End of training
Browse files- README.md +10 -7
- all_results.json +6 -6
- model.safetensors +1 -1
- runs/Apr14_03-13-09_086e6037c7be/events.out.tfevents.1713064397.086e6037c7be.34.0 +3 -0
- runs/Apr14_03-22-36_086e6037c7be/events.out.tfevents.1713064962.086e6037c7be.34.1 +3 -0
- runs/Apr14_03-22-36_086e6037c7be/events.out.tfevents.1713067015.086e6037c7be.34.2 +3 -0
- train_results.json +6 -6
- trainer_state.json +223 -105
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [facebook/hubert-base-ls960](https://huggingface.co/facebook/hubert-base-ls960) on an unknown dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
-
- Loss: 0.
|
| 21 |
-
- Accuracy: 0.
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
@@ -45,17 +45,20 @@ The following hyperparameters were used during training:
|
|
| 45 |
- total_train_batch_size: 8
|
| 46 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
- lr_scheduler_type: linear
|
| 48 |
-
- num_epochs:
|
| 49 |
- mixed_precision_training: Native AMP
|
| 50 |
|
| 51 |
### Training results
|
| 52 |
|
| 53 |
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 54 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 55 |
-
| 0.
|
| 56 |
-
| 0.
|
| 57 |
-
| 0.
|
| 58 |
-
| 0.
|
|
|
|
|
|
|
|
|
|
| 59 |
|
| 60 |
|
| 61 |
### Framework versions
|
|
|
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [facebook/hubert-base-ls960](https://huggingface.co/facebook/hubert-base-ls960) on an unknown dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.1715
|
| 21 |
+
- Accuracy: 0.9737
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
|
|
| 45 |
- total_train_batch_size: 8
|
| 46 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
- lr_scheduler_type: linear
|
| 48 |
+
- num_epochs: 15.0
|
| 49 |
- mixed_precision_training: Native AMP
|
| 50 |
|
| 51 |
### Training results
|
| 52 |
|
| 53 |
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 54 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 55 |
+
| 0.554 | 2.02 | 500 | 0.2904 | 0.9152 |
|
| 56 |
+
| 0.2241 | 4.04 | 1000 | 0.2854 | 0.9313 |
|
| 57 |
+
| 0.1422 | 6.06 | 1500 | 0.3125 | 0.9333 |
|
| 58 |
+
| 0.1245 | 8.08 | 2000 | 0.1867 | 0.9636 |
|
| 59 |
+
| 0.0912 | 10.1 | 2500 | 0.1746 | 0.9697 |
|
| 60 |
+
| 0.0599 | 12.12 | 3000 | 0.2121 | 0.9596 |
|
| 61 |
+
| 0.041 | 14.14 | 3500 | 0.1715 | 0.9737 |
|
| 62 |
|
| 63 |
|
| 64 |
### Framework versions
|
all_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 1979,
|
| 7 |
-
"train_samples_per_second": 14.
|
| 8 |
-
"train_steps_per_second": 1.
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 14.97,
|
| 3 |
+
"total_flos": 9.492004863262944e+17,
|
| 4 |
+
"train_loss": 0.2598070231043858,
|
| 5 |
+
"train_runtime": 2003.5076,
|
| 6 |
"train_samples": 1979,
|
| 7 |
+
"train_samples_per_second": 14.817,
|
| 8 |
+
"train_steps_per_second": 1.849
|
| 9 |
}
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 379890236
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:79c2e792114ef1543892e47165b4b0e3d330d589a64047fbce0e86a38dc9c2a0
|
| 3 |
size 379890236
|
runs/Apr14_03-13-09_086e6037c7be/events.out.tfevents.1713064397.086e6037c7be.34.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:626ac898554fc079ab7f80057c21acc15900d22ecac89f4571396e9cf31a2cac
|
| 3 |
+
size 5854
|
runs/Apr14_03-22-36_086e6037c7be/events.out.tfevents.1713064962.086e6037c7be.34.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5e9928a9b5f6c07bc2f8075693c862485a2384ae3403e57b07716d642f6ef3ed
|
| 3 |
+
size 16274
|
runs/Apr14_03-22-36_086e6037c7be/events.out.tfevents.1713067015.086e6037c7be.34.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a33c501abe08f25b1ce35f260f925b1d9564fa3bd29d6aa8eb0af09785070b83
|
| 3 |
+
size 411
|
train_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 1979,
|
| 7 |
-
"train_samples_per_second": 14.
|
| 8 |
-
"train_steps_per_second": 1.
