Model save
Browse files- README.md +69 -0
- all_results.json +13 -0
- config.json +228 -0
- eval_results.json +8 -0
- model.safetensors +3 -0
- preprocessor_config.json +28 -0
- runs/Mar02_13-37-49_9a8dc51c99f8/events.out.tfevents.1709386701.9a8dc51c99f8.326.0 +3 -0
- runs/Mar02_13-46-44_6f1f307693d9/events.out.tfevents.1709387309.6f1f307693d9.258.0 +3 -0
- runs/Mar02_13-56-30_966e26ef2d5b/events.out.tfevents.1709387957.966e26ef2d5b.289.0 +3 -0
- runs/Mar02_13-56-30_966e26ef2d5b/events.out.tfevents.1709388178.966e26ef2d5b.289.1 +3 -0
- runs/Mar02_14-41-50_04f752329dca/events.out.tfevents.1709390567.04f752329dca.283.0 +3 -0
- train_results.json +8 -0
- trainer_state.json +1056 -0
- training_args.bin +3 -0
README.md
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: facebook/deit-base-distilled-patch16-224
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
metrics:
|
| 7 |
+
- accuracy
|
| 8 |
+
model-index:
|
| 9 |
+
- name: base
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# base
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [facebook/deit-base-distilled-patch16-224](https://huggingface.co/facebook/deit-base-distilled-patch16-224) on an unknown dataset.
|
| 19 |
+
It achieves the following results on the evaluation set:
|
| 20 |
+
- Accuracy: 0.906
|
| 21 |
+
- Loss: 0.3054
|
| 22 |
+
|
| 23 |
+
## Model description
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Intended uses & limitations
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training and evaluation data
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Training procedure
|
| 36 |
+
|
| 37 |
+
### Training hyperparameters
|
| 38 |
+
|
| 39 |
+
The following hyperparameters were used during training:
|
| 40 |
+
- learning_rate: 5e-05
|
| 41 |
+
- train_batch_size: 64
|
| 42 |
+
- eval_batch_size: 1
|
| 43 |
+
- seed: 777
|
| 44 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 45 |
+
- lr_scheduler_type: linear
|
| 46 |
+
- num_epochs: 10
|
| 47 |
+
|
| 48 |
+
### Training results
|
| 49 |
+
|
| 50 |
+
| Training Loss | Epoch | Step | Accuracy | Validation Loss |
|
| 51 |
+
|:-------------:|:-----:|:----:|:--------:|:---------------:|
|
| 52 |
+
| 1.1232 | 1.0 | 782 | 0.8416 | 0.5390 |
|
| 53 |
+
| 0.9017 | 2.0 | 1564 | 0.8699 | 0.4365 |
|
| 54 |
+
| 0.7565 | 3.0 | 2346 | 0.8858 | 0.3678 |
|
| 55 |
+
| 0.706 | 4.0 | 3128 | 0.8952 | 0.3446 |
|
| 56 |
+
| 0.6353 | 5.0 | 3910 | 0.8986 | 0.3331 |
|
| 57 |
+
| 0.5384 | 6.0 | 4692 | 0.9001 | 0.3223 |
|
| 58 |
+
| 0.5004 | 7.0 | 5474 | 0.9018 | 0.3249 |
|
| 59 |
+
| 0.4672 | 8.0 | 6256 | 0.904 | 0.3113 |
|
| 60 |
+
| 0.4526 | 9.0 | 7038 | 0.9054 | 0.3081 |
|
| 61 |
+
| 0.4289 | 10.0 | 7820 | 0.906 | 0.3054 |
|
| 62 |
+
|
| 63 |
+
|
| 64 |
+
### Framework versions
|
| 65 |
+
|
| 66 |
+
- Transformers 4.38.1
|
| 67 |
+
- Pytorch 2.1.0+cu121
|
| 68 |
+
- Datasets 2.18.0
|
| 69 |
+
- Tokenizers 0.15.2
|
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.906,
|
| 4 |
+
"eval_loss": 0.3053509294986725,
|
| 5 |
+
"eval_runtime": 173.1395,
|
| 6 |
+
"eval_samples_per_second": 57.757,
|
| 7 |
+
"eval_steps_per_second": 57.757,
|
| 8 |
+
"total_flos": 3.8815447523328e+19,
|
| 9 |
+
"train_loss": 0.0,
|
| 10 |
+
"train_runtime": 24.522,
|
| 11 |
+
"train_samples_per_second": 20389.834,
|
| 12 |
+
"train_steps_per_second": 318.897
|
| 13 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,228 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/deit-base-distilled-patch16-224",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"DeiTForImageClassificationWithTeacher"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"encoder_stride": 16,
|
| 8 |
+
"finetuning_task": "image-classification",
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.0,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "apple",
|
| 14 |
+
"1": "aquarium_fish",
|
| 15 |
+
"10": "bowl",
|
| 16 |
+
"11": "boy",
|
| 17 |
+
"12": "bridge",
|
| 18 |
+
"13": "bus",
|
| 19 |
+
"14": "butterfly",
|
| 20 |
+
"15": "camel",
|
| 21 |
+
"16": "can",
|
| 22 |
+
"17": "castle",
|
| 23 |
+
"18": "caterpillar",
|
| 24 |
+
"19": "cattle",
|
| 25 |
+
"2": "baby",
|
| 26 |
+
"20": "chair",
|
| 27 |
+
"21": "chimpanzee",
|
| 28 |
+
"22": "clock",
|
| 29 |
+
"23": "cloud",
|
| 30 |
+
"24": "cockroach",
|
| 31 |
+
"25": "couch",
|
| 32 |
+
"26": "cra",
|
| 33 |
+
"27": "crocodile",
|
| 34 |
+
"28": "cup",
|
| 35 |
+
"29": "dinosaur",
|
| 36 |
+
"3": "bear",
|
| 37 |
+
"30": "dolphin",
|
| 38 |
+
"31": "elephant",
|
| 39 |
+
"32": "flatfish",
|
| 40 |
+
"33": "forest",
|
| 41 |
+
"34": "fox",
|
| 42 |
+
"35": "girl",
|
| 43 |
+
"36": "hamster",
|
| 44 |
+
"37": "house",
|
| 45 |
+
"38": "kangaroo",
|
| 46 |
+
"39": "keyboard",
|
| 47 |
+
"4": "beaver",
|
| 48 |
+
"40": "lamp",
|
| 49 |
+
"41": "lawn_mower",
|
| 50 |
+
"42": "leopard",
|
| 51 |
+
"43": "lion",
|
| 52 |
+
"44": "lizard",
|
| 53 |
+
"45": "lobster",
|
| 54 |
+
"46": "man",
|
| 55 |
+
"47": "maple_tree",
|
| 56 |
+
"48": "motorcycle",
|
| 57 |
+
"49": "mountain",
|
| 58 |
+
"5": "bed",
|
| 59 |
+
"50": "mouse",
|
| 60 |
+
"51": "mushroom",
|
| 61 |
+
"52": "oak_tree",
|
| 62 |
+
"53": "orange",
|
| 63 |
+
"54": "orchid",
|
| 64 |
+
"55": "otter",
|
| 65 |
+
"56": "palm_tree",
|
| 66 |
+
"57": "pear",
