Upload plato model files
Browse files- config.json +25 -0
- generation_config.json +5 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +468 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/scratch/gpfs/BG11/logion-resources/cache/LOGION-50k_wordpiece_model",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForMaskedLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 3072,
|
| 13 |
+
"layer_norm_eps": 1e-12,
|
| 14 |
+
"max_position_embeddings": 512,
|
| 15 |
+
"model_type": "bert",
|
| 16 |
+
"num_attention_heads": 12,
|
| 17 |
+
"num_hidden_layers": 12,
|
| 18 |
+
"pad_token_id": 0,
|
| 19 |
+
"position_embedding_type": "absolute",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.48.3",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 50000
|
| 25 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"pad_token_id": 0,
|
| 4 |
+
"transformers_version": "4.48.3"
|
| 5 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b23563782a0cb54a67d2a8b1eac79eae5f8aceaf189fb581ffe3e2caaa485a82
|
| 3 |
+
size 497995232
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eb4e4e063a7c75a711530077a7a0a8d35451a372d9b10d89f3a33a2bd8ac9783
|
| 3 |
+
size 996111994
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f79f75d2ec77195528273b6bb84e6698cab949de9bcaa6d78c7f95a41fda801a
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4c5f5767035b5956a68d4d20c523c3fd1fec730443fa1db2421ec3c4bedf3e7d
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,468 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 1.8379791975021362,
|
| 3 |
+
"best_model_checkpoint": "/scratch/gpfs/BG11/plato/models-plato-lr/model_plato-phil_unaccented_text_10.txt/checkpoint-12586",
|
| 4 |
+
"epoch": 29.0,
|
| 5 |
+
"eval_steps": 108,
|
| 6 |
+
"global_step": 12586,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"grad_norm": 10.017090797424316,
|
| 14 |
+
"learning_rate": 9.950000000000001e-06,
|
| 15 |
+
"loss": 2.1548,
|
| 16 |
+
"step": 434
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 1.0,
|
| 20 |
+
"eval_loss": 1.9816173315048218,
|
| 21 |
+
"eval_runtime": 2.9153,
|
| 22 |
+
"eval_samples_per_second": 264.807,
|
| 23 |
+
"eval_steps_per_second": 33.272,
|
| 24 |
+
"step": 434
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 2.0,
|
| 28 |
+
"grad_norm": 10.668547630310059,
|
| 29 |
+
"learning_rate": 9.9e-06,
|
| 30 |
+
"loss": 2.0793,
|
| 31 |
+
"step": 868
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 2.0,
|
| 35 |
+
"eval_loss": 1.9505406618118286,
|
| 36 |
+
"eval_runtime": 2.9294,
|
| 37 |
+
"eval_samples_per_second": 263.535,
|
| 38 |
+
"eval_steps_per_second": 33.113,
|
| 39 |
+
"step": 868
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 3.0,
|
| 43 |
+
"grad_norm": 11.247385025024414,
|
| 44 |
+
"learning_rate": 9.85e-06,
|
| 45 |
+
"loss": 2.0625,
|
| 46 |
+
"step": 1302
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 3.0,
|
| 50 |
+
"eval_loss": 1.9273583889007568,
|
| 51 |
+
"eval_runtime": 2.9137,
|
| 52 |
+
"eval_samples_per_second": 264.954,
|
| 53 |
+
"eval_steps_per_second": 33.291,
|
| 54 |
+
"step": 1302
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 4.0,
|
| 58 |
+
