Commit
·
c6db5b6
1
Parent(s):
6dee97d
ccit_final
Browse files- config.json +1 -1
- model.safetensors +2 -2
- optimizer.pt +2 -2
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +501 -164
config.json
CHANGED
|
@@ -20,7 +20,7 @@
|
|
| 20 |
"pad_token_id": 1,
|
| 21 |
"position_embedding_type": "absolute",
|
| 22 |
"torch_dtype": "float32",
|
| 23 |
-
"transformers_version": "4.41.
|
| 24 |
"type_vocab_size": 1,
|
| 25 |
"use_cache": true,
|
| 26 |
"vocab_size": 50265
|
|
|
|
| 20 |
"pad_token_id": 1,
|
| 21 |
"position_embedding_type": "absolute",
|
| 22 |
"torch_dtype": "float32",
|
| 23 |
+
"transformers_version": "4.41.2",
|
| 24 |
"type_vocab_size": 1,
|
| 25 |
"use_cache": true,
|
| 26 |
"vocab_size": 50265
|
model.safetensors
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:b6aad7c2f42e665fe68a35f0abf79a2483077c39dd42fba313390c6515423a69
|
| 3 |
+
size 498705152
|
optimizer.pt
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:a7e0a9772004bf6dc7cb7d761cdf5209bcbc125cb4ff8281b6b4af487abc8608
|
| 3 |
+
size 997531514
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4297877051912922800eeda2dfabec7286e3f19a18bb48fac3d835eb91c6dbae
|
| 3 |
size 14244
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:db3e53efd18aae3a98eba7ce2c13a309a6e1555c5a2fe5b6c71d6270566d2fef
|
| 3 |
size 1064
|
trainer_state.json
CHANGED
|
@@ -1,219 +1,556 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
-
"epoch": 0.
|
| 13 |
-
"grad_norm": 14.
|
| 14 |
"learning_rate": 4.97e-05,
|
| 15 |
-
"loss": 1.
|
| 16 |
"step": 500
|
| 17 |
},
|
| 18 |
{
|
| 19 |
-
"epoch": 0.
|
| 20 |
-
"eval_accuracy": 0.
|
| 21 |
-
"eval_f1_class_0": 0.
|
| 22 |
-
"eval_f1_class_1": 0.
|
| 23 |
-
"eval_f1_class_2": 0.
|
| 24 |
-
"eval_f1_class_3": 0.
|
| 25 |
-
"eval_f1_class_4": 0.
|
| 26 |
-
"eval_f1_class_5": 0.
|
| 27 |
-
"eval_f1_class_6": 0.
|
| 28 |
-
"eval_f1_class_7": 0.
|
| 29 |
-
"eval_f1_class_8": 0.
|
| 30 |
-
"
|
| 31 |
-
"
|
| 32 |
-
"
|
| 33 |
-
"
|
| 34 |
-
"
|
| 35 |
-
"
|
| 36 |
-
"
|
| 37 |
-
"
|
|
|
|
| 38 |
"step": 500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
-
"epoch": 1.
|
| 42 |
-
"grad_norm":
|
| 43 |
-
"learning_rate": 4.
|
| 44 |
-
"loss": 0.
|
| 45 |
"step": 1000
|
| 46 |
},
|
| 47 |
{
|
| 48 |
-
"epoch": 1.
|
| 49 |
-
"eval_accuracy": 0.
|
| 50 |
-
"eval_f1_class_0": 0.
|
| 51 |
-
"eval_f1_class_1": 0.
|
| 52 |
-
"eval_f1_class_2": 0.
|
| 53 |
-
"eval_f1_class_3": 0.
|
| 54 |
-
"eval_f1_class_4": 0.
|
| 55 |
-
"eval_f1_class_5": 0.
|
| 56 |
-
"eval_f1_class_6": 0.
|
| 57 |
-
"eval_f1_class_7": 0.
|
| 58 |
-
"eval_f1_class_8": 0.
|
| 59 |
-
"
|
| 60 |
-
"
|
| 61 |
-
"
|
| 62 |
-
"
|
| 63 |
-
"
|
| 64 |
-
"
|
| 65 |
-
"
|
| 66 |
-
"
|
|
|
|
| 67 |
"step": 1000
|
| 68 |
},
|
| 69 |
{
|
| 70 |
-
"epoch":
|
| 71 |
-
"grad_norm":
|
| 72 |
-
"learning_rate": 4.
|
| 73 |
-
"loss": 0.
|
| 74 |
"step": 1500
|
| 75 |
},
|
| 76 |
{
|
| 77 |
-
"epoch":
|
| 78 |
-
"eval_accuracy": 0.
|
| 79 |
-
"eval_f1_class_0": 0.
|
| 80 |
-
"eval_f1_class_1": 0.
|
| 81 |
-
"eval_f1_class_2": 0.
|
| 82 |
-
"eval_f1_class_3": 0.
|
| 83 |
-
"eval_f1_class_4": 0.
|
| 84 |
-
"eval_f1_class_5": 0.
|
| 85 |
-
"eval_f1_class_6": 0.
|
| 86 |
-
"eval_f1_class_7": 0.
|
| 87 |
-
"eval_f1_class_8": 0.
|
| 88 |
-
"
|
| 89 |
-
"
|
| 90 |
-
"
|
| 91 |
-
"
|
| 92 |
-
"
|
| 93 |
-
"
|
| 94 |
-
"
|
| 95 |
-
"
|
|
|
|
| 96 |
"step": 1500
|
| 97 |
},
|
| 98 |
{
|
| 99 |
-
"epoch": 2.
