Upload folder using huggingface_hub
Browse files- checkpoint-1140/config.json +188 -0
- checkpoint-1140/model.safetensors +3 -0
- checkpoint-1140/optimizer.pt +3 -0
- checkpoint-1140/preprocessor_config.json +36 -0
- checkpoint-1140/rng_state.pth +3 -0
- checkpoint-1140/scheduler.pt +3 -0
- checkpoint-1140/trainer_state.json +317 -0
- checkpoint-1140/training_args.bin +3 -0
- model.safetensors +1 -1
- training_args.bin +1 -1
checkpoint-1140/config.json
ADDED
|
@@ -0,0 +1,188 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "dima806/indian_food_image_detection",
|
| 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": "adhirasam",
|
| 13 |
+
"1": "aloo_gobi",
|
| 14 |
+
"2": "aloo_matar",
|
| 15 |
+
"3": "aloo_methi",
|
| 16 |
+
"4": "aloo_shimla_mirch",
|
| 17 |
+
"5": "aloo_tikki",
|
| 18 |
+
"6": "anarsa",
|
| 19 |
+
"7": "ariselu",
|
| 20 |
+
"8": "bandar_laddu",
|
| 21 |
+
"9": "basundi",
|
| 22 |
+
"10": "bhatura",
|
| 23 |
+
"11": "bhindi_masala",
|
| 24 |
+
"12": "biryani",
|
| 25 |
+
"13": "boondi",
|
| 26 |
+
"14": "butter_chicken",
|
| 27 |
+
"15": "chak_hao_kheer",
|
| 28 |
+
"16": "cham_cham",
|
| 29 |
+
"17": "chana_masala",
|
| 30 |
+
"18": "chapati",
|
| 31 |
+
"19": "chhena_kheeri",
|
| 32 |
+
"20": "chicken_razala",
|
| 33 |
+
"21": "chicken_tikka",
|
| 34 |
+
"22": "chicken_tikka_masala",
|
| 35 |
+
"23": "chikki",
|
| 36 |
+
"24": "daal_baati_churma",
|
| 37 |
+
"25": "daal_puri",
|
| 38 |
+
"26": "dal_makhani",
|
| 39 |
+
"27": "dal_tadka",
|
| 40 |
+
"28": "dharwad_pedha",
|
| 41 |
+
"29": "doodhpak",
|
| 42 |
+
"30": "double_ka_meetha",
|
| 43 |
+
"31": "dum_aloo",
|
| 44 |
+
"32": "gajar_ka_halwa",
|
| 45 |
+
"33": "gavvalu",
|
| 46 |
+
"34": "ghevar",
|
| 47 |
+
"35": "gulab_jamun",
|
| 48 |
+
"36": "imarti",
|
| 49 |
+
"37": "jalebi",
|
| 50 |
+
"38": "kachori",
|
| 51 |
+
"39": "kadai_paneer",
|
| 52 |
+
"40": "kadhi_pakoda",
|
| 53 |
+
"41": "kajjikaya",
|
| 54 |
+
"42": "kakinada_khaja",
|
| 55 |
+
"43": "kalakand",
|
| 56 |
+
"44": "karela_bharta",
|
| 57 |
+
"45": "kofta",
|
| 58 |
+
"46": "kuzhi_paniyaram",
|
| 59 |
+
"47": "lassi",
|
| 60 |
+
"48": "ledikeni",
|
| 61 |
+
"49": "litti_chokha",
|
| 62 |
+
"50": "lyangcha",
|
| 63 |
+
"51": "maach_jhol",
|
| 64 |
+
"52": "makki_di_roti_sarson_da_saag",
|
| 65 |
+
"53": "malapua",
|
| 66 |
+
"54": "misi_roti",
|
| 67 |
+
"55": "misti_doi",
|
| 68 |
+
"56": "modak",
|
| 69 |
+
"57": "mysore_pak",
|
| 70 |
+
"58": "naan",
|
| 71 |
+
"59": "navrattan_korma",
|
| 72 |
+
"60": "palak_paneer",
|
| 73 |
+
"61": "paneer_butter_masala",
|
| 74 |
+
"62": "phirni",
|
| 75 |
+
"63": "pithe",
|
| 76 |
+
"64": "poha",
|
| 77 |
+
"65": "poornalu",
|
| 78 |
+
"66": "pootharekulu",
|
| 79 |
+
"67": "qubani_ka_meetha",
|
| 80 |
+
"68": "rabri",
|
| 81 |
+
"69": "ras_malai",
|
| 82 |
+
"70": "rasgulla",
|
| 83 |
+
"71": "sandesh",
|
| 84 |
+
"72": "shankarpali",
|
| 85 |
+
"73": "sheer_korma",
|
