Upload folder using huggingface_hub
Browse files- checkpoint-3502/config.json +52 -0
- checkpoint-3502/model.safetensors +3 -0
- checkpoint-3502/optimizer.pt +3 -0
- checkpoint-3502/preprocessor_config.json +36 -0
- checkpoint-3502/rng_state.pth +3 -0
- checkpoint-3502/scheduler.pt +3 -0
- checkpoint-3502/trainer_state.json +223 -0
- checkpoint-3502/training_args.bin +3 -0
- config.json +52 -0
- model.safetensors +3 -0
- preprocessor_config.json +36 -0
- training_args.bin +3 -0
checkpoint-3502/config.json
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/vit-base-patch16-224-in21k",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"ViTForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"encoder_stride": 16,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.0,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "Crispy Chicken",
|
| 13 |
+
"1": "burger",
|
| 14 |
+
"2": "chole_bhature",
|
| 15 |
+
"3": "chai",
|
| 16 |
+
"4": "chicken_curry",
|
| 17 |
+
"5": "cheesecake",
|
| 18 |
+
"6": "Baked Potato",
|
| 19 |
+
"7": "apple_pie",
|
| 20 |
+
"8": "dal_makhani",
|
| 21 |
+
"9": "butter_naan",
|
| 22 |
+
"10": "dhokla",
|
| 23 |
+
"11": "chapati"
|
| 24 |
+
},
|
| 25 |
+
"image_size": 224,
|
| 26 |
+
"initializer_range": 0.02,
|
| 27 |
+
"intermediate_size": 3072,
|
| 28 |
+
"label2id": {
|
| 29 |
+
"Baked Potato": 6,
|
| 30 |
+
"Crispy Chicken": 0,
|
| 31 |
+
"apple_pie": 7,
|
| 32 |
+
"burger": 1,
|
| 33 |
+
"butter_naan": 9,
|
| 34 |
+
"chai": 3,
|
| 35 |
+
"chapati": 11,
|
| 36 |
+
"cheesecake": 5,
|
| 37 |
+
"chicken_curry": 4,
|
| 38 |
+
"chole_bhature": 2,
|
| 39 |
+
"dal_makhani": 8,
|
| 40 |
+
"dhokla": 10
|
| 41 |
+
},
|
| 42 |
+
"layer_norm_eps": 1e-12,
|
| 43 |
+
"model_type": "vit",
|
| 44 |
+
"num_attention_heads": 12,
|
| 45 |
+
"num_channels": 3,
|
| 46 |
+
"num_hidden_layers": 12,
|
| 47 |
+
"patch_size": 16,
|
| 48 |
+
"problem_type": "single_label_classification",
|
| 49 |
+
"qkv_bias": true,
|
| 50 |
+
"torch_dtype": "float32",
|
| 51 |
+
"transformers_version": "4.39.3"
|
| 52 |
+
}
|
checkpoint-3502/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:636ce365f3f04dee8240276b182eba4d9bfed8cb276b31b559170fb4ece366df
|
| 3 |
+
size 343254736
|
checkpoint-3502/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:25fd17ef0a265b073474f67809d37835fab09f29fcf1e4f42d38bf206dbe8b13
|
| 3 |
+
size 686630330
|
checkpoint-3502/preprocessor_config.json
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_valid_processor_keys": [
|
| 3 |
+
"images",
|
| 4 |
+
"do_resize",
|
| 5 |
+
"size",
|
| 6 |
+
"resample",
|
| 7 |
+
"do_rescale",
|
| 8 |
+
"rescale_factor",
|
| 9 |
+
"do_normalize",
|
| 10 |
+
"image_mean",
|
| 11 |
+
"image_std",
|
| 12 |
+
"return_tensors",
|
| 13 |
+
"data_format",
|
| 14 |
+
"input_data_format"
|
| 15 |
+
],
|
| 16 |
+
"do_normalize": true,
|
| 17 |
+
"do_rescale": true,
|
| 18 |
+
"do_resize": true,
|
| 19 |
+
"image_mean": [
|
| 20 |
+
0.5,
|
| 21 |
+
0.5,
|
| 22 |
+
0.5
|
| 23 |
+
],
|
| 24 |
+
"image_processor_type": "ViTImageProcessor",
|
| 25 |
+
"image_std": [
|
| 26 |
+
0.5,
|
| 27 |
+
0.5,
|
| 28 |
+
0.5
|
| 29 |
+
],
|
| 30 |
+
"resample": 2,
|
| 31 |
+
"rescale_factor": 0.00392156862745098,
|
| 32 |
+
"size": {
|
| 33 |
+
"height": 224,
|
| 34 |
+
"width": 224
|
| 35 |
+
}
|
| 36 |
+
}
|
checkpoint-3502/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0cd5f06eb3b9393c7e0ebebe1ad05cd6f524231a601ee409d53fdcafab6fbedf