|
| 100 |
-
"grad_norm":
|
| 101 |
-
"learning_rate":
|
| 102 |
-
"loss": 0.
|
| 103 |
"step": 2000
|
| 104 |
},
|
| 105 |
{
|
| 106 |
-
"epoch": 2.
|
| 107 |
-
"eval_accuracy": 0.
|
| 108 |
-
"eval_f1_class_0": 0.
|
| 109 |
-
"eval_f1_class_1": 0.
|
| 110 |
-
"eval_f1_class_2": 0.
|
| 111 |
-
"eval_f1_class_3": 0.
|
| 112 |
-
"eval_f1_class_4": 0.
|
| 113 |
-
"eval_f1_class_5": 0.
|
| 114 |
-
"eval_f1_class_6": 0.
|
| 115 |
-
"eval_f1_class_7": 0.
|
| 116 |
-
"eval_f1_class_8": 0.
|
| 117 |
-
"
|
| 118 |
-
"
|
| 119 |
-
"
|
| 120 |
-
"
|
| 121 |
-
"
|
| 122 |
-
"
|
| 123 |
-
"
|
| 124 |
-
"
|
|
|
|
| 125 |
"step": 2000
|
| 126 |
},
|
| 127 |
{
|
| 128 |
-
"epoch":
|
| 129 |
-
"grad_norm":
|
| 130 |
-
"learning_rate": 3.
|
| 131 |
-
"loss": 0.
|
| 132 |
"step": 2500
|
| 133 |
},
|
| 134 |
{
|
| 135 |
-
"epoch":
|
| 136 |
-
"eval_accuracy": 0.
|
| 137 |
-
"eval_f1_class_0": 0.
|
| 138 |
-
"eval_f1_class_1": 0.
|
| 139 |
-
"eval_f1_class_2": 0.
|
| 140 |
-
"eval_f1_class_3": 0.
|
| 141 |
-
"eval_f1_class_4": 0.
|
| 142 |
-
"eval_f1_class_5": 0.
|
| 143 |
-
"eval_f1_class_6": 0.
|
| 144 |
-
"eval_f1_class_7": 0.
|
| 145 |
-
"eval_f1_class_8": 0.
|
| 146 |
-
"
|
| 147 |
-
"
|
| 148 |
-
"
|
| 149 |
-
"
|
| 150 |
-
"
|
| 151 |
-
"
|
| 152 |
-
"
|
| 153 |
-
"
|
|
|
|
| 154 |
"step": 2500
|
| 155 |
},
|
| 156 |
{
|
| 157 |
-
"epoch":
|
| 158 |
-
"grad_norm":
|
| 159 |
-
"learning_rate": 3.
|
| 160 |
-
"loss": 0.
|
| 161 |
"step": 3000
|
| 162 |
},
|
| 163 |
{
|
| 164 |
-
"epoch":
|
| 165 |
-
"eval_accuracy": 0.
|
| 166 |
-
"eval_f1_class_0": 0.
|
| 167 |
-
"eval_f1_class_1": 0.
|
| 168 |
-
"eval_f1_class_2": 0.
|
| 169 |
-
"eval_f1_class_3": 0.
|
| 170 |
-
"eval_f1_class_4": 0.
|
| 171 |
-
"eval_f1_class_5": 0.
|
| 172 |
-
"eval_f1_class_6": 0.
|
| 173 |
-
"eval_f1_class_7": 0.
|
| 174 |
-
"eval_f1_class_8": 0.
|
| 175 |
-
"
|
| 176 |
-
"
|
| 177 |
-
"
|
| 178 |
-
"
|
| 179 |
-
"
|
| 180 |
-
"
|
| 181 |
-
"
|
| 182 |
-
"
|
|
|
|
| 183 |
"step": 3000
|
| 184 |
},
|
| 185 |
{
|
| 186 |
-
"epoch":
|
| 187 |
-
"grad_norm":
|
| 188 |
-
"learning_rate":
|
| 189 |
-
"loss": 0.
|
| 190 |
"step": 3500
|
| 191 |
},
|
| 192 |
{
|
| 193 |
-
"epoch":
|
| 194 |
-
"eval_accuracy": 0.
|
| 195 |
-
"eval_f1_class_0": 0.
|
| 196 |
-
"eval_f1_class_1": 0.
|
| 197 |
-
"eval_f1_class_2": 0.
|
| 198 |
-
"eval_f1_class_3": 0.
|
| 199 |
-
"eval_f1_class_4": 0.
|
| 200 |
-
"eval_f1_class_5": 0.
|
| 201 |
-
"eval_f1_class_6": 0.
|
| 202 |
-
"eval_f1_class_7": 0.
|
| 203 |
-
"eval_f1_class_8": 0.