|
| 67 |
+
"58": "pickup_truck",
|
| 68 |
+
"59": "pine_tree",
|
| 69 |
+
"6": "bee",
|
| 70 |
+
"60": "plain",
|
| 71 |
+
"61": "plate",
|
| 72 |
+
"62": "poppy",
|
| 73 |
+
"63": "porcupine",
|
| 74 |
+
"64": "possum",
|
| 75 |
+
"65": "rabbit",
|
| 76 |
+
"66": "raccoon",
|
| 77 |
+
"67": "ray",
|
| 78 |
+
"68": "road",
|
| 79 |
+
"69": "rocket",
|
| 80 |
+
"7": "beetle",
|
| 81 |
+
"70": "rose",
|
| 82 |
+
"71": "sea",
|
| 83 |
+
"72": "seal",
|
| 84 |
+
"73": "shark",
|
| 85 |
+
"74": "shrew",
|
| 86 |
+
"75": "skunk",
|
| 87 |
+
"76": "skyscraper",
|
| 88 |
+
"77": "snail",
|
| 89 |
+
"78": "snake",
|
| 90 |
+
"79": "spider",
|
| 91 |
+
"8": "bicycle",
|
| 92 |
+
"80": "squirrel",
|
| 93 |
+
"81": "streetcar",
|
| 94 |
+
"82": "sunflower",
|
| 95 |
+
"83": "sweet_pepper",
|
| 96 |
+
"84": "table",
|
| 97 |
+
"85": "tank",
|
| 98 |
+
"86": "telephone",
|
| 99 |
+
"87": "television",
|
| 100 |
+
"88": "tiger",
|
| 101 |
+
"89": "tractor",
|
| 102 |
+
"9": "bottle",
|
| 103 |
+
"90": "train",
|
| 104 |
+
"91": "trout",
|
| 105 |
+
"92": "tulip",
|
| 106 |
+
"93": "turtle",
|
| 107 |
+
"94": "wardrobe",
|
| 108 |
+
"95": "whale",
|
| 109 |
+
"96": "willow_tree",
|
| 110 |
+
"97": "wolf",
|
| 111 |
+
"98": "woman",
|
| 112 |
+
"99": "worm"
|
| 113 |
+
},
|
| 114 |
+
"image_size": 224,
|
| 115 |
+
"initializer_range": 0.02,
|
| 116 |
+
"intermediate_size": 3072,
|
| 117 |
+
"label2id": {
|
| 118 |
+
"apple": "0",
|
| 119 |
+
"aquarium_fish": "1",
|
| 120 |
+
"baby": "2",
|
| 121 |
+
"bear": "3",
|
| 122 |
+
"beaver": "4",
|
| 123 |
+
"bed": "5",
|
| 124 |
+
"bee": "6",
|
| 125 |
+
"beetle": "7",
|
| 126 |
+
"bicycle": "8",
|
| 127 |
+
"bottle": "9",
|
| 128 |
+
"bowl": "10",
|
| 129 |
+
"boy": "11",
|
| 130 |
+
"bridge": "12",
|
| 131 |
+
"bus": "13",
|
| 132 |
+
"butterfly": "14",
|
| 133 |
+
"camel": "15",
|
| 134 |
+
"can": "16",
|
| 135 |
+
"castle": "17",
|
| 136 |
+
"caterpillar": "18",
|
| 137 |
+
"cattle": "19",
|
| 138 |
+
"chair": "20",
|
| 139 |
+
"chimpanzee": "21",
|
| 140 |
+
"clock": "22",
|
| 141 |
+
"cloud": "23",
|
| 142 |
+
"cockroach": "24",
|
| 143 |
+
"couch": "25",
|
| 144 |
+
"cra": "26",
|
| 145 |
+
"crocodile": "27",
|
| 146 |
+
"cup": "28",
|
| 147 |
+
"dinosaur": "29",
|
| 148 |
+
"dolphin": "30",
|
| 149 |
+
"elephant": "31",
|
| 150 |
+
"flatfish": "32",
|
| 151 |
+
"forest": "33",
|
| 152 |
+
"fox": "34",
|
| 153 |
+
"girl": "35",
|
| 154 |
+
"hamster": "36",
|
| 155 |
+
"house": "37",
|
| 156 |
+
"kangaroo": "38",
|
| 157 |
+
"keyboard": "39",
|
| 158 |
+
"lamp": "40",
|
| 159 |
+
"lawn_mower": "41",
|
| 160 |
+
"leopard": "42",
|
| 161 |
+
"lion": "43",
|
| 162 |
+
"lizard": "44",
|
| 163 |
+
"lobster": "45",
|
| 164 |
+
"man": "46",
|
| 165 |
+
"maple_tree": "47",
|
| 166 |
+
"motorcycle": "48",
|
| 167 |
+
"mountain": "49",
|
| 168 |
+
"mouse": "50",
|
| 169 |
+
"mushroom": "51",
|
| 170 |
+
"oak_tree": "52",
|
| 171 |
+
"orange": "53",
|
| 172 |
+
"orchid": "54",
|
| 173 |
+
"otter": "55",
|
| 174 |
+
"palm_tree": "56",
|
| 175 |
+
"pear": "57",
|
| 176 |
+
"pickup_truck": "58",
|
| 177 |
+
"pine_tree": "59",
|
| 178 |
+
"plain": "60",
|
| 179 |
+
"plate": "61",
|
| 180 |
+
"poppy": "62",
|
| 181 |
+
"porcupine": "63",
|
| 182 |
+
"possum": "64",
|
| 183 |
+
"rabbit": "65",
|
| 184 |
+
"raccoon": "66",
|
| 185 |
+
"ray": "67",
|
| 186 |
+
"road": "68",
|
| 187 |
+
"rocket": "69",
|
| 188 |
+
"rose": "70",
|
| 189 |
+
"sea": "71",
|
| 190 |
+
"seal": "72",
|
| 191 |
+
"shark": "73",
|
| 192 |
+
"shrew": "74",
|
| 193 |
+
"skunk": "75",
|
| 194 |
+
"skyscraper": "76",
|
| 195 |
+
"snail": "77",
|
| 196 |
+
"snake": "78",
|
| 197 |
+
"spider": "79",
|
| 198 |
+
"squirrel": "80",
|
| 199 |
+
"streetcar": "81",
|
| 200 |
+
"sunflower": "82",
|
| 201 |
+
"sweet_pepper": "83",
|
| 202 |
+
"table": "84",
|
| 203 |
+
"tank": "85",
|
| 204 |
+
"telephone": "86",
|
| 205 |
+
"television": "87",
|
| 206 |
+
"tiger": "88",
|
| 207 |
+
"tractor": "89",
|
| 208 |
+
"train": "90",
|
| 209 |
+
"trout": "91",
|
| 210 |
+
"tulip": "92",
|
| 211 |
+
"turtle": "93",
|
| 212 |
+
"wardrobe": "94",
|
| 213 |
+
"whale": "95",
|
| 214 |
+
"willow_tree": "96",
|
| 215 |
+
"wolf": "97",
|
| 216 |
+
"woman": "98",
|
| 217 |
+
"worm": "99"
|
| 218 |
+
},
|
| 219 |
+
"layer_norm_eps": 1e-12,
|
| 220 |
+
"model_type": "deit",
|
| 221 |
+
"num_attention_heads": 12,
|
| 222 |
+
"num_channels": 3,
|
| 223 |
+
"num_hidden_layers": 12,
|
| 224 |
+
"patch_size": 16,
|
| 225 |
+
"qkv_bias": true,
|
| 226 |
+
"torch_dtype": "float32",
|
| 227 |
+
"transformers_version": "4.38.1"
|
| 228 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.906,
|
| 4 |
+
"eval_loss": 0.3053509294986725,
|
| 5 |
+
"eval_runtime": 173.1395,
|
| 6 |
+
"eval_samples_per_second": 57.757,
|
| 7 |
+
"eval_steps_per_second": 57.757
|
| 8 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9b71ff9ed44c774c8658235ed01d108cb4daef8da81aac1ace5bd1ce14ba37b0
|
| 3 |
+
size 343839688
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"crop_size": {
|
| 3 |
+
"height": 224,
|
| 4 |
+
"width": 224
|
| 5 |
+
},
|
| 6 |
+
"do_center_crop": true,
|
| 7 |
+
"do_normalize": true,
|
| 8 |
+
"do_rescale": true,
|
| 9 |
+
"do_resize": true,
|
| 10 |
+
"image_mean": [
|
| 11 |
+
0.485,
|