|
| 3 |
+
size 14244
|
checkpoint-3502/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bba373b8ab47143520d807cc317cb7377c13b119312b117b94e4bcec8585033b
|
| 3 |
+
size 1064
|
checkpoint-3502/trainer_state.json
ADDED
|
@@ -0,0 +1,223 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.19758115708827972,
|
| 3 |
+
"best_model_checkpoint": "food_type_image_detection/checkpoint-3502",
|
| 4 |
+
"epoch": 17.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 3502,
|
| 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 |
+
"eval_accuracy": 0.8,
|
| 14 |
+
"eval_loss": 1.4625368118286133,
|
| 15 |
+
"eval_runtime": 62.0201,
|
| 16 |
+
"eval_samples_per_second": 26.524,
|
| 17 |
+
"eval_steps_per_second": 3.322,
|
| 18 |
+
"step": 206
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 2.0,
|
| 22 |
+
"eval_accuracy": 0.9288753799392098,
|
| 23 |
+
"eval_loss": 0.8831868767738342,
|
| 24 |
+
"eval_runtime": 61.6486,
|
| 25 |
+
"eval_samples_per_second": 26.683,
|
| 26 |
+
"eval_steps_per_second": 3.342,
|
| 27 |
+
"step": 412
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 2.43,
|
| 31 |
+
"grad_norm": 1.7900900840759277,
|
| 32 |
+
"learning_rate": 8.894348894348896e-06,
|
| 33 |
+
"loss": 1.4087,
|
| 34 |
+
"step": 500
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 3.0,
|
| 38 |
+
"eval_accuracy": 0.941033434650456,
|
| 39 |
+
"eval_loss": 0.6213602423667908,
|
| 40 |
+
"eval_runtime": 63.4524,
|
| 41 |
+
"eval_samples_per_second": 25.925,
|
| 42 |
+
"eval_steps_per_second": 3.247,
|
| 43 |
+
"step": 618
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 4.0,
|
| 47 |
+
"eval_accuracy": 0.9483282674772037,
|
| 48 |
+
"eval_loss": 0.4874071180820465,
|
| 49 |
+
"eval_runtime": 61.5456,
|
| 50 |
+
"eval_samples_per_second": 26.728,
|
| 51 |
+
"eval_steps_per_second": 3.347,
|
| 52 |
+
"step": 824
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 4.85,
|
| 56 |
+
"grad_norm": 1.8106521368026733,
|
| 57 |
+
"learning_rate": 7.665847665847667e-06,
|
| 58 |
+
"loss": 0.4866,
|
| 59 |
+
"step": 1000
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 5.0,
|
| 63 |
+
"eval_accuracy": 0.9513677811550152,
|
| 64 |
+
"eval_loss": 0.39853978157043457,
|
| 65 |
+
"eval_runtime": 61.6019,
|
| 66 |
+
"eval_samples_per_second": 26.704,
|
| 67 |
+
"eval_steps_per_second": 3.344,
|
| 68 |
+
"step": 1030
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 6.0,
|
| 72 |
+
"eval_accuracy": 0.9556231003039514,
|
| 73 |
+
"eval_loss": 0.34268224239349365,
|
| 74 |
+
"eval_runtime": 61.808,
|
| 75 |
+
"eval_samples_per_second": 26.615,
|
| 76 |
+
"eval_steps_per_second": 3.333,
|
| 77 |
+
"step": 1236
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 7.0,
|
| 81 |
+
"eval_accuracy": 0.9544072948328267,
|
| 82 |
+
"eval_loss": 0.3072179853916168,
|
| 83 |
+
"eval_runtime": 62.1104,
|
| 84 |
+
"eval_samples_per_second": 26.485,
|
| 85 |
+
"eval_steps_per_second": 3.317,
|
| 86 |
+
"step": 1442
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 7.28,
|
| 90 |
+
"grad_norm": 1.1063586473464966,
|
| 91 |
+
"learning_rate": 6.437346437346438e-06,
|
| 92 |
+
"loss": 0.2571,
|
| 93 |
+
"step": 1500
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 8.0,
|
| 97 |
+
"eval_accuracy": 0.9556231003039514,
|
| 98 |
+
"eval_loss": 0.28772830963134766,
|
| 99 |
+
"eval_runtime": 61.8927,
|
| 100 |
+
"eval_samples_per_second": 26.578,
|
| 101 |
+
"eval_steps_per_second": 3.328,