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 14.97,
|
| 3 |
+
"total_flos": 9.492004863262944e+17,
|
| 4 |
+
"train_loss": 0.2598070231043858,
|
| 5 |
+
"train_runtime": 2003.5076,
|
| 6 |
"train_samples": 1979,
|
| 7 |
+
"train_samples_per_second": 14.817,
|
| 8 |
+
"train_steps_per_second": 1.849
|
| 9 |
}
|
trainer_state.json
CHANGED
|
@@ -1,233 +1,351 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric":
|
| 3 |
-
"best_model_checkpoint":
|
| 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": 0.4,
|
| 13 |
-
"grad_norm":
|
| 14 |
-
"learning_rate": 9.
|
| 15 |
-
"loss":
|
| 16 |
"step": 100
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.81,
|
| 20 |
-
"grad_norm":
|
| 21 |
-
"learning_rate": 9.
|
| 22 |
-
"loss":
|
| 23 |
"step": 200
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 1.21,
|
| 27 |
-
"grad_norm":
|
| 28 |
-
"learning_rate":
|
| 29 |
-
"loss": 0.
|
| 30 |
"step": 300
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 1.62,
|
| 34 |
-
"grad_norm":
|
| 35 |
-
"learning_rate": 8.
|
| 36 |
-
"loss": 0.
|
| 37 |
"step": 400
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 2.02,
|
| 41 |
-
"grad_norm":
|
| 42 |
-
"learning_rate":
|
| 43 |
-
"loss": 0.
|
| 44 |
"step": 500
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 2.02,
|
| 48 |
-
"eval_accuracy": 0.
|
| 49 |
-
"eval_loss": 0.
|
| 50 |
-
"eval_runtime":
|
| 51 |
-
"eval_samples_per_second": 21.
|
| 52 |
-
"eval_steps_per_second": 5.
|
| 53 |
"step": 500
|
| 54 |
},
|
| 55 |
{
|
| 56 |
"epoch": 2.42,
|
| 57 |
-
"grad_norm":
|
| 58 |
-
"learning_rate":
|
| 59 |
-
"loss": 0.
|
| 60 |
"step": 600
|
| 61 |
},
|
| 62 |
{
|
| 63 |
"epoch": 2.83,
|
| 64 |
-
"grad_norm":
|
| 65 |
-
"learning_rate":
|
| 66 |
-
"loss": 0.
|
| 67 |
"step": 700
|
| 68 |
},
|
| 69 |
{
|
| 70 |
"epoch": 3.23,
|
| 71 |
-
"grad_norm":
|
| 72 |
-
"learning_rate":
|
| 73 |
-
"loss": 0.
|
| 74 |
"step": 800
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 3.64,
|
| 78 |
-
"grad_norm": 0.
|
| 79 |
-
"learning_rate":
|
| 80 |
-
"loss": 0.
|
| 81 |
"step": 900
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 4.04,
|
| 85 |
-
"grad_norm": 0.
|
| 86 |
-
"learning_rate":
|
| 87 |
-
"loss": 0.
|
| 88 |
"step": 1000
|
| 89 |
},
|
| 90 |
{
|
| 91 |
"epoch": 4.04,
|
| 92 |
-
"eval_accuracy": 0.
|
| 93 |
-
"eval_loss": 0.
|
| 94 |
-
"eval_runtime": 23.
|
| 95 |
-
"eval_samples_per_second": 21.
