loss": 0.5412,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.9128630705394191,
|
| 83 |
+
"grad_norm": 0.22215520376065145,
|
| 84 |
+
"learning_rate": 5e-06,
|
| 85 |
+
"loss": 0.5361,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.995850622406639,
|
| 90 |
+
"grad_norm": 0.20002555613598916,
|
| 91 |
+
"learning_rate": 5e-06,
|
| 92 |
+
"loss": 0.5354,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.995850622406639,
|
| 97 |
+
"eval_loss": 0.5283368229866028,
|
| 98 |
+
"eval_runtime": 121.5237,
|
| 99 |
+
"eval_samples_per_second": 26.703,
|
| 100 |
+
"eval_steps_per_second": 0.42,
|
| 101 |
+
"step": 120
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.0788381742738589,
|
| 105 |
+
"grad_norm": 0.23151869944663353,
|
| 106 |
+
"learning_rate": 5e-06,
|
| 107 |
+
"loss": 0.5334,
|
| 108 |
+
"step": 130
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 1.161825726141079,
|
| 112 |
+
"grad_norm": 0.20416069004838694,
|
| 113 |
+
"learning_rate": 5e-06,
|
| 114 |
+
"loss": 0.5122,
|
| 115 |
+
"step": 140
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 1.2448132780082988,
|
| 119 |
+
"grad_norm": 0.21256654137396935,
|
| 120 |
+
"learning_rate": 5e-06,
|
| 121 |
+
"loss": 0.509,
|
| 122 |
+
"step": 150
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 1.3278008298755186,
|
| 126 |
+
"grad_norm": 0.21018667523519946,
|
| 127 |
+
"learning_rate": 5e-06,
|
| 128 |
+
"loss": 0.5041,
|
| 129 |
+
"step": 160
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 1.4107883817427385,
|
| 133 |
+
"grad_norm": 0.219240042940767,
|
| 134 |
+
"learning_rate": 5e-06,
|
| 135 |
+
"loss": 0.4998,
|
| 136 |
+
"step": 170
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 1.4937759336099585,
|
| 140 |
+
"grad_norm": 0.22681455392212077,
|
| 141 |
+
"learning_rate": 5e-06,
|
| 142 |
+
"loss": 0.5037,
|
| 143 |
+
"step": 180
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.5767634854771784,
|
| 147 |
+
"grad_norm": 0.227133839723048,
|
| 148 |
+
"learning_rate": 5e-06,
|
| 149 |
+
"loss": 0.4977,
|
| 150 |
+
"step": 190
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 1.6597510373443982,
|
| 154 |
+
"grad_norm": 0.21040711904959797,
|
| 155 |
+
"learning_rate": 5e-06,
|
| 156 |
+
"loss": 0.4941,
|
| 157 |
+
"step": 200
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 1.7427385892116183,
|
| 161 |
+
"grad_norm": 0.23482785666403702,
|
| 162 |
+
"learning_rate": 5e-06,
|
| 163 |
+
"loss": 0.4945,
|
| 164 |
+
"step": 210
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 1.8257261410788381,
|
| 168 |
+
"grad_norm": 0.2035179907011211,
|
| 169 |
+
"learning_rate": 5e-06,
|
| 170 |
+
"loss": 0.4904,
|
| 171 |
+
"step": 220
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 1.908713692946058,
|
| 175 |
+
"grad_norm": 0.21720290177963564,
|
| 176 |
+
"learning_rate": 5e-06,
|
| 177 |
+
"loss": 0.491,
|
| 178 |
+
"step": 230
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 1.991701244813278,
|
| 182 |
+
"grad_norm": 0.2214820393037949,
|
| 183 |
+
"learning_rate": 5e-06,
|
| 184 |
+
"loss": 0.4901,
|
| 185 |
+
"step": 240
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 2.0,
|
| 189 |
+
"eval_loss": 0.5045989155769348,
|
| 190 |
+
"eval_runtime": 122.3281,
|
| 191 |
+
"eval_samples_per_second": 26.527,
|
| 192 |
+
"eval_steps_per_second": 0.417,
|
| 193 |
+
"step": 241
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 2.074688796680498,
|
| 197 |
+
"grad_norm": 0.25601226331965665,
|
| 198 |
+
"learning_rate": 5e-06,
|
| 199 |
+
"loss": 0.4945,
|
| 200 |
+
"step": 250