| 12 |
+
0.456,
|
| 13 |
+
0.406
|
| 14 |
+
],
|
| 15 |
+
"image_processor_type": "DeiTImageProcessor",
|
| 16 |
+
"image_std": [
|
| 17 |
+
0.229,
|
| 18 |
+
0.224,
|
| 19 |
+
0.225
|
| 20 |
+
],
|
| 21 |
+
"resample": 3,
|
| 22 |
+
"rescale_factor": 0.00392156862745098,
|
| 23 |
+
"size": {
|
| 24 |
+
"height": 256,
|
| 25 |
+
"shortest_edge": 224,
|
| 26 |
+
"width": 256
|
| 27 |
+
}
|
| 28 |
+
}
|
runs/Mar02_13-37-49_9a8dc51c99f8/events.out.tfevents.1709386701.9a8dc51c99f8.326.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7b0dbced4ab36e03e88e46ef91d72555e172f4d34044b31a7e7ecd7fa938a4a9
|
| 3 |
+
size 8685
|
runs/Mar02_13-46-44_6f1f307693d9/events.out.tfevents.1709387309.6f1f307693d9.258.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f642a8349b7c92b81885a73fe6c9bac0e8bafb6d67b0d1fe831cd80d77c82587
|
| 3 |
+
size 9115
|
runs/Mar02_13-56-30_966e26ef2d5b/events.out.tfevents.1709387957.966e26ef2d5b.289.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be78f5c8a81fff068bf10b4b76f5771ec5380914718819f77accaee79e005ba9
|
| 3 |
+
size 9115
|
runs/Mar02_13-56-30_966e26ef2d5b/events.out.tfevents.1709388178.966e26ef2d5b.289.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:483e318391ab1d971902cd0c3fd8be437a3416ce7325c17bd50199b8a53ef5d5
|
| 3 |
+
size 9515
|
runs/Mar02_14-41-50_04f752329dca/events.out.tfevents.1709390567.04f752329dca.283.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d9c96a2100754ead05455d791e7fa94ac8e9591244dbfc141238edcc76345906
|
| 3 |
+
size 9115
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"total_flos": 3.8815447523328e+19,
|
| 4 |
+
"train_loss": 0.0,
|
| 5 |
+
"train_runtime": 24.522,
|
| 6 |
+
"train_samples_per_second": 20389.834,
|
| 7 |
+
"train_steps_per_second": 318.897
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1056 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.3053509294986725,
|
| 3 |
+
"best_model_checkpoint": "/content/drive/MyDrive/Sumdu/PhD/LGViT/DeiT-base/cifar100/base/checkpoint-7820",
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 7820,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.06,
|
| 13 |
+
"learning_rate": 4.9680306905370846e-05,
|
| 14 |
+
"loss": 4.3042,
|
| 15 |
+
"step": 50
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.13,
|
| 19 |
+
"learning_rate": 4.936061381074169e-05,
|
| 20 |
+
"loss": 3.2558,
|
| 21 |
+
"step": 100
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.19,
|
| 25 |
+
"learning_rate": 4.9040920716112534e-05,
|
| 26 |
+
"loss": 2.3793,
|
| 27 |
+
"step": 150
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.26,
|
| 31 |
+
"learning_rate": 4.872122762148338e-05,
|
| 32 |
+
"loss": 1.898,
|
| 33 |
+
"step": 200
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.32,
|
| 37 |
+
"learning_rate": 4.840153452685422e-05,
|
| 38 |
+
"loss": 1.6523,
|
| 39 |
+
"step": 250
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.38,
|
| 43 |
+
"learning_rate": 4.8081841432225067e-05,
|
| 44 |
+
"loss": 1.5372,
|
| 45 |
+
"step": 300
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.45,
|
| 49 |
+
"learning_rate": 4.776214833759591e-05,
|
| 50 |
+
"loss": 1.3641,
|
| 51 |
+
"step": 350
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.51,
|
| 55 |
+
"learning_rate": 4.7442455242966755e-05,
|
| 56 |
+
"loss": 1.3355,
|
| 57 |
+
"step": 400
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.58,
|
| 61 |
+
"learning_rate": 4.71227621483376e-05,
|
| 62 |
+
"loss": 1.2267,
|
| 63 |
+
"step": 450
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.64,
|
| 67 |
+
"learning_rate": 4.680306905370844e-05,
|
| 68 |
+
"loss": 1.1976,
|
| 69 |
+
"step": 500
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.7,
|
| 73 |
+
"learning_rate": 4.6483375959079287e-05,
|
| 74 |
+
"loss": 1.1752,
|
| 75 |
+
"step": 550
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.77,
|
| 79 |
+
"learning_rate": 4.616368286445013e-05,
|
| 80 |
+
"loss": 1.178,
|
| 81 |
+
"step": 600
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.83,
|
| 85 |
+
"learning_rate": 4.5843989769820975e-05,
|
| 86 |
+
"loss": 1.1759,
|
| 87 |
+
"step": 650
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.9,
|
| 91 |
+
"learning_rate": 4.552429667519182e-05,
|
| 92 |
+
"loss": 1.1176,
|
| 93 |
+
"step": 700
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.96,
|
| 97 |
+
"learning_rate": 4.520460358056266e-05,
|
| 98 |
+
"loss": 1.1232,
|
| 99 |
+
"step": 750
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 1.0,
|
| 103 |
+
"eval_accuracy": 0.8416,
|
| 104 |
+
"eval_loss": 0.539007306098938,
|
| 105 |
+
"eval_runtime": 176.056,
|
| 106 |
+
"eval_samples_per_second": 56.8,
|
| 107 |
+
"eval_steps_per_second": 56.8,
|
| 108 |
+
"step": 782
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 1.02,
|
| 112 |
+
"learning_rate": 4.488491048593351e-05,
|
| 113 |
+
"loss": 1.0464,
|
| 114 |
+
"step": 800
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 1.09,
|
| 118 |
+
"learning_rate": 4.456521739130435e-05,
|
| 119 |
+
"loss": 0.9967,
|
| 120 |
+
"step": 850
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 1.15,
|
| 124 |
+
"learning_rate": 4.4245524296675195e-05,
|
| 125 |
+
"loss": 0.9914,
|
| 126 |
+
"step": 900