|
| 96 |
-
"eval_steps_per_second": 5.
|
| 97 |
"step": 1000
|
| 98 |
},
|
| 99 |
{
|
| 100 |
"epoch": 4.44,
|
| 101 |
-
"grad_norm":
|
| 102 |
-
"learning_rate":
|
| 103 |
-
"loss": 0.
|
| 104 |
"step": 1100
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 4.85,
|
| 108 |
-
"grad_norm":
|
| 109 |
-
"learning_rate":
|
| 110 |
-
"loss": 0.
|
| 111 |
"step": 1200
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 5.25,
|
| 115 |
-
"grad_norm":
|
| 116 |
-
"learning_rate":
|
| 117 |
-
"loss": 0.
|
| 118 |
"step": 1300
|
| 119 |
},
|
| 120 |
{
|
| 121 |
"epoch": 5.66,
|
| 122 |
-
"grad_norm":
|
| 123 |
-
"learning_rate":
|
| 124 |
-
"loss": 0.
|
| 125 |
"step": 1400
|
| 126 |
},
|
| 127 |
{
|
| 128 |
"epoch": 6.06,
|
| 129 |
-
"grad_norm":
|
| 130 |
-
"learning_rate":
|
| 131 |
-
"loss": 0.
|
| 132 |
"step": 1500
|
| 133 |
},
|
| 134 |
{
|
| 135 |
"epoch": 6.06,
|
| 136 |
-
"eval_accuracy": 0.
|
| 137 |
-
"eval_loss": 0.
|
| 138 |
-
"eval_runtime": 23.
|
| 139 |
-
"eval_samples_per_second": 21.
|
| 140 |
"eval_steps_per_second": 5.39,
|
| 141 |
"step": 1500
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 6.46,
|
| 145 |
-
"grad_norm":
|
| 146 |
-
"learning_rate":
|
| 147 |
-
"loss": 0.
|
| 148 |
"step": 1600
|
| 149 |
},
|
| 150 |
{
|
| 151 |
"epoch": 6.87,
|
| 152 |
-
"grad_norm": 0.
|
| 153 |
-
"learning_rate":
|
| 154 |
-
"loss": 0.
|
| 155 |
"step": 1700
|
| 156 |
},
|
| 157 |
{
|
| 158 |
"epoch": 7.27,
|
| 159 |
-
"grad_norm":
|
| 160 |
-
"learning_rate":
|
| 161 |
-
"loss": 0.
|
| 162 |
"step": 1800
|
| 163 |
},
|
| 164 |
{
|
| 165 |
"epoch": 7.68,
|
| 166 |
-
"grad_norm": 0.
|
| 167 |
-
"learning_rate":
|
| 168 |
-
"loss": 0.
|
| 169 |
"step": 1900
|
| 170 |
},
|
| 171 |
{
|
| 172 |
"epoch": 8.08,
|
| 173 |
-
"grad_norm":
|
| 174 |
-
"learning_rate":
|
| 175 |
-
"loss": 0.
|
| 176 |
"step": 2000
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 8.08,
|
| 180 |
-
"eval_accuracy": 0.
|
| 181 |
-
"eval_loss": 0.
|
| 182 |
-
"eval_runtime": 23.
|
| 183 |
-
"eval_samples_per_second": 21.
|
| 184 |
-
"eval_steps_per_second": 5.
|
| 185 |
"step": 2000
|
| 186 |
},
|
| 187 |
{
|
| 188 |
"epoch": 8.48,
|
| 189 |
-
"grad_norm":
|
| 190 |
-
"learning_rate":
|
| 191 |
-
"loss": 0.
|
| 192 |
"step": 2100
|
| 193 |
},
|
| 194 |
{
|
| 195 |
"epoch": 8.89,
|
| 196 |
-
"grad_norm":
|
| 197 |
-
"learning_rate":
|
| 198 |
-
"loss": 0.