| 86 |
+
"74": "sheera",
|
| 87 |
+
"75": "shrikhand",
|
| 88 |
+
"76": "sohan_halwa",
|
| 89 |
+
"77": "sohan_papdi",
|
| 90 |
+
"78": "sutar_feni",
|
| 91 |
+
"79": "unni_appam"
|
| 92 |
+
},
|
| 93 |
+
"image_size": 224,
|
| 94 |
+
"initializer_range": 0.02,
|
| 95 |
+
"intermediate_size": 3072,
|
| 96 |
+
"label2id": {
|
| 97 |
+
"adhirasam": 0,
|
| 98 |
+
"aloo_gobi": 1,
|
| 99 |
+
"aloo_matar": 2,
|
| 100 |
+
"aloo_methi": 3,
|
| 101 |
+
"aloo_shimla_mirch": 4,
|
| 102 |
+
"aloo_tikki": 5,
|
| 103 |
+
"anarsa": 6,
|
| 104 |
+
"ariselu": 7,
|
| 105 |
+
"bandar_laddu": 8,
|
| 106 |
+
"basundi": 9,
|
| 107 |
+
"bhatura": 10,
|
| 108 |
+
"bhindi_masala": 11,
|
| 109 |
+
"biryani": 12,
|
| 110 |
+
"boondi": 13,
|
| 111 |
+
"butter_chicken": 14,
|
| 112 |
+
"chak_hao_kheer": 15,
|
| 113 |
+
"cham_cham": 16,
|
| 114 |
+
"chana_masala": 17,
|
| 115 |
+
"chapati": 18,
|
| 116 |
+
"chhena_kheeri": 19,
|
| 117 |
+
"chicken_razala": 20,
|
| 118 |
+
"chicken_tikka": 21,
|
| 119 |
+
"chicken_tikka_masala": 22,
|
| 120 |
+
"chikki": 23,
|
| 121 |
+
"daal_baati_churma": 24,
|
| 122 |
+
"daal_puri": 25,
|
| 123 |
+
"dal_makhani": 26,
|
| 124 |
+
"dal_tadka": 27,
|
| 125 |
+
"dharwad_pedha": 28,
|
| 126 |
+
"doodhpak": 29,
|
| 127 |
+
"double_ka_meetha": 30,
|
| 128 |
+
"dum_aloo": 31,
|
| 129 |
+
"gajar_ka_halwa": 32,
|
| 130 |
+
"gavvalu": 33,
|
| 131 |
+
"ghevar": 34,
|
| 132 |
+
"gulab_jamun": 35,
|
| 133 |
+
"imarti": 36,
|
| 134 |
+
"jalebi": 37,
|
| 135 |
+
"kachori": 38,
|
| 136 |
+
"kadai_paneer": 39,
|
| 137 |
+
"kadhi_pakoda": 40,
|
| 138 |
+
"kajjikaya": 41,
|
| 139 |
+
"kakinada_khaja": 42,
|
| 140 |
+
"kalakand": 43,
|
| 141 |
+
"karela_bharta": 44,
|
| 142 |
+
"kofta": 45,
|
| 143 |
+
"kuzhi_paniyaram": 46,
|
| 144 |
+
"lassi": 47,
|
| 145 |
+
"ledikeni": 48,
|
| 146 |
+
"litti_chokha": 49,
|
| 147 |
+
"lyangcha": 50,
|
| 148 |
+
"maach_jhol": 51,
|
| 149 |
+
"makki_di_roti_sarson_da_saag": 52,
|
| 150 |
+
"malapua": 53,
|
| 151 |
+
"misi_roti": 54,
|
| 152 |
+
"misti_doi": 55,
|
| 153 |
+
"modak": 56,
|
| 154 |
+
"mysore_pak": 57,
|
| 155 |
+
"naan": 58,
|
| 156 |
+
"navrattan_korma": 59,
|
| 157 |
+
"palak_paneer": 60,
|
| 158 |
+
"paneer_butter_masala": 61,
|
| 159 |
+
"phirni": 62,
|
| 160 |
+
"pithe": 63,
|
| 161 |
+
"poha": 64,
|
| 162 |
+
"poornalu": 65,
|
| 163 |
+
"pootharekulu": 66,
|
| 164 |
+
"qubani_ka_meetha": 67,
|
| 165 |
+
"rabri": 68,
|
| 166 |
+
"ras_malai": 69,
|
| 167 |
+
"rasgulla": 70,
|
| 168 |
+
"sandesh": 71,
|
| 169 |
+
"shankarpali": 72,
|
| 170 |
+
"sheer_korma": 73,
|
| 171 |
+
"sheera": 74,
|
| 172 |
+
"shrikhand": 75,
|
| 173 |
+
"sohan_halwa": 76,
|
| 174 |
+
"sohan_papdi": 77,
|
| 175 |
+
"sutar_feni": 78,
|
| 176 |
+
"unni_appam": 79
|
| 177 |
+
},
|
| 178 |
+
"layer_norm_eps": 1e-12,
|
| 179 |
+
"model_type": "vit",
|
| 180 |
+
"num_attention_heads": 12,
|
| 181 |
+
"num_channels": 3,
|
| 182 |
+
"num_hidden_layers": 12,
|