"grad_norm": 9.709423065185547,
|
| 59 |
+
"learning_rate": 9.800000000000001e-06,
|
| 60 |
+
"loss": 2.0344,
|
| 61 |
+
"step": 1736
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 4.0,
|
| 65 |
+
"eval_loss": 1.9325604438781738,
|
| 66 |
+
"eval_runtime": 2.9068,
|
| 67 |
+
"eval_samples_per_second": 265.582,
|
| 68 |
+
"eval_steps_per_second": 33.37,
|
| 69 |
+
"step": 1736
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 5.0,
|
| 73 |
+
"grad_norm": 10.718835830688477,
|
| 74 |
+
"learning_rate": 9.75e-06,
|
| 75 |
+
"loss": 2.009,
|
| 76 |
+
"step": 2170
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 5.0,
|
| 80 |
+
"eval_loss": 1.9463964700698853,
|
| 81 |
+
"eval_runtime": 2.9049,
|
| 82 |
+
"eval_samples_per_second": 265.755,
|
| 83 |
+
"eval_steps_per_second": 33.391,
|
| 84 |
+
"step": 2170
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 6.0,
|
| 88 |
+
"grad_norm": 9.384885787963867,
|
| 89 |
+
"learning_rate": 9.7e-06,
|
| 90 |
+
"loss": 1.9735,
|
| 91 |
+
"step": 2604
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 6.0,
|
| 95 |
+
"eval_loss": 1.896444320678711,
|
| 96 |
+
"eval_runtime": 2.9325,
|
| 97 |
+
"eval_samples_per_second": 263.257,
|
| 98 |
+
"eval_steps_per_second": 33.078,
|
| 99 |
+
"step": 2604
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 7.0,
|
| 103 |
+
"grad_norm": 10.999958038330078,
|
| 104 |
+
"learning_rate": 9.65e-06,
|
| 105 |
+
"loss": 1.9355,
|
| 106 |
+
"step": 3038
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 7.0,
|
| 110 |
+
"eval_loss": 1.9304955005645752,
|
| 111 |
+
"eval_runtime": 2.9309,
|
| 112 |
+
"eval_samples_per_second": 263.399,
|
| 113 |
+
"eval_steps_per_second": 33.095,
|
| 114 |
+
"step": 3038
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 8.0,
|
| 118 |
+
"grad_norm": 10.52430534362793,
|
| 119 |
+
"learning_rate": 9.600000000000001e-06,
|
| 120 |
+
"loss": 1.9248,
|
| 121 |
+
"step": 3472
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 8.0,
|
| 125 |
+
"eval_loss": 1.9371752738952637,
|
| 126 |
+
"eval_runtime": 2.9216,
|
| 127 |
+
"eval_samples_per_second": 264.237,
|
| 128 |
+
"eval_steps_per_second": 33.201,
|
| 129 |
+
"step": 3472
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 9.0,
|
| 133 |
+
"grad_norm": 9.388583183288574,
|
| 134 |
+
"learning_rate": 9.55e-06,
|
| 135 |
+
"loss": 1.9172,
|
| 136 |
+
"step": 3906
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 9.0,
|
| 140 |
+
"eval_loss": 1.9058644771575928,
|
| 141 |
+
"eval_runtime": 2.9052,
|
| 142 |
+
"eval_samples_per_second": 265.727,
|
| 143 |
+
"eval_steps_per_second": 33.388,
|
| 144 |
+
"step": 3906
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 10.0,
|
| 148 |
+
"grad_norm": 9.09006118774414,
|
| 149 |
+
"learning_rate": 9.5e-06,
|
| 150 |
+
"loss": 1.8741,
|
| 151 |
+
"step": 4340
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 10.0,
|
| 155 |
+
"eval_loss": 1.8592596054077148,
|
| 156 |
+
"eval_runtime": 2.9141,
|
| 157 |
+
"eval_samples_per_second": 264.922,
|
| 158 |
+
"eval_steps_per_second": 33.287,
|
| 159 |
+
"step": 4340
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 11.0,
|
| 163 |
+