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 1.21,
|
| 130 |
+
"learning_rate": 4.392583120204604e-05,
|
| 131 |
+
"loss": 0.944,
|
| 132 |
+
"step": 950
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 1.28,
|
| 136 |
+
"learning_rate": 4.360613810741688e-05,
|
| 137 |
+
"loss": 0.9537,
|
| 138 |
+
"step": 1000
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 1.34,
|
| 142 |
+
"learning_rate": 4.328644501278773e-05,
|
| 143 |
+
"loss": 0.9067,
|
| 144 |
+
"step": 1050
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 1.41,
|
| 148 |
+
"learning_rate": 4.296675191815857e-05,
|
| 149 |
+
"loss": 0.9325,
|
| 150 |
+
"step": 1100
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 1.47,
|
| 154 |
+
"learning_rate": 4.2647058823529415e-05,
|
| 155 |
+
"loss": 0.9125,
|
| 156 |
+
"step": 1150
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 1.53,
|
| 160 |
+
"learning_rate": 4.232736572890026e-05,
|
| 161 |
+
"loss": 0.9322,
|
| 162 |
+
"step": 1200
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 1.6,
|
| 166 |
+
"learning_rate": 4.20076726342711e-05,
|
| 167 |
+
"loss": 0.9293,
|
| 168 |
+
"step": 1250
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.66,
|
| 172 |
+
"learning_rate": 4.168797953964195e-05,
|
| 173 |
+
"loss": 0.8524,
|
| 174 |
+
"step": 1300
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 1.73,
|
| 178 |
+
"learning_rate": 4.136828644501279e-05,
|
| 179 |
+
"loss": 0.942,
|
| 180 |
+
"step": 1350
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 1.79,
|
| 184 |
+
"learning_rate": 4.1048593350383635e-05,
|
| 185 |
+
"loss": 0.9373,
|
| 186 |
+
"step": 1400
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 1.85,
|
| 190 |
+
"learning_rate": 4.072890025575448e-05,
|
| 191 |
+
"loss": 0.8504,
|
| 192 |
+
"step": 1450
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 1.92,
|
| 196 |
+
"learning_rate": 4.040920716112532e-05,
|
| 197 |
+
"loss": 0.9145,
|
| 198 |
+
"step": 1500
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 1.98,
|
| 202 |
+
"learning_rate": 4.008951406649617e-05,
|
| 203 |
+
"loss": 0.9017,
|
| 204 |
+
"step": 1550
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 2.0,
|
| 208 |
+
"eval_accuracy": 0.8699,
|
| 209 |
+
"eval_loss": 0.4364722669124603,
|
| 210 |
+
"eval_runtime": 178.2596,
|
| 211 |
+
"eval_samples_per_second": 56.098,
|
| 212 |
+
"eval_steps_per_second": 56.098,
|
| 213 |
+
"step": 1564
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 2.05,
|
| 217 |
+
"learning_rate": 3.976982097186701e-05,
|
| 218 |
+
"loss": 0.8527,
|
| 219 |
+
"step": 1600
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 2.11,
|
| 223 |
+
"learning_rate": 3.9450127877237855e-05,
|
| 224 |
+
"loss": 0.7803,
|
| 225 |
+
"step": 1650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 2.17,
|
| 229 |
+
"learning_rate": 3.91304347826087e-05,
|
| 230 |
+
"loss": 0.799,
|
| 231 |
+
"step": 1700
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 2.24,
|
| 235 |
+
"learning_rate": 3.881074168797954e-05,
|
| 236 |
+
"loss": 0.7873,
|
| 237 |
+
"step": 1750
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 2.3,
|
| 241 |
+
"learning_rate": 3.849104859335039e-05,
|
| 242 |
+
"loss": 0.7999,
|
| 243 |
+
"step": 1800
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 2.37,
|
| 247 |
+
"learning_rate": 3.817135549872123e-05,
|
| 248 |
+
"loss": 0.7911,
|
| 249 |
+
"step": 1850
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 2.43,
|
| 253 |
+
"learning_rate": 3.7851662404092075e-05,
|
| 254 |
+
"loss": 0.7987,
|
| 255 |
+
"step": 1900
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 2.49,
|
| 259 |
+
"learning_rate": 3.753196930946292e-05,
|
| 260 |
+
"loss": 0.7923,
|
| 261 |
+
"step": 1950
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 2.56,
|
| 265 |
+
"learning_rate": 3.721227621483376e-05,
|
| 266 |
+
"loss": 0.8116,
|
| 267 |
+
"step": 2000
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 2.62,
|
| 271 |
+
"learning_rate": 3.689258312020461e-05,
|
| 272 |
+
"loss": 0.7709,
|
| 273 |
+
"step": 2050
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 2.69,
|
| 277 |
+
"learning_rate": 3.657289002557545e-05,
|
| 278 |
+
"loss": 0.764,
|
| 279 |
+
"step": 2100
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 2.75,
|
| 283 |
+
"learning_rate": 3.6253196930946295e-05,
|
| 284 |
+
"loss": 0.8047,
|
| 285 |
+
"step": 2150
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 2.81,
|
| 289 |
+
"learning_rate": 3.593350383631714e-05,
|
| 290 |
+
"loss": 0.7357,
|
| 291 |
+
"step": 2200
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 2.88,
|
| 295 |
+
"learning_rate": 3.561381074168798e-05,
|
| 296 |
+
"loss": 0.7779,
|
| 297 |
+
"step": 2250
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 2.94,
|
| 301 |
+
"learning_rate": 3.529411764705883e-05,
|
| 302 |
+
"loss": 0.7565,
|
| 303 |
+
"step": 2300
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 3.0,
|
| 307 |
+
"eval_accuracy": 0.8858,
|
| 308 |
+
"eval_loss": 0.3678203225135803,
|
| 309 |
+
"eval_runtime": 176.8651,
|
| 310 |
+
"eval_samples_per_second": 56.54,
|
| 311 |
+
"eval_steps_per_second": 56.54,
|
| 312 |
+
"step": 2346
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 3.01,
|
| 316 |
+