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 2.1576763485477177,
|
| 204 |
+
"grad_norm": 0.2537099080076595,
|
| 205 |
+
"learning_rate": 5e-06,
|
| 206 |
+
"loss": 0.4617,
|
| 207 |
+
"step": 260
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 2.240663900414938,
|
| 211 |
+
"grad_norm": 0.2445352596834903,
|
| 212 |
+
"learning_rate": 5e-06,
|
| 213 |
+
"loss": 0.4648,
|
| 214 |
+
"step": 270
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 2.323651452282158,
|
| 218 |
+
"grad_norm": 0.24195048816699535,
|
| 219 |
+
"learning_rate": 5e-06,
|
| 220 |
+
"loss": 0.4688,
|
| 221 |
+
"step": 280
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 2.4066390041493775,
|
| 225 |
+
"grad_norm": 0.3297443855710949,
|
| 226 |
+
"learning_rate": 5e-06,
|
| 227 |
+
"loss": 0.46,
|
| 228 |
+
"step": 290
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 2.4896265560165975,
|
| 232 |
+
"grad_norm": 0.2227067008121754,
|
| 233 |
+
"learning_rate": 5e-06,
|
| 234 |
+
"loss": 0.4679,
|
| 235 |
+
"step": 300
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 2.572614107883817,
|
| 239 |
+
"grad_norm": 0.24268677689146825,
|
| 240 |
+
"learning_rate": 5e-06,
|
| 241 |
+
"loss": 0.4642,
|
| 242 |
+
"step": 310
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 2.6556016597510372,
|
| 246 |
+
"grad_norm": 0.24131530500929413,
|
| 247 |
+
"learning_rate": 5e-06,
|
| 248 |
+
"loss": 0.4597,
|
| 249 |
+
"step": 320
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 2.7385892116182573,
|
| 253 |
+
"grad_norm": 0.22997089130920098,
|
| 254 |
+
"learning_rate": 5e-06,
|
| 255 |
+
"loss": 0.4617,
|
| 256 |
+
"step": 330
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 2.821576763485477,
|
| 260 |
+
"grad_norm": 0.23994756278793414,
|
| 261 |
+
"learning_rate": 5e-06,
|
| 262 |
+
"loss": 0.4597,
|
| 263 |
+
"step": 340
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 2.904564315352697,
|
| 267 |
+
"grad_norm": 0.23257285232469585,
|
| 268 |
+
"learning_rate": 5e-06,
|
| 269 |
+
"loss": 0.4545,
|
| 270 |
+
"step": 350
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 2.987551867219917,
|
| 274 |
+
"grad_norm": 0.22525776234601527,
|
| 275 |
+
"learning_rate": 5e-06,
|
| 276 |
+
"loss": 0.4618,
|
| 277 |
+
"step": 360
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 2.987551867219917,
|
| 281 |
+
"eval_loss": 0.49428611993789673,
|
| 282 |
+
"eval_runtime": 121.9138,
|
| 283 |
+
"eval_samples_per_second": 26.617,
|
| 284 |
+
"eval_steps_per_second": 0.418,
|
| 285 |
+
"step": 360
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 2.987551867219917,
|
| 289 |
+
"step": 360,
|
| 290 |
+
"total_flos": 602804028702720.0,
|
| 291 |
+
"train_loss": 0.5124640332327949,
|
| 292 |
+
"train_runtime": 20041.158,
|
| 293 |
+
"train_samples_per_second": 9.227,
|
| 294 |
+
"train_steps_per_second": 0.018
|
| 295 |
+
}
|
| 296 |
+
],
|
| 297 |
+
"logging_steps": 10,
|
| 298 |
+
"max_steps": 360,
|
| 299 |
+
"num_input_tokens_seen": 0,
|
| 300 |
+
"num_train_epochs": 3,
|
| 301 |
+
"save_steps": 500,
|
| 302 |
+
"stateful_callbacks": {
|
| 303 |
+
"TrainerControl": {
|
| 304 |
+
"args": {
|
| 305 |
+
"should_epoch_stop": false,
|
| 306 |
+
"should_evaluate": false,
|
| 307 |
+
"should_log": false,
|
| 308 |
+
"should_save": true,
|
| 309 |
+
"should_training_stop": true
|
| 310 |
+
},
|
| 311 |
+
"attributes": {}
|
| 312 |
+
}
|
| 313 |
+
},
|
| 314 |
+
"total_flos": 602804028702720.0,
|
| 315 |
+
"train_batch_size": 8,
|
| 316 |
+
"trial_name": null,
|
| 317 |
+
"trial_params": null
|
| 318 |
+
}
|
training_eval_loss.png
ADDED
|
training_loss.png
ADDED
|