"grad_norm": 9.535520553588867,
|
| 164 |
+
"learning_rate": 9.450000000000001e-06,
|
| 165 |
+
"loss": 1.8718,
|
| 166 |
+
"step": 4774
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 11.0,
|
| 170 |
+
"eval_loss": 1.932305097579956,
|
| 171 |
+
"eval_runtime": 2.9127,
|
| 172 |
+
"eval_samples_per_second": 265.042,
|
| 173 |
+
"eval_steps_per_second": 33.302,
|
| 174 |
+
"step": 4774
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 12.0,
|
| 178 |
+
"grad_norm": 9.936247825622559,
|
| 179 |
+
"learning_rate": 9.4e-06,
|
| 180 |
+
"loss": 1.8406,
|
| 181 |
+
"step": 5208
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 12.0,
|
| 185 |
+
"eval_loss": 1.9319616556167603,
|
| 186 |
+
"eval_runtime": 2.9091,
|
| 187 |
+
"eval_samples_per_second": 265.374,
|
| 188 |
+
"eval_steps_per_second": 33.344,
|
| 189 |
+
"step": 5208
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 13.0,
|
| 193 |
+
"grad_norm": 9.894285202026367,
|
| 194 |
+
"learning_rate": 9.350000000000002e-06,
|
| 195 |
+
"loss": 1.8332,
|
| 196 |
+
"step": 5642
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 13.0,
|
| 200 |
+
"eval_loss": 1.9167243242263794,
|
| 201 |
+
"eval_runtime": 2.9034,
|
| 202 |
+
"eval_samples_per_second": 265.898,
|
| 203 |
+
"eval_steps_per_second": 33.41,
|
| 204 |
+
"step": 5642
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 14.0,
|
| 208 |
+
"grad_norm": 10.511141777038574,
|
| 209 |
+
"learning_rate": 9.3e-06,
|
| 210 |
+
"loss": 1.8258,
|
| 211 |
+
"step": 6076
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 14.0,
|
| 215 |
+
"eval_loss": 1.9056346416473389,
|
| 216 |
+
"eval_runtime": 2.9265,
|
| 217 |
+
"eval_samples_per_second": 263.8,
|
| 218 |
+
"eval_steps_per_second": 33.146,
|
| 219 |
+
"step": 6076
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 15.0,
|
| 223 |
+
"grad_norm": 11.17342758178711,
|
| 224 |
+
"learning_rate": 9.250000000000001e-06,
|
| 225 |
+
"loss": 1.8045,
|
| 226 |
+
"step": 6510
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 15.0,
|
| 230 |
+
"eval_loss": 1.8958162069320679,
|
| 231 |
+
"eval_runtime": 2.9122,
|
| 232 |
+
"eval_samples_per_second": 265.094,
|
| 233 |
+
"eval_steps_per_second": 33.308,
|
| 234 |
+
"step": 6510
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 16.0,
|
| 238 |
+
"grad_norm": 10.152159690856934,
|
| 239 |
+
"learning_rate": 9.200000000000002e-06,
|
| 240 |
+
"loss": 1.7862,
|
| 241 |
+
"step": 6944
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 16.0,
|
| 245 |
+
"eval_loss": 1.9146229028701782,
|
| 246 |
+
"eval_runtime": 2.9007,
|
| 247 |
+
"eval_samples_per_second": 266.146,
|
| 248 |
+
"eval_steps_per_second": 33.441,
|
| 249 |
+
"step": 6944
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 17.0,
|
| 253 |
+
"grad_norm": 9.818968772888184,
|
| 254 |
+
"learning_rate": 9.15e-06,
|
| 255 |
+
"loss": 1.7761,
|
| 256 |
+
"step": 7378
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 17.0,
|
| 260 |
+
"eval_loss": 1.9073622226715088,
|
| 261 |
+
"eval_runtime": 2.9044,
|
| 262 |
+
"eval_samples_per_second": 265.807,
|
| 263 |
+
"eval_steps_per_second": 33.398,
|
| 264 |
+
"step": 7378
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 18.0,