"learning_rate": 3.497442455242967e-05,
|
| 317 |
+
"loss": 0.7418,
|
| 318 |
+
"step": 2350
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 3.07,
|
| 322 |
+
"learning_rate": 3.4654731457800515e-05,
|
| 323 |
+
"loss": 0.709,
|
| 324 |
+
"step": 2400
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 3.13,
|
| 328 |
+
"learning_rate": 3.433503836317136e-05,
|
| 329 |
+
"loss": 0.6461,
|
| 330 |
+
"step": 2450
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 3.2,
|
| 334 |
+
"learning_rate": 3.40153452685422e-05,
|
| 335 |
+
"loss": 0.6895,
|
| 336 |
+
"step": 2500
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 3.26,
|
| 340 |
+
"learning_rate": 3.369565217391305e-05,
|
| 341 |
+
"loss": 0.6676,
|
| 342 |
+
"step": 2550
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 3.32,
|
| 346 |
+
"learning_rate": 3.337595907928389e-05,
|
| 347 |
+
"loss": 0.6769,
|
| 348 |
+
"step": 2600
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 3.39,
|
| 352 |
+
"learning_rate": 3.3056265984654735e-05,
|
| 353 |
+
"loss": 0.7018,
|
| 354 |
+
"step": 2650
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 3.45,
|
| 358 |
+
"learning_rate": 3.273657289002558e-05,
|
| 359 |
+
"loss": 0.6686,
|
| 360 |
+
"step": 2700
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 3.52,
|
| 364 |
+
"learning_rate": 3.241687979539642e-05,
|
| 365 |
+
"loss": 0.7211,
|
| 366 |
+
"step": 2750
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 3.58,
|
| 370 |
+
"learning_rate": 3.209718670076726e-05,
|
| 371 |
+
"loss": 0.7097,
|
| 372 |
+
"step": 2800
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 3.64,
|
| 376 |
+
"learning_rate": 3.177749360613811e-05,
|
| 377 |
+
"loss": 0.6939,
|
| 378 |
+
"step": 2850
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 3.71,
|
| 382 |
+
"learning_rate": 3.145780051150895e-05,
|
| 383 |
+
"loss": 0.6593,
|
| 384 |
+
"step": 2900
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 3.77,
|
| 388 |
+
"learning_rate": 3.11381074168798e-05,
|
| 389 |
+
"loss": 0.6931,
|
| 390 |
+
"step": 2950
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 3.84,
|
| 394 |
+
"learning_rate": 3.081841432225064e-05,
|
| 395 |
+
"loss": 0.711,
|
| 396 |
+
"step": 3000
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 3.9,
|
| 400 |
+
"learning_rate": 3.0498721227621484e-05,
|
| 401 |
+
"loss": 0.6439,
|
| 402 |
+
"step": 3050
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 3.96,
|
| 406 |
+
"learning_rate": 3.0179028132992325e-05,
|
| 407 |
+
"loss": 0.706,
|
| 408 |
+
"step": 3100
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 4.0,
|
| 412 |
+
"eval_accuracy": 0.8952,
|
| 413 |
+
"eval_loss": 0.3446212708950043,
|
| 414 |
+
"eval_runtime": 176.3518,
|
| 415 |
+
"eval_samples_per_second": 56.705,
|
| 416 |
+
"eval_steps_per_second": 56.705,
|
| 417 |
+
"step": 3128
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 4.03,
|
| 421 |
+
"learning_rate": 2.9859335038363172e-05,
|
| 422 |
+
"loss": 0.6984,
|
| 423 |
+
"step": 3150
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 4.09,
|
| 427 |
+
"learning_rate": 2.9539641943734013e-05,
|
| 428 |
+
"loss": 0.6062,
|
| 429 |
+
"step": 3200
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 4.16,
|
| 433 |
+
"learning_rate": 2.921994884910486e-05,
|
| 434 |
+
"loss": 0.6542,
|
| 435 |
+
"step": 3250
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 4.22,
|
| 439 |
+
"learning_rate": 2.89002557544757e-05,
|
| 440 |
+
"loss": 0.6092,
|
| 441 |
+
"step": 3300
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 4.28,
|
| 445 |
+
"learning_rate": 2.8580562659846548e-05,
|
| 446 |
+
"loss": 0.6325,
|
| 447 |
+
"step": 3350
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 4.35,
|
| 451 |
+
"learning_rate": 2.826086956521739e-05,
|
| 452 |
+
"loss": 0.6296,
|
| 453 |
+
"step": 3400
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 4.41,
|
| 457 |
+
"learning_rate": 2.7941176470588236e-05,
|
| 458 |
+
"loss": 0.6493,
|
| 459 |
+
"step": 3450
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 4.48,
|
| 463 |
+
"learning_rate": 2.7621483375959077e-05,
|
| 464 |
+
"loss": 0.609,
|
| 465 |
+
"step": 3500
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 4.54,
|
| 469 |
+
"learning_rate": 2.7301790281329924e-05,
|
| 470 |
+
"loss": 0.6173,
|
| 471 |
+
"step": 3550
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 4.6,
|
| 475 |
+
"learning_rate": 2.6982097186700765e-05,
|
| 476 |
+
"loss": 0.6263,
|
| 477 |
+
"step": 3600
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 4.67,
|
| 481 |
+
"learning_rate": 2.6662404092071612e-05,
|
| 482 |
+
"loss": 0.6264,
|
| 483 |
+
"step": 3650
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 4.73,
|
| 487 |
+
"learning_rate": 2.6342710997442456e-05,
|
| 488 |
+
"loss": 0.601,
|
| 489 |
+
"step": 3700
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 4.8,
|
| 493 |
+
"learning_rate": 2.60230179028133e-05,
|
| 494 |
+
"loss": 0.5905,
|
| 495 |
+
"step": 3750
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 4.86,
|
| 499 |
+
"learning_rate": 2.5703324808184144e-05,
|
| 500 |
+
"loss": 0.609,
|
| 501 |
+
"step": 3800
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 4.92,
|
| 505 |
+