|
| 199 |
"step": 2200
|
| 200 |
},
|
| 201 |
{
|
| 202 |
"epoch": 9.29,
|
| 203 |
-
"grad_norm":
|
| 204 |
-
"learning_rate":
|
| 205 |
-
"loss": 0.
|
| 206 |
"step": 2300
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 9.7,
|
| 210 |
-
"grad_norm": 0.
|
| 211 |
-
"learning_rate":
|
| 212 |
-
"loss": 0.
|
| 213 |
"step": 2400
|
| 214 |
},
|
| 215 |
{
|
| 216 |
-
"epoch":
|
| 217 |
-
"
|
| 218 |
-
"
|
| 219 |
-
"
|
| 220 |
-
"
|
| 221 |
-
|
| 222 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 223 |
}
|
| 224 |
],
|
| 225 |
"logging_steps": 100,
|
| 226 |
-
"max_steps":
|
| 227 |
"num_input_tokens_seen": 0,
|
| 228 |
-
"num_train_epochs":
|
| 229 |
"save_steps": 500,
|
| 230 |
-
"total_flos":
|
| 231 |
"train_batch_size": 4,
|
| 232 |
"trial_name": null,
|
| 233 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.17145082354545593,
|
| 3 |
+
"best_model_checkpoint": "/kaggle/working/ckpts/checkpoint-3500",
|
| 4 |
+
"epoch": 14.969696969696969,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 3705,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.4,
|
| 13 |
+
"grad_norm": 2.2636642456054688,
|
| 14 |
+
"learning_rate": 9.730094466936572e-06,
|
| 15 |
+
"loss": 1.5236,
|
| 16 |
"step": 100
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.81,
|
| 20 |
+
"grad_norm": 2.942427396774292,
|
| 21 |
+
"learning_rate": 9.460188933873145e-06,
|
| 22 |
+
"loss": 1.2036,
|
| 23 |
"step": 200
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 1.21,
|
| 27 |
+
"grad_norm": 4.616508960723877,
|
| 28 |
+
"learning_rate": 9.192982456140351e-06,
|
| 29 |
+
"loss": 0.8129,
|
| 30 |
"step": 300
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 1.62,
|
| 34 |
+
"grad_norm": 6.923274517059326,
|
| 35 |
+
"learning_rate": 8.923076923076925e-06,
|
| 36 |
+
"loss": 0.603,
|
| 37 |
"step": 400
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 2.02,
|
| 41 |
+
"grad_norm": 14.605295181274414,
|
| 42 |
+
"learning_rate": 8.65587044534413e-06,
|
| 43 |
+
"loss": 0.554,
|
| 44 |
"step": 500
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 2.02,
|
| 48 |
+
"eval_accuracy": 0.9151515364646912,
|
| 49 |
+
"eval_loss": 0.29042723774909973,
|
| 50 |
+
"eval_runtime": 22.943,
|
| 51 |
+
"eval_samples_per_second": 21.575,
|
| 52 |
+
"eval_steps_per_second": 5.405,
|
| 53 |
"step": 500
|
| 54 |
},
|
| 55 |
{
|
| 56 |
"epoch": 2.42,
|
| 57 |
+
"grad_norm": 8.531092643737793,
|
| 58 |
+
"learning_rate": 8.385964912280704e-06,
|
| 59 |
+
"loss": 0.4044,
|
| 60 |
"step": 600
|
| 61 |
},
|
| 62 |
{
|
| 63 |
"epoch": 2.83,
|
| 64 |
+
"grad_norm": 17.11962890625,
|
| 65 |
+
"learning_rate": 8.118758434547908e-06,
|
| 66 |
+
"loss": 0.3846,
|
| 67 |
"step": 700
|
| 68 |
},
|
| 69 |
{
|
| 70 |
"epoch": 3.23,
|
| 71 |
+
"grad_norm": 40.550960540771484,
|
| 72 |
+
"learning_rate": 7.848852901484481e-06,
|
| 73 |
+