|
| 204 |
-
"
|
| 205 |
-
"
|
| 206 |
-
"
|
| 207 |
-
"
|
| 208 |
-
"
|
| 209 |
-
"
|
| 210 |
-
"
|
| 211 |
-
"
|
|
|
|
| 212 |
"step": 3500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 213 |
}
|
| 214 |
],
|
| 215 |
"logging_steps": 500,
|
| 216 |
-
"max_steps":
|
| 217 |
"num_input_tokens_seen": 0,
|
| 218 |
"num_train_epochs": 10,
|
| 219 |
"save_steps": 500,
|
|
@@ -229,7 +566,7 @@
|
|
| 229 |
"attributes": {}
|
| 230 |
}
|
| 231 |
},
|
| 232 |
-
"total_flos":
|
| 233 |
"train_batch_size": 16,
|
| 234 |
"trial_name": null,
|
| 235 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 9.433962264150944,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 9000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.5241090146750524,
|
| 13 |
+
"grad_norm": 14.80826473236084,
|
| 14 |
"learning_rate": 4.97e-05,
|
| 15 |
+
"loss": 1.8113,
|
| 16 |
"step": 500
|
| 17 |
},
|
| 18 |
{
|
| 19 |
+
"epoch": 0.5241090146750524,
|
| 20 |
+
"eval_accuracy": 0.6428945988463556,
|
| 21 |
+
"eval_f1_class_0": 0.632183908045977,
|
| 22 |
+
"eval_f1_class_1": 0.9472222222222223,
|
| 23 |
+
"eval_f1_class_2": 0.3657331136738056,
|
| 24 |
+
"eval_f1_class_3": 0.5542168674698795,
|
| 25 |
+
"eval_f1_class_4": 0.6165254237288136,
|
| 26 |
+
"eval_f1_class_5": 0.6261808367071524,
|
| 27 |
+
"eval_f1_class_6": 0.6833541927409262,
|
| 28 |
+
"eval_f1_class_7": 0.5899053627760252,
|
| 29 |
+
"eval_f1_class_8": 0.662280701754386,
|
| 30 |
+
"eval_f1_class_9": 0.6956521739130435,
|
| 31 |
+
"eval_f1_macro": 0.6373254803032232,
|
| 32 |
+
"eval_f1_micro": 0.6428945988463556,
|
| 33 |
+
"eval_f1_weighted": 0.6376303124409798,
|
| 34 |
+
"eval_loss": 1.1190885305404663,
|
| 35 |
+
"eval_recall_weighted": 0.6428945988463556,
|
| 36 |
+
"eval_runtime": 18.5604,
|
| 37 |
+
"eval_samples_per_second": 205.491,
|
| 38 |
+
"eval_steps_per_second": 6.465,
|
| 39 |
"step": 500
|
| 40 |
},
|
| 41 |
{
|
| 42 |
+
"epoch": 1.0482180293501049,
|
| 43 |
+
"grad_norm": 6.733876705169678,
|
| 44 |
+
"learning_rate": 4.725110619469027e-05,
|
| 45 |
+
"loss": 0.885,
|
| 46 |
"step": 1000
|
| 47 |
},
|
| 48 |
{
|
| 49 |
+
"epoch": 1.0482180293501049,
|
| 50 |
+
"eval_accuracy": 0.8185631882538018,
|
| 51 |
+
"eval_f1_class_0": 0.7974522292993631,
|
| 52 |
+
"eval_f1_class_1": 0.9651324965132496,
|
| 53 |
+
"eval_f1_class_2": 0.6306569343065694,
|
| 54 |
+
"eval_f1_class_3": 0.7193229901269393,
|
| 55 |
+
"eval_f1_class_4": 0.8341584158415841,
|
| 56 |
+
"eval_f1_class_5": 0.807061790668348,
|
| 57 |
+
"eval_f1_class_6": 0.8753246753246754,
|
| 58 |
+
"eval_f1_class_7": 0.8368794326241134,
|
| 59 |
+
"eval_f1_class_8": 0.8787446504992867,
|
| 60 |
+
"eval_f1_class_9": 0.8255528255528255,
|
| 61 |
+
"eval_f1_macro": 0.8170286440756953,
|
| 62 |
+
"eval_f1_micro": 0.8185631882538018,
|
| 63 |
+
"eval_f1_weighted": 0.8175319894578669,
|
| 64 |
+
"eval_loss": 0.6289433836936951,
|
| 65 |
+
"eval_recall_weighted": 0.8185631882538018,
|
| 66 |
+
"eval_runtime": 18.5795,
|
| 67 |
+
"eval_samples_per_second": 205.28,
|
| 68 |
+
"eval_steps_per_second": 6.459,
|
| 69 |
"step": 1000
|
| 70 |
},
|
| 71 |
{
|
| 72 |
+
"epoch": 1.5723270440251573,
|
| 73 |
+
"grad_norm": 21.78145408630371,
|
| 74 |
+
"learning_rate": 4.449115044247788e-05,
|
| 75 |
+
"loss": 0.4651,
|
| 76 |
"step": 1500
|
| 77 |
},
|
| 78 |
{
|
| 79 |
+
"epoch": 1.5723270440251573,
|
| 80 |
+
"eval_accuracy": 0.8615626638699528,
|
| 81 |
+
"eval_f1_class_0": 0.8236877523553162,
|
| 82 |
+
"eval_f1_class_1": 0.9653259361997226,
|
| 83 |
+
"eval_f1_class_2": 0.6986128625472887,
|
| 84 |
+
"eval_f1_class_3": 0.8308457711442785,
|
| 85 |
+
"eval_f1_class_4": 0.8667563930013459,
|
| 86 |
+
"eval_f1_class_5": 0.8346456692913385,
|
| 87 |
+
"eval_f1_class_6": 0.8898305084745762,
|
| 88 |
+
"eval_f1_class_7": 0.9136420525657072,
|
| 89 |
+
"eval_f1_class_8": 0.8668280871670703,
|
| 90 |
+
"eval_f1_class_9": 0.9410150891632374,
|
| 91 |
+