"learning_rate": 2.538363171355499e-05,
|
| 506 |
+
"loss": 0.5939,
|
| 507 |
+
"step": 3850
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 4.99,
|
| 511 |
+
"learning_rate": 2.5063938618925832e-05,
|
| 512 |
+
"loss": 0.6353,
|
| 513 |
+
"step": 3900
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 5.0,
|
| 517 |
+
"eval_accuracy": 0.8986,
|
| 518 |
+
"eval_loss": 0.333119660615921,
|
| 519 |
+
"eval_runtime": 177.2116,
|
| 520 |
+
"eval_samples_per_second": 56.43,
|
| 521 |
+
"eval_steps_per_second": 56.43,
|
| 522 |
+
"step": 3910
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 5.05,
|
| 526 |
+
"learning_rate": 2.4744245524296676e-05,
|
| 527 |
+
"loss": 0.6056,
|
| 528 |
+
"step": 3950
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 5.12,
|
| 532 |
+
"learning_rate": 2.442455242966752e-05,
|
| 533 |
+
"loss": 0.5218,
|
| 534 |
+
"step": 4000
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 5.18,
|
| 538 |
+
"learning_rate": 2.4104859335038364e-05,
|
| 539 |
+
"loss": 0.5656,
|
| 540 |
+
"step": 4050
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 5.24,
|
| 544 |
+
"learning_rate": 2.378516624040921e-05,
|
| 545 |
+
"loss": 0.5728,
|
| 546 |
+
"step": 4100
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 5.31,
|
| 550 |
+
"learning_rate": 2.3465473145780052e-05,
|
| 551 |
+
"loss": 0.5803,
|
| 552 |
+
"step": 4150
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 5.37,
|
| 556 |
+
"learning_rate": 2.3145780051150897e-05,
|
| 557 |
+
"loss": 0.5907,
|
| 558 |
+
"step": 4200
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 5.43,
|
| 562 |
+
"learning_rate": 2.282608695652174e-05,
|
| 563 |
+
"loss": 0.567,
|
| 564 |
+
"step": 4250
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 5.5,
|
| 568 |
+
"learning_rate": 2.2506393861892585e-05,
|
| 569 |
+
"loss": 0.5585,
|
| 570 |
+
"step": 4300
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 5.56,
|
| 574 |
+
"learning_rate": 2.218670076726343e-05,
|
| 575 |
+
"loss": 0.5634,
|
| 576 |
+
"step": 4350
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 5.63,
|
| 580 |
+
"learning_rate": 2.1867007672634273e-05,
|
| 581 |
+
"loss": 0.5398,
|
| 582 |
+
"step": 4400
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 5.69,
|
| 586 |
+
"learning_rate": 2.1547314578005117e-05,
|
| 587 |
+
"loss": 0.5353,
|
| 588 |
+
"step": 4450
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 5.75,
|
| 592 |
+
"learning_rate": 2.122762148337596e-05,
|
| 593 |
+
"loss": 0.5341,
|
| 594 |
+
"step": 4500
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 5.82,
|
| 598 |
+
"learning_rate": 2.0907928388746805e-05,
|
| 599 |
+
"loss": 0.5758,
|
| 600 |
+
"step": 4550
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 5.88,
|
| 604 |
+
"learning_rate": 2.058823529411765e-05,
|
| 605 |
+
"loss": 0.5647,
|
| 606 |
+
"step": 4600
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 5.95,
|
| 610 |
+
"learning_rate": 2.0268542199488493e-05,
|
| 611 |
+
"loss": 0.5384,
|
| 612 |
+
"step": 4650
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 6.0,
|
| 616 |
+
"eval_accuracy": 0.9001,
|
| 617 |
+
"eval_loss": 0.3222615420818329,
|
| 618 |
+
"eval_runtime": 176.5584,
|
| 619 |
+
"eval_samples_per_second": 56.638,
|
| 620 |
+
"eval_steps_per_second": 56.638,
|
| 621 |
+
"step": 4692
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 6.01,
|
| 625 |
+
"learning_rate": 1.9948849104859337e-05,
|
| 626 |
+
"loss": 0.5534,
|
| 627 |
+
"step": 4700
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 6.07,
|
| 631 |
+
"learning_rate": 1.962915601023018e-05,
|
| 632 |
+
"loss": 0.5361,
|
| 633 |
+
"step": 4750
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 6.14,
|
| 637 |
+
"learning_rate": 1.9309462915601025e-05,
|
| 638 |
+
"loss": 0.5419,
|
| 639 |
+
"step": 4800
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 6.2,
|
| 643 |
+
"learning_rate": 1.898976982097187e-05,
|
| 644 |
+
"loss": 0.5235,
|
| 645 |
+
"step": 4850
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 6.27,
|
| 649 |
+
"learning_rate": 1.8670076726342713e-05,
|
| 650 |
+
"loss": 0.5153,
|
| 651 |
+
"step": 4900
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 6.33,
|
| 655 |
+
"learning_rate": 1.8350383631713557e-05,
|
| 656 |
+
"loss": 0.5362,
|
| 657 |
+
"step": 4950
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 6.39,
|
| 661 |
+
"learning_rate": 1.80306905370844e-05,
|
| 662 |
+
"loss": 0.4884,
|
| 663 |
+
"step": 5000
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 6.46,
|
| 667 |
+
"learning_rate": 1.7710997442455245e-05,
|
| 668 |
+
"loss": 0.532,
|
| 669 |
+
"step": 5050
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 6.52,
|
| 673 |
+
"learning_rate": 1.739130434782609e-05,
|
| 674 |
+
"loss": 0.4974,
|
| 675 |
+
"step": 5100
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 6.59,
|
| 679 |
+
"learning_rate": 1.7071611253196933e-05,
|
| 680 |
+
"loss": 0.5375,
|
| 681 |
+
"step": 5150
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 6.65,
|
| 685 |
+
"learning_rate": 1.6751918158567777e-05,
|
| 686 |
+
"loss": 0.5208,
|
| 687 |
+
"step": 5200
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 6.71,
|
| 691 |
+
"learning_rate": 1.643222506393862e-05,