"loss": 0.3334,
|
| 74 |
"step": 800
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 3.64,
|
| 78 |
+
"grad_norm": 0.589801549911499,
|
| 79 |
+
"learning_rate": 7.578947368421054e-06,
|
| 80 |
+
"loss": 0.2088,
|
| 81 |
"step": 900
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 4.04,
|
| 85 |
+
"grad_norm": 0.7292308211326599,
|
| 86 |
+
"learning_rate": 7.309041835357625e-06,
|
| 87 |
+
"loss": 0.2241,
|
| 88 |
"step": 1000
|
| 89 |
},
|
| 90 |
{
|
| 91 |
"epoch": 4.04,
|
| 92 |
+
"eval_accuracy": 0.931313157081604,
|
| 93 |
+
"eval_loss": 0.2854190170764923,
|
| 94 |
+
"eval_runtime": 23.1163,
|
| 95 |
+
"eval_samples_per_second": 21.413,
|
| 96 |
+
"eval_steps_per_second": 5.364,
|
| 97 |
"step": 1000
|
| 98 |
},
|
| 99 |
{
|
| 100 |
"epoch": 4.44,
|
| 101 |
+
"grad_norm": 4.302298069000244,
|
| 102 |
+
"learning_rate": 7.039136302294197e-06,
|
| 103 |
+
"loss": 0.209,
|
| 104 |
"step": 1100
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 4.85,
|
| 108 |
+
"grad_norm": 8.639167785644531,
|
| 109 |
+
"learning_rate": 6.76923076923077e-06,
|
| 110 |
+
"loss": 0.2459,
|
| 111 |
"step": 1200
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 5.25,
|
| 115 |
+
"grad_norm": 14.07341194152832,
|
| 116 |
+
"learning_rate": 6.502024291497976e-06,
|
| 117 |
+
"loss": 0.2448,
|
| 118 |
"step": 1300
|
| 119 |
},
|
| 120 |
{
|
| 121 |
"epoch": 5.66,
|
| 122 |
+
"grad_norm": 25.380226135253906,
|
| 123 |
+
"learning_rate": 6.232118758434549e-06,
|
| 124 |
+
"loss": 0.2045,
|
| 125 |
"step": 1400
|
| 126 |
},
|
| 127 |
{
|
| 128 |
"epoch": 6.06,
|
| 129 |
+
"grad_norm": 20.343219757080078,
|
| 130 |
+
"learning_rate": 5.962213225371121e-06,
|
| 131 |
+
"loss": 0.1422,
|
| 132 |
"step": 1500
|
| 133 |
},
|
| 134 |
{
|
| 135 |
"epoch": 6.06,
|
| 136 |
+
"eval_accuracy": 0.9333333373069763,
|
| 137 |
+
"eval_loss": 0.31245771050453186,
|
| 138 |
+
"eval_runtime": 23.007,
|
| 139 |
+
"eval_samples_per_second": 21.515,
|
| 140 |
"eval_steps_per_second": 5.39,
|
| 141 |
"step": 1500
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 6.46,
|
| 145 |
+
"grad_norm": 48.67534255981445,
|
| 146 |
+
"learning_rate": 5.692307692307692e-06,
|
| 147 |
+
"loss": 0.1553,
|
| 148 |
"step": 1600
|
| 149 |
},
|
| 150 |
{
|
| 151 |
"epoch": 6.87,
|
| 152 |
+
"grad_norm": 0.05280930921435356,
|
| 153 |
+
"learning_rate": 5.4224021592442655e-06,
|
| 154 |
+
"loss": 0.1939,
|
| 155 |
"step": 1700
|
| 156 |
},
|
| 157 |
{
|
| 158 |
"epoch": 7.27,
|
| 159 |
+
"grad_norm": 2.396838426589966,
|
| 160 |
+
"learning_rate": 5.152496626180837e-06,
|
| 161 |
+
"loss": 0.1421,
|
| 162 |
"step": 1800
|
| 163 |
},
|
| 164 |
{
|
| 165 |
"epoch": 7.68,
|
| 166 |
+
"grad_norm": 0.47542768716812134,
|
| 167 |
+
"learning_rate": 4.882591093117409e-06,
|
| 168 |
+
"loss": 0.1711,
|
| 169 |
"step": 1900
|
| 170 |
},
|
| 171 |
{
|
| 172 |