"eval_f1_macro": 0.8631190121909882,
|
| 92 |
+
"eval_f1_micro": 0.8615626638699528,
|
| 93 |
+
"eval_f1_weighted": 0.8631956720576096,
|
| 94 |
+
"eval_loss": 0.5186640024185181,
|
| 95 |
+
"eval_recall_weighted": 0.8615626638699528,
|
| 96 |
+
"eval_runtime": 18.5317,
|
| 97 |
+
"eval_samples_per_second": 205.809,
|
| 98 |
+
"eval_steps_per_second": 6.475,
|
| 99 |
"step": 1500
|
| 100 |
},
|
| 101 |
{
|
| 102 |
+
"epoch": 2.0964360587002098,
|
| 103 |
+
"grad_norm": 6.887778282165527,
|
| 104 |
+
"learning_rate": 4.172566371681416e-05,
|
| 105 |
+
"loss": 0.3292,
|
| 106 |
"step": 2000
|
| 107 |
},
|
| 108 |
{
|
| 109 |
+
"epoch": 2.0964360587002098,
|
| 110 |
+
"eval_accuracy": 0.9003670686942842,
|
| 111 |
+
"eval_f1_class_0": 0.8623188405797102,
|
| 112 |
+
"eval_f1_class_1": 0.9615384615384616,
|
| 113 |
+
"eval_f1_class_2": 0.7830045523520485,
|
| 114 |
+
"eval_f1_class_3": 0.9002624671916011,
|
| 115 |
+
"eval_f1_class_4": 0.8952618453865336,
|
| 116 |
+
"eval_f1_class_5": 0.8938271604938272,
|
| 117 |
+
"eval_f1_class_6": 0.9402173913043479,
|
| 118 |
+
"eval_f1_class_7": 0.9341935483870967,
|
| 119 |
+
"eval_f1_class_8": 0.9295039164490861,
|
| 120 |
+
"eval_f1_class_9": 0.8950131233595802,
|
| 121 |
+
"eval_f1_macro": 0.8995141307042294,
|
| 122 |
+
"eval_f1_micro": 0.9003670686942842,
|
| 123 |
+
"eval_f1_weighted": 0.899989630913346,
|
| 124 |
+
"eval_loss": 0.37930119037628174,
|
| 125 |
+
"eval_recall_weighted": 0.9003670686942842,
|
| 126 |
+
"eval_runtime": 18.5937,
|
| 127 |
+
"eval_samples_per_second": 205.123,
|
| 128 |
+
"eval_steps_per_second": 6.454,
|
| 129 |
"step": 2000
|
| 130 |
},
|
| 131 |
{
|
| 132 |
+
"epoch": 2.620545073375262,
|
| 133 |
+
"grad_norm": 18.899120330810547,
|
| 134 |
+
"learning_rate": 3.896017699115044e-05,
|
| 135 |
+
"loss": 0.2064,
|
| 136 |
"step": 2500
|
| 137 |
},
|
| 138 |
{
|
| 139 |
+
"epoch": 2.620545073375262,
|
| 140 |
+
"eval_accuracy": 0.916885159937074,
|
| 141 |
+
"eval_f1_class_0": 0.8660049627791564,
|
| 142 |
+
"eval_f1_class_1": 0.9670329670329672,
|
| 143 |
+
"eval_f1_class_2": 0.8037889039242219,
|
| 144 |
+
"eval_f1_class_3": 0.9132653061224489,
|
| 145 |
+
"eval_f1_class_4": 0.9252217997465145,
|
| 146 |
+
"eval_f1_class_5": 0.896719319562576,
|
| 147 |
+
"eval_f1_class_6": 0.9539295392953929,
|
| 148 |
+
"eval_f1_class_7": 0.9450000000000001,
|
| 149 |
+
"eval_f1_class_8": 0.9398601398601398,
|
| 150 |
+
"eval_f1_class_9": 0.9660056657223797,
|
| 151 |
+
"eval_f1_macro": 0.9176828604045797,
|
| 152 |
+
"eval_f1_micro": 0.916885159937074,
|
| 153 |
+
"eval_f1_weighted": 0.9176253934359663,
|
| 154 |
+
"eval_loss": 0.34075525403022766,
|
| 155 |
+
"eval_recall_weighted": 0.916885159937074,
|
| 156 |
+
"eval_runtime": 18.4847,
|
| 157 |
+
"eval_samples_per_second": 206.333,
|
| 158 |
+
"eval_steps_per_second": 6.492,
|
| 159 |
"step": 2500
|
| 160 |
},
|
| 161 |
{
|
| 162 |
+
"epoch": 3.1446540880503147,
|
| 163 |
+
"grad_norm": 1.2899887561798096,
|
| 164 |
+
"learning_rate": 3.619469026548672e-05,
|
| 165 |
+
"loss": 0.1562,
|
| 166 |
"step": 3000
|
| 167 |
},
|
| 168 |
{
|
| 169 |
+
"epoch": 3.1446540880503147,
|
| 170 |
+
"eval_accuracy": 0.9326166754063975,
|
| 171 |
+
"eval_f1_class_0": 0.8903061224489796,
|
| 172 |
+
"eval_f1_class_1": 0.9594594594594594,
|
| 173 |
+
"eval_f1_class_2": 0.8427128427128427,
|
| 174 |
+
"eval_f1_class_3": 0.9363867684478372,
|
| 175 |
+
"eval_f1_class_4": 0.9300000000000002,
|
| 176 |
+
"eval_f1_class_5": 0.9108433734939759,
|
| 177 |
+
"eval_f1_class_6": 0.9784366576819407,
|
| 178 |
+
"eval_f1_class_7": 0.9619289340101522,
|
| 179 |
+
"eval_f1_class_8": 0.9572192513368984,
|
| 180 |
+
"eval_f1_class_9": 0.9567642956764295,
|
| 181 |
+
"eval_f1_macro": 0.9324057705268516,
|
| 182 |
+
"eval_f1_micro": 0.9326166754063975,
|
| 183 |
+
"eval_f1_weighted": 0.9323951969454738,
|
| 184 |
+
"eval_loss": 0.29545533657073975,
|
| 185 |
+