|
| 692 |
+
"loss": 0.5572,
|
| 693 |
+
"step": 5250
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 6.78,
|
| 697 |
+
"learning_rate": 1.6112531969309465e-05,
|
| 698 |
+
"loss": 0.5224,
|
| 699 |
+
"step": 5300
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 6.84,
|
| 703 |
+
"learning_rate": 1.579283887468031e-05,
|
| 704 |
+
"loss": 0.5319,
|
| 705 |
+
"step": 5350
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 6.91,
|
| 709 |
+
"learning_rate": 1.5473145780051153e-05,
|
| 710 |
+
"loss": 0.5071,
|
| 711 |
+
"step": 5400
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 6.97,
|
| 715 |
+
"learning_rate": 1.5153452685421995e-05,
|
| 716 |
+
"loss": 0.5004,
|
| 717 |
+
"step": 5450
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 7.0,
|
| 721 |
+
"eval_accuracy": 0.9018,
|
| 722 |
+
"eval_loss": 0.32489562034606934,
|
| 723 |
+
"eval_runtime": 170.5926,
|
| 724 |
+
"eval_samples_per_second": 58.619,
|
| 725 |
+
"eval_steps_per_second": 58.619,
|
| 726 |
+
"step": 5474
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 7.03,
|
| 730 |
+
"learning_rate": 1.483375959079284e-05,
|
| 731 |
+
"loss": 0.5231,
|
| 732 |
+
"step": 5500
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 7.1,
|
| 736 |
+
"learning_rate": 1.4514066496163683e-05,
|
| 737 |
+
"loss": 0.5106,
|
| 738 |
+
"step": 5550
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 7.16,
|
| 742 |
+
"learning_rate": 1.4194373401534527e-05,
|
| 743 |
+
"loss": 0.4962,
|
| 744 |
+
"step": 5600
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 7.23,
|
| 748 |
+
"learning_rate": 1.3874680306905371e-05,
|
| 749 |
+
"loss": 0.4761,
|
| 750 |
+
"step": 5650
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 7.29,
|
| 754 |
+
"learning_rate": 1.3554987212276215e-05,
|
| 755 |
+
"loss": 0.5139,
|
| 756 |
+
"step": 5700
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 7.35,
|
| 760 |
+
"learning_rate": 1.323529411764706e-05,
|
| 761 |
+
"loss": 0.4848,
|
| 762 |
+
"step": 5750
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 7.42,
|
| 766 |
+
"learning_rate": 1.2915601023017903e-05,
|
| 767 |
+
"loss": 0.5141,
|
| 768 |
+
"step": 5800
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 7.48,
|
| 772 |
+
"learning_rate": 1.2595907928388747e-05,
|
| 773 |
+
"loss": 0.4613,
|
| 774 |
+
"step": 5850
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 7.54,
|
| 778 |
+
"learning_rate": 1.2276214833759591e-05,
|
| 779 |
+
"loss": 0.4606,
|
| 780 |
+
"step": 5900
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 7.61,
|
| 784 |
+
"learning_rate": 1.1956521739130435e-05,
|
| 785 |
+
"loss": 0.4778,
|
| 786 |
+
"step": 5950
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 7.67,
|
| 790 |
+
"learning_rate": 1.163682864450128e-05,
|
| 791 |
+
"loss": 0.497,
|
| 792 |
+
"step": 6000
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 7.74,
|
| 796 |
+
"learning_rate": 1.1317135549872123e-05,
|
| 797 |
+
"loss": 0.5005,
|
| 798 |
+
"step": 6050
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 7.8,
|
| 802 |
+
"learning_rate": 1.0997442455242967e-05,
|
| 803 |
+
"loss": 0.5184,
|
| 804 |
+
"step": 6100
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 7.86,
|
| 808 |
+
"learning_rate": 1.0677749360613812e-05,
|
| 809 |
+
"loss": 0.5045,
|
| 810 |
+
"step": 6150
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 7.93,
|
| 814 |
+
"learning_rate": 1.0358056265984656e-05,
|
| 815 |
+
"loss": 0.4755,
|
| 816 |
+
"step": 6200
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 7.99,
|
| 820 |
+
"learning_rate": 1.00383631713555e-05,
|
| 821 |
+
"loss": 0.4672,
|
| 822 |
+
"step": 6250
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 8.0,
|
| 826 |
+
"eval_accuracy": 0.904,
|
| 827 |
+
"eval_loss": 0.31126102805137634,
|
| 828 |
+
"eval_runtime": 170.2673,
|
| 829 |
+
"eval_samples_per_second": 58.731,
|
| 830 |
+
"eval_steps_per_second": 58.731,
|
| 831 |
+
"step": 6256
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 8.06,
|
| 835 |
+
"learning_rate": 9.718670076726342e-06,
|
| 836 |
+
"loss": 0.4827,
|
| 837 |
+
"step": 6300
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 8.12,
|
| 841 |
+
"learning_rate": 9.398976982097186e-06,
|
| 842 |
+
"loss": 0.4801,
|
| 843 |
+
"step": 6350
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 8.18,
|
| 847 |
+
"learning_rate": 9.07928388746803e-06,
|
| 848 |
+
"loss": 0.4566,
|
| 849 |
+
"step": 6400
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 8.25,
|
| 853 |
+
"learning_rate": 8.759590792838874e-06,
|
| 854 |
+
"loss": 0.4837,
|
| 855 |
+
"step": 6450
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 8.31,
|
| 859 |
+
"learning_rate": 8.439897698209718e-06,
|
| 860 |
+
"loss": 0.4406,
|
| 861 |
+
"step": 6500
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 8.38,
|
| 865 |
+
"learning_rate": 8.120204603580562e-06,
|
| 866 |
+
"loss": 0.4885,
|
| 867 |
+
"step": 6550
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 8.44,
|
| 871 |
+
"learning_rate": 7.800511508951406e-06,
|
| 872 |
+
"loss": 0.4286,
|
| 873 |
+
"step": 6600
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 8.5,
|
| 877 |
+
"learning_rate": 7.480818414322251e-06,
|
| 878 |
+
"loss": 0.4652,