|
| 268 |
+
"grad_norm": 9.413209915161133,
|
| 269 |
+
"learning_rate": 9.100000000000001e-06,
|
| 270 |
+
"loss": 1.778,
|
| 271 |
+
"step": 7812
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 18.0,
|
| 275 |
+
"eval_loss": 1.9236599206924438,
|
| 276 |
+
"eval_runtime": 2.893,
|
| 277 |
+
"eval_samples_per_second": 266.85,
|
| 278 |
+
"eval_steps_per_second": 33.529,
|
| 279 |
+
"step": 7812
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 19.0,
|
| 283 |
+
"grad_norm": 9.5354642868042,
|
| 284 |
+
"learning_rate": 9.050000000000001e-06,
|
| 285 |
+
"loss": 1.7452,
|
| 286 |
+
"step": 8246
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 19.0,
|
| 290 |
+
"eval_loss": 1.899939775466919,
|
| 291 |
+
"eval_runtime": 2.91,
|
| 292 |
+
"eval_samples_per_second": 265.293,
|
| 293 |
+
"eval_steps_per_second": 33.334,
|
| 294 |
+
"step": 8246
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 20.0,
|
| 298 |
+
"grad_norm": 11.661208152770996,
|
| 299 |
+
"learning_rate": 9e-06,
|
| 300 |
+
"loss": 1.718,
|
| 301 |
+
"step": 8680
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 20.0,
|
| 305 |
+
"eval_loss": 1.9721119403839111,
|
| 306 |
+
"eval_runtime": 2.9167,
|
| 307 |
+
"eval_samples_per_second": 264.683,
|
| 308 |
+
"eval_steps_per_second": 33.257,
|
| 309 |
+
"step": 8680
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 21.0,
|
| 313 |
+
"grad_norm": 11.493612289428711,
|
| 314 |
+
"learning_rate": 8.95e-06,
|
| 315 |
+
"loss": 1.7184,
|
| 316 |
+
"step": 9114
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 21.0,
|
| 320 |
+
"eval_loss": 1.916871428489685,
|
| 321 |
+
"eval_runtime": 2.9076,
|
| 322 |
+
"eval_samples_per_second": 265.509,
|
| 323 |
+
"eval_steps_per_second": 33.361,
|
| 324 |
+
"step": 9114
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 22.0,
|
| 328 |
+
"grad_norm": 11.153796195983887,
|
| 329 |
+
"learning_rate": 8.900000000000001e-06,
|
| 330 |
+
"loss": 1.7141,
|
| 331 |
+
"step": 9548
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 22.0,
|
| 335 |
+
"eval_loss": 1.8576736450195312,
|
| 336 |
+
"eval_runtime": 2.9127,
|
| 337 |
+
"eval_samples_per_second": 265.049,
|
| 338 |
+
"eval_steps_per_second": 33.303,
|
| 339 |
+
"step": 9548
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 23.0,
|
| 343 |
+
"grad_norm": 10.186408042907715,
|
| 344 |
+
"learning_rate": 8.85e-06,
|
| 345 |
+
"loss": 1.7025,
|
| 346 |
+
"step": 9982
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 23.0,
|
| 350 |
+
"eval_loss": 1.9040586948394775,
|
| 351 |
+
"eval_runtime": 2.9122,
|
| 352 |
+
"eval_samples_per_second": 265.094,
|
| 353 |
+
"eval_steps_per_second": 33.308,
|
| 354 |
+
"step": 9982
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 24.0,
|
| 358 |
+
"grad_norm": 10.842204093933105,
|
| 359 |
+
"learning_rate": 8.8e-06,
|
| 360 |
+
"loss": 1.6861,
|
| 361 |
+
"step": 10416
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 24.0,
|
| 365 |
+
"eval_loss": 1.8773962259292603,
|
| 366 |
+
"eval_runtime": 2.9089,
|
| 367 |
+
"eval_samples_per_second": 265.391,
|
| 368 |
+
"eval_steps_per_second": 33.346,
|
| 369 |
+
"step": 10416