"eval_recall_weighted": 0.9326166754063975,
|
| 186 |
+
"eval_runtime": 18.4597,
|
| 187 |
+
"eval_samples_per_second": 206.612,
|
| 188 |
+
"eval_steps_per_second": 6.501,
|
| 189 |
"step": 3000
|
| 190 |
},
|
| 191 |
{
|
| 192 |
+
"epoch": 3.668763102725367,
|
| 193 |
+
"grad_norm": 7.331942558288574,
|
| 194 |
+
"learning_rate": 3.342920353982301e-05,
|
| 195 |
+
"loss": 0.1097,
|
| 196 |
"step": 3500
|
| 197 |
},
|
| 198 |
{
|
| 199 |
+
"epoch": 3.668763102725367,
|
| 200 |
+
"eval_accuracy": 0.936811746198217,
|
| 201 |
+
"eval_f1_class_0": 0.8847497089639115,
|
| 202 |
+
"eval_f1_class_1": 0.967391304347826,
|
| 203 |
+
"eval_f1_class_2": 0.8646153846153847,
|
| 204 |
+
"eval_f1_class_3": 0.9375830013280213,
|
| 205 |
+
"eval_f1_class_4": 0.9284818067754077,
|
| 206 |
+
"eval_f1_class_5": 0.9315068493150684,
|
| 207 |
+
"eval_f1_class_6": 0.9733333333333333,
|
| 208 |
+
"eval_f1_class_7": 0.9563046192259677,
|
| 209 |
+
"eval_f1_class_8": 0.9731903485254693,
|
| 210 |
+
"eval_f1_class_9": 0.9495225102319237,
|
| 211 |
+
"eval_f1_macro": 0.9366678866662314,
|
| 212 |
+
"eval_f1_micro": 0.936811746198217,
|
| 213 |
+
"eval_f1_weighted": 0.9365912559521494,
|
| 214 |
+
"eval_loss": 0.3290887176990509,
|
| 215 |
+
"eval_recall_weighted": 0.936811746198217,
|
| 216 |
+
"eval_runtime": 18.5309,
|
| 217 |
+
"eval_samples_per_second": 205.818,
|
| 218 |
+
"eval_steps_per_second": 6.476,
|
| 219 |
"step": 3500
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 4.1928721174004195,
|
| 223 |
+
"grad_norm": 0.010240758769214153,
|
| 224 |
+
"learning_rate": 3.066924778761062e-05,
|
| 225 |
+
"loss": 0.1123,
|
| 226 |
+
"step": 4000
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 4.1928721174004195,
|
| 230 |
+
"eval_accuracy": 0.9431043523859465,
|
| 231 |
+
"eval_f1_class_0": 0.9055118110236221,
|
| 232 |
+
"eval_f1_class_1": 0.9780821917808219,
|
| 233 |
+
"eval_f1_class_2": 0.8661870503597122,
|
| 234 |
+
"eval_f1_class_3": 0.9431524547803619,
|
| 235 |
+
"eval_f1_class_4": 0.9369592088998764,
|
| 236 |
+
"eval_f1_class_5": 0.9245742092457422,
|
| 237 |
+
"eval_f1_class_6": 0.9734042553191489,
|
| 238 |
+
"eval_f1_class_7": 0.958904109589041,
|
| 239 |
+
"eval_f1_class_8": 0.9736842105263158,
|
| 240 |
+
"eval_f1_class_9": 0.9680998613037449,
|
| 241 |
+
"eval_f1_macro": 0.9428559362828388,
|
| 242 |
+
"eval_f1_micro": 0.9431043523859465,
|
| 243 |
+
"eval_f1_weighted": 0.9427303581875908,
|
| 244 |
+
"eval_loss": 0.3067891001701355,
|
| 245 |
+
"eval_recall_weighted": 0.9431043523859465,
|
| 246 |
+
"eval_runtime": 18.589,
|
| 247 |
+
"eval_samples_per_second": 205.175,
|
| 248 |
+
"eval_steps_per_second": 6.455,
|
| 249 |
+
"step": 4000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 4.716981132075472,
|
| 253 |
+
"grad_norm": 0.4483562707901001,
|
| 254 |
+
"learning_rate": 2.7903761061946903e-05,
|
| 255 |
+
"loss": 0.0732,
|
| 256 |
+
"step": 4500
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 4.716981132075472,
|
| 260 |
+
"eval_accuracy": 0.940744625065548,
|
| 261 |
+
"eval_f1_class_0": 0.8982826948480845,
|
| 262 |
+
"eval_f1_class_1": 0.9741496598639455,
|
| 263 |
+
"eval_f1_class_2": 0.8587570621468926,
|
| 264 |
+
"eval_f1_class_3": 0.940127388535032,
|
| 265 |
+
"eval_f1_class_4": 0.9477707006369427,
|
| 266 |
+
"eval_f1_class_5": 0.9290012033694345,
|
| 267 |
+
"eval_f1_class_6": 0.972972972972973,
|
| 268 |
+
"eval_f1_class_7": 0.9485294117647058,
|
| 269 |
+
"eval_f1_class_8": 0.9693741677762983,
|
| 270 |
+
"eval_f1_class_9": 0.9666666666666666,
|
| 271 |
+
"eval_f1_macro": 0.9405631928580975,
|
| 272 |
+
"eval_f1_micro": 0.940744625065548,
|
| 273 |
+
"eval_f1_weighted": 0.9405207858122142,
|
| 274 |
+
"eval_loss": 0.3208290636539459,
|
| 275 |
+
"eval_recall_weighted": 0.940744625065548,
|
| 276 |
+
"eval_runtime": 18.5774,
|
| 277 |
+
"eval_samples_per_second": 205.303,
|
| 278 |
+
"eval_steps_per_second": 6.459,
|