"epoch": 8.08,
|
| 173 |
+
"grad_norm": 24.91346549987793,
|
| 174 |
+
"learning_rate": 4.6126855600539814e-06,
|
| 175 |
+
"loss": 0.1245,
|
| 176 |
"step": 2000
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 8.08,
|
| 180 |
+
"eval_accuracy": 0.9636363387107849,
|
| 181 |
+
"eval_loss": 0.18670035898685455,
|
| 182 |
+
"eval_runtime": 23.2005,
|
| 183 |
+
"eval_samples_per_second": 21.336,
|
| 184 |
+
"eval_steps_per_second": 5.345,
|
| 185 |
"step": 2000
|
| 186 |
},
|
| 187 |
{
|
| 188 |
"epoch": 8.48,
|
| 189 |
+
"grad_norm": 103.44571685791016,
|
| 190 |
+
"learning_rate": 4.342780026990554e-06,
|
| 191 |
+
"loss": 0.1178,
|
| 192 |
"step": 2100
|
| 193 |
},
|
| 194 |
{
|
| 195 |
"epoch": 8.89,
|
| 196 |
+
"grad_norm": 3.599931478500366,
|
| 197 |
+
"learning_rate": 4.072874493927126e-06,
|
| 198 |
+
"loss": 0.1012,
|
| 199 |
"step": 2200
|
| 200 |
},
|
| 201 |
{
|
| 202 |
"epoch": 9.29,
|
| 203 |
+
"grad_norm": 13.552335739135742,
|
| 204 |
+
"learning_rate": 3.8029689608636982e-06,
|
| 205 |
+
"loss": 0.1241,
|
| 206 |
"step": 2300
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 9.7,
|
| 210 |
+
"grad_norm": 0.7136185169219971,
|
| 211 |
+
"learning_rate": 3.53306342780027e-06,
|
| 212 |
+
"loss": 0.1054,
|
| 213 |
"step": 2400
|
| 214 |
},
|
| 215 |
{
|
| 216 |
+
"epoch": 10.1,
|
| 217 |
+
"grad_norm": 0.019901130348443985,
|
| 218 |
+
"learning_rate": 3.2631578947368423e-06,
|
| 219 |
+
"loss": 0.0912,
|
| 220 |
+
"step": 2500
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 10.1,
|
| 224 |
+
"eval_accuracy": 0.9696969985961914,
|
| 225 |
+
"eval_loss": 0.1746375858783722,
|
| 226 |
+
"eval_runtime": 22.9664,
|
| 227 |
+
"eval_samples_per_second": 21.553,
|
| 228 |
+
"eval_steps_per_second": 5.399,
|
| 229 |
+
"step": 2500
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 10.51,
|
| 233 |
+
"grad_norm": 0.025977592915296555,
|
| 234 |
+
"learning_rate": 2.9932523616734146e-06,
|
| 235 |
+
"loss": 0.0672,
|
| 236 |
+
"step": 2600
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 10.91,
|
| 240 |
+
"grad_norm": 0.022668780758976936,
|
| 241 |
+
"learning_rate": 2.723346828609987e-06,
|
| 242 |
+
"loss": 0.1198,
|
| 243 |
+
"step": 2700
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 11.31,
|
| 247 |
+
"grad_norm": 18.671525955200195,
|
| 248 |
+
"learning_rate": 2.453441295546559e-06,
|
| 249 |
+
"loss": 0.0758,
|
| 250 |
+
"step": 2800
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 11.72,
|
| 254 |
+
"grad_norm": 10.470223426818848,
|
| 255 |
+
"learning_rate": 2.183535762483131e-06,
|
| 256 |
+
"loss": 0.1329,
|
| 257 |
+
"step": 2900
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 12.12,
|
| 261 |
+
"grad_norm": 0.030979245901107788,
|
| 262 |
+
"learning_rate": 1.913630229419703e-06,
|
| 263 |
+
"loss": 0.0599,
|
| 264 |
+
"step": 3000