|
| 102 |
+
"step": 1648
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 9.0,
|
| 106 |
+
"eval_accuracy": 0.9592705167173252,
|
| 107 |
+
"eval_loss": 0.2588224411010742,
|
| 108 |
+
"eval_runtime": 61.5242,
|
| 109 |
+
"eval_samples_per_second": 26.737,
|
| 110 |
+
"eval_steps_per_second": 3.348,
|
| 111 |
+
"step": 1854
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 9.71,
|
| 115 |
+
"grad_norm": 0.31184545159339905,
|
| 116 |
+
"learning_rate": 5.208845208845209e-06,
|
| 117 |
+
"loss": 0.1647,
|
| 118 |
+
"step": 2000
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 10.0,
|
| 122 |
+
"eval_accuracy": 0.9604863221884499,
|
| 123 |
+
"eval_loss": 0.24028564989566803,
|
| 124 |
+
"eval_runtime": 61.688,
|
| 125 |
+
"eval_samples_per_second": 26.666,
|
| 126 |
+
"eval_steps_per_second": 3.339,
|
| 127 |
+
"step": 2060
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 11.0,
|
| 131 |
+
"eval_accuracy": 0.9495440729483283,
|
| 132 |
+
"eval_loss": 0.25100070238113403,
|
| 133 |
+
"eval_runtime": 62.0182,
|
| 134 |
+
"eval_samples_per_second": 26.524,
|
| 135 |
+
"eval_steps_per_second": 3.322,
|
| 136 |
+
"step": 2266
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 12.0,
|
| 140 |
+
"eval_accuracy": 0.9610942249240122,
|
| 141 |
+
"eval_loss": 0.22012543678283691,
|
| 142 |
+
"eval_runtime": 61.7287,
|
| 143 |
+
"eval_samples_per_second": 26.649,
|
| 144 |
+
"eval_steps_per_second": 3.337,
|
| 145 |
+
"step": 2472
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 12.14,
|
| 149 |
+
"grad_norm": 0.26199012994766235,
|
| 150 |
+
"learning_rate": 3.980343980343981e-06,
|
| 151 |
+
"loss": 0.1153,
|
| 152 |
+
"step": 2500
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 13.0,
|
| 156 |
+
"eval_accuracy": 0.9586626139817629,
|
| 157 |
+
"eval_loss": 0.21760769188404083,
|
| 158 |
+
"eval_runtime": 61.6059,
|
| 159 |
+
"eval_samples_per_second": 26.702,
|
| 160 |
+
"eval_steps_per_second": 3.344,
|
| 161 |
+
"step": 2678
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 14.0,
|
| 165 |
+
"eval_accuracy": 0.964741641337386,
|
| 166 |
+
"eval_loss": 0.20447726547718048,
|
| 167 |
+
"eval_runtime": 62.4469,
|
| 168 |
+
"eval_samples_per_second": 26.342,
|
| 169 |
+
"eval_steps_per_second": 3.299,
|
| 170 |
+
"step": 2884
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 14.56,
|
| 174 |
+
"grad_norm": 0.23101642727851868,
|
| 175 |
+
"learning_rate": 2.751842751842752e-06,
|
| 176 |
+
"loss": 0.091,
|
| 177 |
+
"step": 3000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 15.0,
|
| 181 |
+
"eval_accuracy": 0.9544072948328267,
|
| 182 |
+
"eval_loss": 0.2252843677997589,
|
| 183 |
+
"eval_runtime": 61.6312,
|
| 184 |
+
"eval_samples_per_second": 26.691,
|
| 185 |
+
"eval_steps_per_second": 3.342,
|
| 186 |
+
"step": 3090
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 16.0,
|
| 190 |
+
"eval_accuracy": 0.9617021276595744,
|
| 191 |
+
"eval_loss": 0.202474907040596,
|
| 192 |
+
"eval_runtime": 62.1996,
|
| 193 |
+
"eval_samples_per_second": 26.447,
|
| 194 |
+
"eval_steps_per_second": 3.312,
|
| 195 |
+
"step": 3296
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 16.99,
|
| 199 |
+
"grad_norm": 0.17727512121200562,
|
| 200 |
+
"learning_rate": 1.5233415233415235e-06,
|
| 201 |
+
"loss": 0.0762,
|
| 202 |
+
"step": 3500
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 17.0,
|
| 206 |
+
"eval_accuracy": 0.9623100303951367,