| 279 |
+
"step": 4500
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 5.241090146750524,
|
| 283 |
+
"grad_norm": 0.00789484940469265,
|
| 284 |
+
"learning_rate": 2.5138274336283185e-05,
|
| 285 |
+
"loss": 0.0582,
|
| 286 |
+
"step": 5000
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 5.241090146750524,
|
| 290 |
+
"eval_accuracy": 0.9444153120083901,
|
| 291 |
+
"eval_f1_class_0": 0.8981132075471698,
|
| 292 |
+
"eval_f1_class_1": 0.9780821917808219,
|
| 293 |
+
"eval_f1_class_2": 0.8748241912798874,
|
| 294 |
+
"eval_f1_class_3": 0.9479166666666665,
|
| 295 |
+
"eval_f1_class_4": 0.9436795994993743,
|
| 296 |
+
"eval_f1_class_5": 0.942189421894219,
|
| 297 |
+
"eval_f1_class_6": 0.9613259668508287,
|
| 298 |
+
"eval_f1_class_7": 0.9551122194513716,
|
| 299 |
+
"eval_f1_class_8": 0.9713541666666666,
|
| 300 |
+
"eval_f1_class_9": 0.9721448467966574,
|
| 301 |
+
"eval_f1_macro": 0.9444742478433662,
|
| 302 |
+
"eval_f1_micro": 0.9444153120083901,
|
| 303 |
+
"eval_f1_weighted": 0.9443976005879655,
|
| 304 |
+
"eval_loss": 0.3178301155567169,
|
| 305 |
+
"eval_recall_weighted": 0.9444153120083901,
|
| 306 |
+
"eval_runtime": 18.6189,
|
| 307 |
+
"eval_samples_per_second": 204.845,
|
| 308 |
+
"eval_steps_per_second": 6.445,
|
| 309 |
+
"step": 5000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 5.765199161425577,
|
| 313 |
+
"grad_norm": 0.2585061490535736,
|
| 314 |
+
"learning_rate": 2.237278761061947e-05,
|
| 315 |
+
"loss": 0.0533,
|
| 316 |
+
"step": 5500
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 5.765199161425577,
|
| 320 |
+
"eval_accuracy": 0.951232302045097,
|
| 321 |
+
"eval_f1_class_0": 0.9144316730523627,
|
| 322 |
+
"eval_f1_class_1": 0.9754768392370572,
|
| 323 |
+
"eval_f1_class_2": 0.8732782369146006,
|
| 324 |
+
"eval_f1_class_3": 0.9662337662337663,
|
| 325 |
+
"eval_f1_class_4": 0.958974358974359,
|
| 326 |
+
"eval_f1_class_5": 0.9410288582183187,
|
| 327 |
+
"eval_f1_class_6": 0.9840848806366048,
|
| 328 |
+
"eval_f1_class_7": 0.9597989949748743,
|
| 329 |
+
"eval_f1_class_8": 0.9712793733681462,
|
| 330 |
+
"eval_f1_class_9": 0.9667590027700832,
|
| 331 |
+
"eval_f1_macro": 0.9511345984380173,
|
| 332 |
+
"eval_f1_micro": 0.9512323020450971,
|
| 333 |
+
"eval_f1_weighted": 0.9512610025073973,
|
| 334 |
+
"eval_loss": 0.30544909834861755,
|
| 335 |
+
"eval_recall_weighted": 0.951232302045097,
|
| 336 |
+
"eval_runtime": 18.5675,
|
| 337 |
+
"eval_samples_per_second": 205.412,
|
| 338 |
+
"eval_steps_per_second": 6.463,
|
| 339 |
+
"step": 5500
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 6.289308176100629,
|
| 343 |
+
"grad_norm": 0.020399658009409904,
|
| 344 |
+
"learning_rate": 1.9607300884955755e-05,
|
| 345 |
+
"loss": 0.0424,
|
| 346 |
+
"step": 6000
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 6.289308176100629,
|
| 350 |
+
"eval_accuracy": 0.9520188778185632,
|
| 351 |
+
"eval_f1_class_0": 0.9041450777202072,
|
| 352 |
+
"eval_f1_class_1": 0.9782016348773842,
|
| 353 |
+
"eval_f1_class_2": 0.8834019204389574,
|
| 354 |
+
"eval_f1_class_3": 0.9584415584415584,
|
| 355 |
+
"eval_f1_class_4": 0.9491094147582698,
|
| 356 |
+
"eval_f1_class_5": 0.9521472392638036,
|
| 357 |
+
"eval_f1_class_6": 0.9824086603518267,
|
| 358 |
+
"eval_f1_class_7": 0.9638854296388544,
|
| 359 |
+
"eval_f1_class_8": 0.9814323607427056,
|
| 360 |
+
"eval_f1_class_9": 0.9669421487603306,
|
| 361 |
+
"eval_f1_macro": 0.9520115444993896,
|
| 362 |
+
"eval_f1_micro": 0.9520188778185632,
|
| 363 |
+
"eval_f1_weighted": 0.9520110205225185,
|
| 364 |
+
"eval_loss": 0.30924680829048157,
|
| 365 |
+
"eval_recall_weighted": 0.9520188778185632,
|
| 366 |
+
"eval_runtime": 18.5678,
|
| 367 |
+
"eval_samples_per_second": 205.41,
|
| 368 |
+
"eval_steps_per_second": 6.463,
|
| 369 |
+
"step": 6000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 6.813417190775681,
|
| 373 |
+
"grad_norm": 3.3116018772125244,