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 12.12,
|
| 268 |
+
"eval_accuracy": 0.9595959782600403,
|
| 269 |
+
"eval_loss": 0.21210218966007233,
|
| 270 |
+
"eval_runtime": 22.8997,
|
| 271 |
+
"eval_samples_per_second": 21.616,
|
| 272 |
+
"eval_steps_per_second": 5.415,
|
| 273 |
+
"step": 3000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 12.53,
|
| 277 |
+
"grad_norm": 0.3461657166481018,
|
| 278 |
+
"learning_rate": 1.6437246963562757e-06,
|
| 279 |
+
"loss": 0.0821,
|
| 280 |
+
"step": 3100
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 12.93,
|
| 284 |
+
"grad_norm": 38.45835876464844,
|
| 285 |
+
"learning_rate": 1.3738191632928477e-06,
|
| 286 |
+
"loss": 0.0919,
|
| 287 |
+
"step": 3200
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 13.33,
|
| 291 |
+
"grad_norm": 0.8387115001678467,
|
| 292 |
+
"learning_rate": 1.106612685560054e-06,
|
| 293 |
+
"loss": 0.0814,
|
| 294 |
+
"step": 3300
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 13.74,
|
| 298 |
+
"grad_norm": 0.22132369875907898,
|
| 299 |
+
"learning_rate": 8.367071524966262e-07,
|
| 300 |
+
"loss": 0.0954,
|
| 301 |
+
"step": 3400
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 14.14,
|
| 305 |
+
"grad_norm": 0.05546307936310768,
|
| 306 |
+
"learning_rate": 5.668016194331984e-07,
|
| 307 |
+
"loss": 0.041,
|
| 308 |
+
"step": 3500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 14.14,
|
| 312 |
+
"eval_accuracy": 0.973737359046936,
|
| 313 |
+
"eval_loss": 0.17145082354545593,
|
| 314 |
+
"eval_runtime": 23.372,
|
| 315 |
+
"eval_samples_per_second": 21.179,
|
| 316 |
+
"eval_steps_per_second": 5.305,
|
| 317 |
+
"step": 3500
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 14.55,
|
| 321 |
+
"grad_norm": 17.49441146850586,
|
| 322 |
+
"learning_rate": 2.968960863697706e-07,
|
| 323 |
+
"loss": 0.0978,
|
| 324 |
+
"step": 3600
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 14.95,
|
| 328 |
+
"grad_norm": 0.08890076726675034,
|
| 329 |
+
"learning_rate": 2.6990553306342783e-08,
|
| 330 |
+
"loss": 0.0463,
|
| 331 |
+
"step": 3700
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 14.97,
|
| 335 |
+
"step": 3705,
|
| 336 |
+
"total_flos": 9.492004863262944e+17,
|
| 337 |
+
"train_loss": 0.2598070231043858,
|
| 338 |
+
"train_runtime": 2003.5076,
|
| 339 |
+
"train_samples_per_second": 14.817,
|
| 340 |
+
"train_steps_per_second": 1.849
|
| 341 |
}
|
| 342 |
],
|
| 343 |
"logging_steps": 100,
|
| 344 |
+
"max_steps": 3705,
|
| 345 |
"num_input_tokens_seen": 0,
|
| 346 |
+
"num_train_epochs": 15,
|
| 347 |
"save_steps": 500,
|
| 348 |
+
"total_flos": 9.492004863262944e+17,
|
| 349 |
"train_batch_size": 4,
|
| 350 |
"trial_name": null,
|
| 351 |
"trial_params": null
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:130caa78aae0916c3a7db0f96d6974f82f39539ced1fb8a646a4918394f60256
|
| 3 |
size 4984
|