| 183 |
+
"patch_size": 16,
|
| 184 |
+
"problem_type": "single_label_classification",
|
| 185 |
+
"qkv_bias": true,
|
| 186 |
+
"torch_dtype": "float32",
|
| 187 |
+
"transformers_version": "4.41.0"
|
| 188 |
+
}
|
checkpoint-1140/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ed6113566cd92744fa8be6bfa6d67a9422ba8440eeedcf18c2bed7cd7d77108a
|
| 3 |
+
size 343463912
|
checkpoint-1140/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:24075e850a993d8f837b74c31d389f95700ba98503d6fe95c4f817516e1514db
|
| 3 |
+
size 687048325
|
checkpoint-1140/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-1140/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f2eb24122a5e9d619c3f35025da2c7039b089d75bd715cb9170006fca673d86c
|
| 3 |
+
size 14575
|
checkpoint-1140/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5626ac0d707892a1d437e95ee8f2e92e59e84e3f45fc27a226ea380db45d4568
|
| 3 |
+
size 627
|
checkpoint-1140/trainer_state.json
ADDED
|
@@ -0,0 +1,317 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 3.9748706817626953,
|
| 3 |
+
"best_model_checkpoint": "indian_food_image_detection/checkpoint-1140",
|
| 4 |
+
"epoch": 30.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1140,
|
| 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": 0.3025,
|
| 14 |
+
"eval_loss": 4.151414394378662,
|
| 15 |
+
"eval_runtime": 24.372,
|
| 16 |
+
"eval_samples_per_second": 65.649,
|
| 17 |
+
"eval_steps_per_second": 2.052,
|
| 18 |
+
"step": 38
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 2.0,
|
| 22 |
+
"eval_accuracy": 0.313125,
|
| 23 |
+
"eval_loss": 4.140285015106201,
|
| 24 |
+
"eval_runtime": 24.4394,
|
| 25 |
+
"eval_samples_per_second": 65.468,
|
| 26 |
+
"eval_steps_per_second": 2.046,
|
| 27 |
+
"step": 76
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 3.0,
|
| 31 |
+
"eval_accuracy": 0.321875,
|
| 32 |
+
"eval_loss": 4.129441261291504,
|
| 33 |
+
"eval_runtime": 24.5066,
|
| 34 |
+
"eval_samples_per_second": 65.288,
|
| 35 |
+
"eval_steps_per_second": 2.04,
|
| 36 |
+
"step": 114
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 4.0,
|
| 40 |
+
"eval_accuracy": 0.331875,
|
| 41 |
+
"eval_loss": 4.1186909675598145,
|
| 42 |
+
"eval_runtime": 24.4129,
|
| 43 |
+
"eval_samples_per_second": 65.539,
|
| 44 |
+
"eval_steps_per_second": 2.048,
|
| 45 |
+
"step": 152
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 5.0,
|
| 49 |
+
"eval_accuracy": 0.3425,
|
| 50 |
+
"eval_loss": 4.108262538909912,
|
| 51 |
+
"eval_runtime": 24.4749,
|
| 52 |
+
"eval_samples_per_second": 65.373,
|
| 53 |
+
"eval_steps_per_second": 2.043,
|
| 54 |
+
"step": 190
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 6.0,
|
| 58 |
+
"eval_accuracy": 0.35,
|
| 59 |
+
"eval_loss": 4.098318099975586,
|
| 60 |
+
"eval_runtime": 24.5662,
|
| 61 |
+
"eval_samples_per_second": 65.13,
|
| 62 |
+
"eval_steps_per_second": 2.035,
|
| 63 |
+
"step": 228