|
| 207 |
+
"eval_loss": 0.19758115708827972,
|
| 208 |
+
"eval_runtime": 62.4322,
|
| 209 |
+
"eval_samples_per_second": 26.349,
|
| 210 |
+
"eval_steps_per_second": 3.3,
|
| 211 |
+
"step": 3502
|
| 212 |
+
}
|
| 213 |
+
],
|
| 214 |
+
"logging_steps": 500,
|
| 215 |
+
"max_steps": 4120,
|
| 216 |
+
"num_input_tokens_seen": 0,
|
| 217 |
+
"num_train_epochs": 20,
|
| 218 |
+
"save_steps": 500,
|
| 219 |
+
"total_flos": 8.66376093889423e+18,
|
| 220 |
+
"train_batch_size": 32,
|
| 221 |
+
"trial_name": null,
|
| 222 |
+
"trial_params": null
|
| 223 |
+
}
|
checkpoint-3502/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6650882927ec3600b36a44f3745d63dc84e8699f59d5ce1f6c4aead7d5590226
|
| 3 |
+
size 4920
|
config.json
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/vit-base-patch16-224-in21k",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"ViTForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"encoder_stride": 16,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.0,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "Crispy Chicken",
|
| 13 |
+
"1": "burger",
|
| 14 |
+
"2": "chole_bhature",
|
| 15 |
+
"3": "chai",
|
| 16 |
+
"4": "chicken_curry",
|
| 17 |
+
"5": "cheesecake",
|
| 18 |
+
"6": "Baked Potato",
|
| 19 |
+
"7": "apple_pie",
|
| 20 |
+
"8": "dal_makhani",
|
| 21 |
+
"9": "butter_naan",
|
| 22 |
+
"10": "dhokla",
|
| 23 |
+
"11": "chapati"
|
| 24 |
+
},
|
| 25 |
+
"image_size": 224,
|
| 26 |
+
"initializer_range": 0.02,
|
| 27 |
+
"intermediate_size": 3072,
|
| 28 |
+
"label2id": {
|
| 29 |
+
"Baked Potato": 6,
|
| 30 |
+
"Crispy Chicken": 0,
|
| 31 |
+
"apple_pie": 7,
|
| 32 |
+
"burger": 1,
|
| 33 |
+
"butter_naan": 9,
|
| 34 |
+
"chai": 3,
|
| 35 |
+
"chapati": 11,
|
| 36 |
+
"cheesecake": 5,
|
| 37 |
+
"chicken_curry": 4,
|
| 38 |
+
"chole_bhature": 2,
|
| 39 |
+
"dal_makhani": 8,
|
| 40 |
+
"dhokla": 10
|
| 41 |
+
},
|
| 42 |
+
"layer_norm_eps": 1e-12,
|
| 43 |
+
"model_type": "vit",
|
| 44 |
+
"num_attention_heads": 12,
|
| 45 |
+
"num_channels": 3,
|
| 46 |
+
"num_hidden_layers": 12,
|
| 47 |
+
"patch_size": 16,
|
| 48 |
+
"problem_type": "single_label_classification",
|
| 49 |
+
"qkv_bias": true,
|
| 50 |
+
"torch_dtype": "float32",
|
| 51 |
+
"transformers_version": "4.39.3"
|
| 52 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:636ce365f3f04dee8240276b182eba4d9bfed8cb276b31b559170fb4ece366df
|
| 3 |
+
size 343254736
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_valid_processor_keys": [
|
| 3 |
+
"images",
|
| 4 |
+
"do_resize",
|
| 5 |
+
"size",
|
| 6 |
+
"resample",
|
| 7 |
+
"do_rescale",
|
| 8 |
+
"rescale_factor",
|
| 9 |
+
"do_normalize",
|
| 10 |
+
"image_mean",
|
| 11 |
+
"image_std",
|
| 12 |
+
"return_tensors",
|
| 13 |
+
"data_format",
|
| 14 |
+
"input_data_format"
|
| 15 |
+
],
|
| 16 |
+
"do_normalize": true,
|
| 17 |
+
"do_rescale": true,
|
| 18 |
+
"do_resize": true,
|
| 19 |
+
"image_mean": [
|
| 20 |
+
0.5,
|
| 21 |
+
0.5,
|
| 22 |
+
0.5
|
| 23 |
+
],
|
| 24 |
+
"image_processor_type": "ViTImageProcessor",
|
| 25 |
+
"image_std": [
|
| 26 |
+
0.5,
|
| 27 |
+
0.5,
|
| 28 |
+
0.5
|
| 29 |
+
],
|
| 30 |
+
"resample": 2,
|
| 31 |
+
"rescale_factor": 0.00392156862745098,
|
| 32 |
+
"size": {
|
| 33 |
+
"height": 224,
|
| 34 |
+
"width": 224
|
| 35 |
+
}
|
| 36 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6650882927ec3600b36a44f3745d63dc84e8699f59d5ce1f6c4aead7d5590226
|
| 3 |
+
size 4920
|