|
| 879 |
+
"step": 6650
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 8.57,
|
| 883 |
+
"learning_rate": 7.161125319693095e-06,
|
| 884 |
+
"loss": 0.4914,
|
| 885 |
+
"step": 6700
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 8.63,
|
| 889 |
+
"learning_rate": 6.841432225063939e-06,
|
| 890 |
+
"loss": 0.4436,
|
| 891 |
+
"step": 6750
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 8.7,
|
| 895 |
+
"learning_rate": 6.521739130434783e-06,
|
| 896 |
+
"loss": 0.4366,
|
| 897 |
+
"step": 6800
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 8.76,
|
| 901 |
+
"learning_rate": 6.202046035805627e-06,
|
| 902 |
+
"loss": 0.4481,
|
| 903 |
+
"step": 6850
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 8.82,
|
| 907 |
+
"learning_rate": 5.882352941176471e-06,
|
| 908 |
+
"loss": 0.4397,
|
| 909 |
+
"step": 6900
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 8.89,
|
| 913 |
+
"learning_rate": 5.562659846547315e-06,
|
| 914 |
+
"loss": 0.4823,
|
| 915 |
+
"step": 6950
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 8.95,
|
| 919 |
+
"learning_rate": 5.242966751918159e-06,
|
| 920 |
+
"loss": 0.4526,
|
| 921 |
+
"step": 7000
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 9.0,
|
| 925 |
+
"eval_accuracy": 0.9054,
|
| 926 |
+
"eval_loss": 0.3081473708152771,
|
| 927 |
+
"eval_runtime": 171.7477,
|
| 928 |
+
"eval_samples_per_second": 58.225,
|
| 929 |
+
"eval_steps_per_second": 58.225,
|
| 930 |
+
"step": 7038
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 9.02,
|
| 934 |
+
"learning_rate": 4.923273657289003e-06,
|
| 935 |
+
"loss": 0.4519,
|
| 936 |
+
"step": 7050
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 9.08,
|
| 940 |
+
"learning_rate": 4.603580562659847e-06,
|
| 941 |
+
"loss": 0.4074,
|
| 942 |
+
"step": 7100
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 9.14,
|
| 946 |
+
"learning_rate": 4.283887468030691e-06,
|
| 947 |
+
"loss": 0.4026,
|
| 948 |
+
"step": 7150
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 9.21,
|
| 952 |
+
"learning_rate": 3.964194373401535e-06,
|
| 953 |
+
"loss": 0.4629,
|
| 954 |
+
"step": 7200
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 9.27,
|
| 958 |
+
"learning_rate": 3.644501278772379e-06,
|
| 959 |
+
"loss": 0.4267,
|
| 960 |
+
"step": 7250
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 9.34,
|
| 964 |
+
"learning_rate": 3.324808184143223e-06,
|
| 965 |
+
"loss": 0.4086,
|
| 966 |
+
"step": 7300
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 9.4,
|
| 970 |
+
"learning_rate": 3.0051150895140667e-06,
|
| 971 |
+
"loss": 0.4037,
|
| 972 |
+
"step": 7350
|
| 973 |
+
},
|
| 974 |
+
{
|
| 975 |
+
"epoch": 9.46,
|
| 976 |
+
"learning_rate": 2.6854219948849107e-06,
|
| 977 |
+
"loss": 0.4321,
|
| 978 |
+
"step": 7400
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 9.53,
|
| 982 |
+
"learning_rate": 2.3657289002557547e-06,
|
| 983 |
+
"loss": 0.4712,
|
| 984 |
+
"step": 7450
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 9.59,
|
| 988 |
+
"learning_rate": 2.0460358056265987e-06,
|
| 989 |
+
"loss": 0.4478,
|
| 990 |
+
"step": 7500
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 9.65,
|
| 994 |
+
"learning_rate": 1.7263427109974423e-06,
|
| 995 |
+
"loss": 0.4272,
|
| 996 |
+
"step": 7550
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 9.72,
|
| 1000 |
+
"learning_rate": 1.4066496163682865e-06,
|
| 1001 |
+
"loss": 0.4293,
|
| 1002 |
+
"step": 7600
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 9.78,
|
| 1006 |
+
"learning_rate": 1.0869565217391306e-06,
|
| 1007 |
+
"loss": 0.4139,
|
| 1008 |
+
"step": 7650
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 9.85,
|
| 1012 |
+
"learning_rate": 7.672634271099745e-07,
|
| 1013 |
+
"loss": 0.441,
|
| 1014 |
+
"step": 7700
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 9.91,
|
| 1018 |
+
"learning_rate": 4.4757033248081843e-07,
|
| 1019 |
+
"loss": 0.4479,
|
| 1020 |
+
"step": 7750
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 9.97,
|
| 1024 |
+
"learning_rate": 1.2787723785166242e-07,
|
| 1025 |
+
"loss": 0.4289,
|
| 1026 |
+
"step": 7800
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 10.0,
|
| 1030 |
+
"eval_accuracy": 0.906,
|
| 1031 |
+
"eval_loss": 0.3053509294986725,
|
| 1032 |
+
"eval_runtime": 174.6627,
|
| 1033 |
+
"eval_samples_per_second": 57.253,
|
| 1034 |
+
"eval_steps_per_second": 57.253,
|
| 1035 |
+
"step": 7820
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 10.0,
|
| 1039 |
+
"step": 7820,
|
| 1040 |
+
"total_flos": 3.8815447523328e+19,
|
| 1041 |
+
"train_loss": 0.0,
|
| 1042 |
+
"train_runtime": 24.522,
|
| 1043 |
+
"train_samples_per_second": 20389.834,
|
| 1044 |
+
"train_steps_per_second": 318.897
|
| 1045 |
+
}
|
| 1046 |
+
],
|
| 1047 |
+
"logging_steps": 50,
|
| 1048 |
+
"max_steps": 7820,
|
| 1049 |
+
"num_input_tokens_seen": 0,
|
| 1050 |
+
"num_train_epochs": 10,
|
| 1051 |
+
"save_steps": 500,
|
| 1052 |
+
"total_flos": 3.8815447523328e+19,
|
| 1053 |
+
"train_batch_size": 64,
|
| 1054 |
+
"trial_name": null,
|
| 1055 |
+
"trial_params": null
|
| 1056 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:055c36d00af3944358dbbaf22e3ec19450b0c7f9931658e230d9a7e39859bc36
|
| 3 |
+
size 5112
|