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 7.0,
|
| 67 |
+
"eval_accuracy": 0.361875,
|
| 68 |
+
"eval_loss": 4.0884175300598145,
|
| 69 |
+
"eval_runtime": 24.4798,
|
| 70 |
+
"eval_samples_per_second": 65.36,
|
| 71 |
+
"eval_steps_per_second": 2.043,
|
| 72 |
+
"step": 266
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 8.0,
|
| 76 |
+
"eval_accuracy": 0.3725,
|
| 77 |
+
"eval_loss": 4.079204082489014,
|
| 78 |
+
"eval_runtime": 24.5132,
|
| 79 |
+
"eval_samples_per_second": 65.271,
|
| 80 |
+
"eval_steps_per_second": 2.04,
|
| 81 |
+
"step": 304
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 9.0,
|
| 85 |
+
"eval_accuracy": 0.38125,
|
| 86 |
+
"eval_loss": 4.069871425628662,
|
| 87 |
+
"eval_runtime": 24.3786,
|
| 88 |
+
"eval_samples_per_second": 65.631,
|
| 89 |
+
"eval_steps_per_second": 2.051,
|
| 90 |
+
"step": 342
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 10.0,
|
| 94 |
+
"eval_accuracy": 0.38625,
|
| 95 |
+
"eval_loss": 4.061194896697998,
|
| 96 |
+
"eval_runtime": 24.4377,
|
| 97 |
+
"eval_samples_per_second": 65.473,
|
| 98 |
+
"eval_steps_per_second": 2.046,
|
| 99 |
+
"step": 380
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 11.0,
|
| 103 |
+
"eval_accuracy": 0.398125,
|
| 104 |
+
"eval_loss": 4.05281400680542,
|
| 105 |
+
"eval_runtime": 24.8276,
|
| 106 |
+
"eval_samples_per_second": 64.444,
|
| 107 |
+
"eval_steps_per_second": 2.014,
|
| 108 |
+
"step": 418
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 12.0,
|
| 112 |
+
"eval_accuracy": 0.406875,
|
| 113 |
+
"eval_loss": 4.045093536376953,
|
| 114 |
+
"eval_runtime": 24.4234,
|
| 115 |
+
"eval_samples_per_second": 65.511,
|
| 116 |
+
"eval_steps_per_second": 2.047,
|
| 117 |
+
"step": 456
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 13.0,
|
| 121 |
+
"eval_accuracy": 0.4175,
|
| 122 |
+
"eval_loss": 4.037683486938477,
|
| 123 |
+
"eval_runtime": 24.4923,
|
| 124 |
+
"eval_samples_per_second": 65.327,
|
| 125 |
+
"eval_steps_per_second": 2.041,
|
| 126 |
+
"step": 494
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 13.157894736842104,
|
| 130 |
+
"grad_norm": 1.2188057899475098,
|
| 131 |
+
"learning_rate": 1.1743119266055047e-06,
|
| 132 |
+
"loss": 4.0651,
|
| 133 |
+
"step": 500
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 14.0,
|
| 137 |
+
"eval_accuracy": 0.42125,
|
| 138 |
+
"eval_loss": 4.0303263664245605,
|
| 139 |
+
"eval_runtime": 24.5981,
|
| 140 |
+
"eval_samples_per_second": 65.046,
|
| 141 |
+
"eval_steps_per_second": 2.033,
|
| 142 |
+
"step": 532
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 15.0,
|
| 146 |
+
"eval_accuracy": 0.425625,
|
| 147 |
+
"eval_loss": 4.023651599884033,
|
| 148 |
+
"eval_runtime": 24.5314,
|
| 149 |
+
"eval_samples_per_second": 65.223,
|
| 150 |
+
"eval_steps_per_second": 2.038,
|
| 151 |
+
"step": 570
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 16.0,
|
| 155 |
+
"eval_accuracy": 0.434375,