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 25.0,
|
| 373 |
+
"grad_norm": 10.846117973327637,
|
| 374 |
+
"learning_rate": 8.750000000000001e-06,
|
| 375 |
+
"loss": 1.6627,
|
| 376 |
+
"step": 10850
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 25.0,
|
| 380 |
+
"eval_loss": 1.8711638450622559,
|
| 381 |
+
"eval_runtime": 2.912,
|
| 382 |
+
"eval_samples_per_second": 265.113,
|
| 383 |
+
"eval_steps_per_second": 33.311,
|
| 384 |
+
"step": 10850
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 26.0,
|
| 388 |
+
"grad_norm": 11.196520805358887,
|
| 389 |
+
"learning_rate": 8.700000000000001e-06,
|
| 390 |
+
"loss": 1.6625,
|
| 391 |
+
"step": 11284
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 26.0,
|
| 395 |
+
"eval_loss": 1.946097731590271,
|
| 396 |
+
"eval_runtime": 2.9133,
|
| 397 |
+
"eval_samples_per_second": 264.989,
|
| 398 |
+
"eval_steps_per_second": 33.295,
|
| 399 |
+
"step": 11284
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 27.0,
|
| 403 |
+
"grad_norm": 9.261242866516113,
|
| 404 |
+
"learning_rate": 8.65e-06,
|
| 405 |
+
"loss": 1.6464,
|
| 406 |
+
"step": 11718
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 27.0,
|
| 410 |
+
"eval_loss": 1.888836145401001,
|
| 411 |
+
"eval_runtime": 2.9213,
|
| 412 |
+
"eval_samples_per_second": 264.264,
|
| 413 |
+
"eval_steps_per_second": 33.204,
|
| 414 |
+
"step": 11718
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 28.0,
|
| 418 |
+
"grad_norm": 8.242238998413086,
|
| 419 |
+
"learning_rate": 8.6e-06,
|
| 420 |
+
"loss": 1.625,
|
| 421 |
+
"step": 12152
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"epoch": 28.0,
|
| 425 |
+
"eval_loss": 1.9059230089187622,
|
| 426 |
+
"eval_runtime": 2.9279,
|
| 427 |
+
"eval_samples_per_second": 263.673,
|
| 428 |
+
"eval_steps_per_second": 33.13,
|
| 429 |
+
"step": 12152
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 29.0,
|
| 433 |
+
"grad_norm": 10.768765449523926,
|
| 434 |
+
"learning_rate": 8.550000000000001e-06,
|
| 435 |
+
"loss": 1.6194,
|
| 436 |
+
"step": 12586
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 29.0,
|
| 440 |
+
"eval_loss": 1.8379791975021362,
|
| 441 |
+
"eval_runtime": 2.91,
|
| 442 |
+
"eval_samples_per_second": 265.295,
|
| 443 |
+
"eval_steps_per_second": 33.334,
|
| 444 |
+
"step": 12586
|
| 445 |
+
}
|
| 446 |
+
],
|
| 447 |
+
"logging_steps": 108,
|
| 448 |
+
"max_steps": 86800,
|
| 449 |
+
"num_input_tokens_seen": 0,
|
| 450 |
+
"num_train_epochs": 200,
|
| 451 |
+
"save_steps": 108,
|
| 452 |
+
"stateful_callbacks": {
|
| 453 |
+
"TrainerControl": {
|
| 454 |
+
"args": {
|
| 455 |
+
"should_epoch_stop": false,
|
| 456 |
+
"should_evaluate": false,
|
| 457 |
+
"should_log": false,
|
| 458 |
+
"should_save": true,
|
| 459 |
+
"should_training_stop": false
|
| 460 |
+
},
|
| 461 |
+
"attributes": {}
|
| 462 |
+
}
|
| 463 |
+
},
|
| 464 |
+
"total_flos": 5.29846435273769e+16,
|
| 465 |
+
"train_batch_size": 16,
|
| 466 |
+
"trial_name": null,
|
| 467 |
+
"trial_params": null
|
| 468 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:69f375ae1f316f8dded0bd6962a67bed1da28c220664d46ed1fc1e4503897e07
|
| 3 |
+
size 5368
|