|
| 374 |
+
"learning_rate": 1.6841814159292034e-05,
|
| 375 |
+
"loss": 0.0378,
|
| 376 |
+
"step": 6500
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 6.813417190775681,
|
| 380 |
+
"eval_accuracy": 0.9530676455165181,
|
| 381 |
+
"eval_f1_class_0": 0.9155107187894074,
|
| 382 |
+
"eval_f1_class_1": 0.9768707482993196,
|
| 383 |
+
"eval_f1_class_2": 0.8838526912181304,
|
| 384 |
+
"eval_f1_class_3": 0.9488491048593349,
|
| 385 |
+
"eval_f1_class_4": 0.951885565669701,
|
| 386 |
+
"eval_f1_class_5": 0.9431680773881501,
|
| 387 |
+
"eval_f1_class_6": 0.9813829787234043,
|
| 388 |
+
"eval_f1_class_7": 0.9669211195928754,
|
| 389 |
+
"eval_f1_class_8": 0.9868073878627968,
|
| 390 |
+
"eval_f1_class_9": 0.9750000000000001,
|
| 391 |
+
"eval_f1_macro": 0.953024839240312,
|
| 392 |
+
"eval_f1_micro": 0.9530676455165181,
|
| 393 |
+
"eval_f1_weighted": 0.952958033459228,
|
| 394 |
+
"eval_loss": 0.30203908681869507,
|
| 395 |
+
"eval_recall_weighted": 0.9530676455165181,
|
| 396 |
+
"eval_runtime": 18.5121,
|
| 397 |
+
"eval_samples_per_second": 206.028,
|
| 398 |
+
"eval_steps_per_second": 6.482,
|
| 399 |
+
"step": 6500
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 7.337526205450734,
|
| 403 |
+
"grad_norm": 0.005160727072507143,
|
| 404 |
+
"learning_rate": 1.4081858407079645e-05,
|
| 405 |
+
"loss": 0.0305,
|
| 406 |
+
"step": 7000
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 7.337526205450734,
|
| 410 |
+
"eval_accuracy": 0.954116413214473,
|
| 411 |
+
"eval_f1_class_0": 0.9171817058096415,
|
| 412 |
+
"eval_f1_class_1": 0.9822646657571624,
|
| 413 |
+
"eval_f1_class_2": 0.8885672937771346,
|
| 414 |
+
"eval_f1_class_3": 0.95822454308094,
|
| 415 |
+
"eval_f1_class_4": 0.952020202020202,
|
| 416 |
+
"eval_f1_class_5": 0.9428918590522478,
|
| 417 |
+
"eval_f1_class_6": 0.9761273209549071,
|
| 418 |
+
"eval_f1_class_7": 0.9647355163727961,
|
| 419 |
+
"eval_f1_class_8": 0.9866310160427807,
|
| 420 |
+
"eval_f1_class_9": 0.9721448467966574,
|
| 421 |
+
"eval_f1_macro": 0.9540788969664471,
|
| 422 |
+
"eval_f1_micro": 0.954116413214473,
|
| 423 |
+
"eval_f1_weighted": 0.9540130488123523,
|
| 424 |
+
"eval_loss": 0.28602519631385803,
|
| 425 |
+
"eval_recall_weighted": 0.954116413214473,
|
| 426 |
+
"eval_runtime": 18.523,
|
| 427 |
+
"eval_samples_per_second": 205.906,
|
| 428 |
+
"eval_steps_per_second": 6.478,
|
| 429 |
+
"step": 7000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 7.861635220125786,
|
| 433 |
+
"grad_norm": 0.02748439833521843,
|
| 434 |
+
"learning_rate": 1.131637168141593e-05,
|
| 435 |
+
"loss": 0.0216,
|
| 436 |
+
"step": 7500
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 7.861635220125786,
|
| 440 |
+
"eval_accuracy": 0.9562139486103828,
|
| 441 |
+
"eval_f1_class_0": 0.9250317662007626,
|
| 442 |
+
"eval_f1_class_1": 0.9794801641586868,
|
| 443 |
+
"eval_f1_class_2": 0.8888888888888887,
|
| 444 |
+
"eval_f1_class_3": 0.9527458492975734,
|
| 445 |
+
"eval_f1_class_4": 0.9604086845466155,
|
| 446 |
+
"eval_f1_class_5": 0.9476248477466506,
|
| 447 |
+
"eval_f1_class_6": 0.9853137516688919,
|
| 448 |
+
"eval_f1_class_7": 0.9682337992376113,
|
| 449 |
+
"eval_f1_class_8": 0.9790575916230366,
|
| 450 |
+
"eval_f1_class_9": 0.9736477115117891,
|
| 451 |
+
"eval_f1_macro": 0.9560433054880507,
|
| 452 |
+
"eval_f1_micro": 0.9562139486103828,
|
| 453 |
+
"eval_f1_weighted": 0.956078385428201,
|
| 454 |
+
"eval_loss": 0.303732693195343,
|
| 455 |
+
"eval_recall_weighted": 0.9562139486103828,
|
| 456 |
+
"eval_runtime": 18.6679,
|
| 457 |
+
"eval_samples_per_second": 204.308,
|
| 458 |
+
"eval_steps_per_second": 6.428,
|
| 459 |
+
"step": 7500
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 8.385744234800839,
|
| 463 |
+
"grad_norm": 0.017909903079271317,
|
| 464 |
+
"learning_rate": 8.550884955752212e-06,
|
| 465 |
+
"loss": 0.0198,
|
| 466 |
+
"step": 8000