|
| 156 |
+
"eval_loss": 4.017561435699463,
|
| 157 |
+
"eval_runtime": 24.6545,
|
| 158 |
+
"eval_samples_per_second": 64.897,
|
| 159 |
+
"eval_steps_per_second": 2.028,
|
| 160 |
+
"step": 608
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 17.0,
|
| 164 |
+
"eval_accuracy": 0.439375,
|
| 165 |
+
"eval_loss": 4.011629104614258,
|
| 166 |
+
"eval_runtime": 24.5053,
|
| 167 |
+
"eval_samples_per_second": 65.292,
|
| 168 |
+
"eval_steps_per_second": 2.04,
|
| 169 |
+
"step": 646
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 18.0,
|
| 173 |
+
"eval_accuracy": 0.446875,
|
| 174 |
+
"eval_loss": 4.006289482116699,
|
| 175 |
+
"eval_runtime": 24.5387,
|
| 176 |
+
"eval_samples_per_second": 65.203,
|
| 177 |
+
"eval_steps_per_second": 2.038,
|
| 178 |
+
"step": 684
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 19.0,
|
| 182 |
+
"eval_accuracy": 0.45375,
|
| 183 |
+
"eval_loss": 4.00118350982666,
|
| 184 |
+
"eval_runtime": 24.6877,
|
| 185 |
+
"eval_samples_per_second": 64.81,
|
| 186 |
+
"eval_steps_per_second": 2.025,
|
| 187 |
+
"step": 722
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 20.0,
|
| 191 |
+
"eval_accuracy": 0.4575,
|
| 192 |
+
"eval_loss": 3.9966068267822266,
|
| 193 |
+
"eval_runtime": 24.6967,
|
| 194 |
+
"eval_samples_per_second": 64.786,
|
| 195 |
+
"eval_steps_per_second": 2.025,
|
| 196 |
+
"step": 760
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 21.0,
|
| 200 |
+
"eval_accuracy": 0.46,
|
| 201 |
+
"eval_loss": 3.9925529956817627,
|
| 202 |
+
"eval_runtime": 24.8697,
|
| 203 |
+
"eval_samples_per_second": 64.335,
|
| 204 |
+
"eval_steps_per_second": 2.01,
|
| 205 |
+
"step": 798
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 22.0,
|
| 209 |
+
"eval_accuracy": 0.463125,
|
| 210 |
+
"eval_loss": 3.9888575077056885,
|
| 211 |
+
"eval_runtime": 24.9595,
|
| 212 |
+
"eval_samples_per_second": 64.104,
|
| 213 |
+
"eval_steps_per_second": 2.003,
|
| 214 |
+
"step": 836
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 23.0,
|
| 218 |
+
"eval_accuracy": 0.464375,
|
| 219 |
+
"eval_loss": 3.9854891300201416,
|
| 220 |
+
"eval_runtime": 24.5799,
|
| 221 |
+
"eval_samples_per_second": 65.094,
|
| 222 |
+
"eval_steps_per_second": 2.034,
|
| 223 |
+
"step": 874
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 24.0,
|
| 227 |
+
"eval_accuracy": 0.466875,
|
| 228 |
+
"eval_loss": 3.9826958179473877,
|
| 229 |
+
"eval_runtime": 24.5912,
|
| 230 |
+
"eval_samples_per_second": 65.064,
|
| 231 |
+
"eval_steps_per_second": 2.033,
|
| 232 |
+
"step": 912
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 25.0,
|
| 236 |
+
"eval_accuracy": 0.469375,
|
| 237 |
+
"eval_loss": 3.980361223220825,
|
| 238 |
+
"eval_runtime": 24.6399,
|
| 239 |
+
"eval_samples_per_second": 64.935,
|
| 240 |
+
"eval_steps_per_second": 2.029,
|
| 241 |
+
"step": 950
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 26.0,
|
| 245 |
+
"eval_accuracy": 0.470625,
|
| 246 |
+
"eval_loss": 3.9784324169158936,
|
| 247 |
+
"eval_runtime": 24.6458,
|
| 248 |
+
"eval_samples_per_second": 64.92,
|
| 249 |
+
"eval_steps_per_second": 2.029,
|
| 250 |
+
"step": 988
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 26.31578947368421,
|
| 254 |
+
"grad_norm": 1.1505932807922363,
|
| 255 |
+
"learning_rate": 2.568807339449541e-07,
|
| 256 |
+
"loss": 3.9504,
|
| 257 |
+
"step": 1000
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 27.0,
|
| 261 |
+
"eval_accuracy": 0.470625,
|
| 262 |
+
"eval_loss": 3.976879358291626,
|
| 263 |
+
"eval_runtime": 24.7593,
|
| 264 |
+
"eval_samples_per_second": 64.622,
|
| 265 |
+
"eval_steps_per_second": 2.019,
|
| 266 |
+
"step": 1026
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 28.0,
|
| 270 |
+
"eval_accuracy": 0.473125,
|
| 271 |
+
"eval_loss": 3.9757637977600098,
|
| 272 |
+
"eval_runtime": 24.4731,
|
| 273 |
+
"eval_samples_per_second": 65.378,
|
| 274 |
+
"eval_steps_per_second": 2.043,
|
| 275 |
+
"step": 1064
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 29.0,
|
| 279 |
+
"eval_accuracy": 0.473125,
|
| 280 |
+
"eval_loss": 3.9750988483428955,
|
| 281 |
+
"eval_runtime": 24.4848,
|
| 282 |
+
"eval_samples_per_second": 65.347,
|
| 283 |
+
"eval_steps_per_second": 2.042,
|
| 284 |
+
"step": 1102
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 30.0,
|
| 288 |
+
"eval_accuracy": 0.47375,
|
| 289 |
+
"eval_loss": 3.9748706817626953,
|
| 290 |
+
"eval_runtime": 24.6946,
|
| 291 |
+
"eval_samples_per_second": 64.791,
|
| 292 |
+
"eval_steps_per_second": 2.025,
|
| 293 |
+
"step": 1140
|
| 294 |
+
}
|
| 295 |
+
],
|
| 296 |
+
"logging_steps": 500,
|
| 297 |
+
"max_steps": 1140,
|
| 298 |
+
"num_input_tokens_seen": 0,
|
| 299 |
+
"num_train_epochs": 30,
|
| 300 |
+
"save_steps": 500,
|
| 301 |
+
"stateful_callbacks": {
|
| 302 |
+
"TrainerControl": {
|
| 303 |
+
"args": {
|
| 304 |
+
"should_epoch_stop": false,
|
| 305 |
+
"should_evaluate": false,
|
| 306 |
+
"should_log": false,
|
| 307 |
+
"should_save": true,
|
| 308 |
+
"should_training_stop": true
|
| 309 |
+
},
|
| 310 |
+
"attributes": {}
|
| 311 |
+
}
|
| 312 |
+
},
|
| 313 |
+
"total_flos": 5.583323767504896e+18,
|
| 314 |
+
"train_batch_size": 64,
|
| 315 |
+
"trial_name": null,
|
| 316 |
+
"trial_params": null
|
| 317 |
+
}
|
checkpoint-1140/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:95fe0cc363c17b82270e3c3348089686d420c74e46259d9642de869e8363cdf9
|
| 3 |
+
size 4603
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 343463912
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ed6113566cd92744fa8be6bfa6d67a9422ba8440eeedcf18c2bed7cd7d77108a
|
| 3 |
size 343463912
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4603
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:95fe0cc363c17b82270e3c3348089686d420c74e46259d9642de869e8363cdf9
|
| 3 |
size 4603
|