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 8.385744234800839,
|
| 470 |
+
"eval_accuracy": 0.9562139486103828,
|
| 471 |
+
"eval_f1_class_0": 0.9207547169811321,
|
| 472 |
+
"eval_f1_class_1": 0.9795361527967258,
|
| 473 |
+
"eval_f1_class_2": 0.8829337094499294,
|
| 474 |
+
"eval_f1_class_3": 0.9623865110246432,
|
| 475 |
+
"eval_f1_class_4": 0.9592875318066159,
|
| 476 |
+
"eval_f1_class_5": 0.9451887941534713,
|
| 477 |
+
"eval_f1_class_6": 0.9826897470039946,
|
| 478 |
+
"eval_f1_class_7": 0.9707006369426752,
|
| 479 |
+
"eval_f1_class_8": 0.9855072463768115,
|
| 480 |
+
"eval_f1_class_9": 0.9721448467966574,
|
| 481 |
+
"eval_f1_macro": 0.9561129893332657,
|
| 482 |
+
"eval_f1_micro": 0.9562139486103828,
|
| 483 |
+
"eval_f1_weighted": 0.9561806063383737,
|
| 484 |
+
"eval_loss": 0.2868480682373047,
|
| 485 |
+
"eval_recall_weighted": 0.9562139486103828,
|
| 486 |
+
"eval_runtime": 18.5749,
|
| 487 |
+
"eval_samples_per_second": 205.331,
|
| 488 |
+
"eval_steps_per_second": 6.46,
|
| 489 |
+
"step": 8000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 8.90985324947589,
|
| 493 |
+
"grad_norm": 0.019022395834326744,
|
| 494 |
+
"learning_rate": 5.785398230088496e-06,
|
| 495 |
+
"loss": 0.0145,
|
| 496 |
+
"step": 8500
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 8.90985324947589,
|
| 500 |
+
"eval_accuracy": 0.9580492920818039,
|
| 501 |
+
"eval_f1_class_0": 0.9221260815822002,
|
| 502 |
+
"eval_f1_class_1": 0.9754768392370572,
|
| 503 |
+
"eval_f1_class_2": 0.8932584269662921,
|
| 504 |
+
"eval_f1_class_3": 0.9621903520208606,
|
| 505 |
+
"eval_f1_class_4": 0.9541984732824428,
|
| 506 |
+
"eval_f1_class_5": 0.957920792079208,
|
| 507 |
+
"eval_f1_class_6": 0.9879518072289156,
|
| 508 |
+
"eval_f1_class_7": 0.9671717171717172,
|
| 509 |
+
"eval_f1_class_8": 0.9867724867724869,
|
| 510 |
+
"eval_f1_class_9": 0.9735006973500697,
|
| 511 |
+
"eval_f1_macro": 0.9580567673691249,
|
| 512 |
+
"eval_f1_micro": 0.9580492920818039,
|
| 513 |
+
"eval_f1_weighted": 0.9581079328217166,
|
| 514 |
+
"eval_loss": 0.29160091280937195,
|
| 515 |
+
"eval_recall_weighted": 0.9580492920818039,
|
| 516 |
+
"eval_runtime": 18.5713,
|
| 517 |
+
"eval_samples_per_second": 205.371,
|
| 518 |
+
"eval_steps_per_second": 6.462,
|
| 519 |
+
"step": 8500
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 9.433962264150944,
|
| 523 |
+
"grad_norm": 0.0038110397290438414,
|
| 524 |
+
"learning_rate": 3.019911504424779e-06,
|
| 525 |
+
"loss": 0.014,
|
| 526 |
+
"step": 9000
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 9.433962264150944,
|
| 530 |
+
"eval_accuracy": 0.959884635553225,
|
| 531 |
+
"eval_f1_class_0": 0.9228855721393034,
|
| 532 |
+
"eval_f1_class_1": 0.9809264305177112,
|
| 533 |
+
"eval_f1_class_2": 0.9080459770114943,
|
| 534 |
+
"eval_f1_class_3": 0.9571984435797665,
|
| 535 |
+
"eval_f1_class_4": 0.9566326530612245,
|
| 536 |
+
"eval_f1_class_5": 0.9585365853658537,
|
| 537 |
+
"eval_f1_class_6": 0.9840848806366048,
|
| 538 |
+
"eval_f1_class_7": 0.9745547073791349,
|
| 539 |
+
"eval_f1_class_8": 0.9868073878627968,
|
| 540 |
+
"eval_f1_class_9": 0.9680998613037449,
|
| 541 |
+
"eval_f1_macro": 0.9597772498857635,
|
| 542 |
+
"eval_f1_micro": 0.959884635553225,
|
| 543 |
+
"eval_f1_weighted": 0.9597604965241316,
|
| 544 |
+
"eval_loss": 0.28916341066360474,
|
| 545 |
+
"eval_recall_weighted": 0.959884635553225,
|
| 546 |
+
"eval_runtime": 18.5578,
|
| 547 |
+
"eval_samples_per_second": 205.52,
|
| 548 |
+
"eval_steps_per_second": 6.466,
|
| 549 |
+
"step": 9000
|
| 550 |
}
|
| 551 |
],
|
| 552 |
"logging_steps": 500,
|
| 553 |
+
"max_steps": 9540,
|
| 554 |
"num_input_tokens_seen": 0,
|
| 555 |
"num_train_epochs": 10,
|
| 556 |
"save_steps": 500,
|
|
|
|
| 566 |
"attributes": {}
|
| 567 |
}
|
| 568 |
},
|
| 569 |
+
"total_flos": 3.787923857566925e+16,
|
| 570 |
"train_batch_size": 16,
|
| 571 |
"trial_name": null,
|
| 572 |
"trial_params": null
|