Upload folder using huggingface_hub
Browse files- checkpoint-5640/config.json +129 -0
- checkpoint-5640/model.safetensors +3 -0
- checkpoint-5640/optimizer.pt +3 -0
- checkpoint-5640/preprocessor_config.json +23 -0
- checkpoint-5640/rng_state.pth +3 -0
- checkpoint-5640/scheduler.pt +3 -0
- checkpoint-5640/trainer_state.json +511 -0
- checkpoint-5640/training_args.bin +3 -0
- model.safetensors +1 -1
- training_args.bin +1 -1
checkpoint-5640/config.json
ADDED
|
@@ -0,0 +1,129 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"ViTForImageClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.0,
|
| 6 |
+
"encoder_stride": 16,
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.0,
|
| 9 |
+
"hidden_size": 768,
|
| 10 |
+
"id2label": {
|
| 11 |
+
"0": "adidas_forum_high",
|
| 12 |
+
"1": "adidas_forum_low",
|
| 13 |
+
"2": "adidas_gazelle",
|
| 14 |
+
"3": "adidas_nmd_r1",
|
| 15 |
+
"4": "adidas_samba",
|
| 16 |
+
"5": "adidas_stan_smith",
|
| 17 |
+
"6": "adidas_superstar",
|
| 18 |
+
"7": "adidas_ultraboost",
|
| 19 |
+
"8": "asics_gel-lyte_iii",
|
| 20 |
+
"9": "converse_chuck_70_high",
|
| 21 |
+
"10": "converse_chuck_70_low",
|
| 22 |
+
"11": "converse_chuck_taylor_all-star_high",
|
| 23 |
+
"12": "converse_chuck_taylor_all-star_low",
|
| 24 |
+
"13": "converse_one_star",
|
| 25 |
+
"14": "new_balance_327",
|
| 26 |
+
"15": "new_balance_550",
|
| 27 |
+
"16": "new_balance_574",
|
| 28 |
+
"17": "new_balance_990",
|
| 29 |
+
"18": "new_balance_992",
|
| 30 |
+
"19": "nike_air_force_1_high",
|
| 31 |
+
"20": "nike_air_force_1_low",
|
| 32 |
+
"21": "nike_air_force_1_mid",
|
| 33 |
+
"22": "nike_air_jordan_11",
|
| 34 |
+
"23": "nike_air_jordan_1_high",
|
| 35 |
+
"24": "nike_air_jordan_1_low",
|
| 36 |
+
"25": "nike_air_jordan_3",
|
| 37 |
+
"26": "nike_air_jordan_4",
|
| 38 |
+
"27": "nike_air_max_1",
|
| 39 |
+
"28": "nike_air_max_270",
|
| 40 |
+
"29": "nike_air_max_90",
|
| 41 |
+
"30": "nike_air_max_95",
|
| 42 |
+
"31": "nike_air_max_97",
|
| 43 |
+
"32": "nike_air_max_plus_(tn)",
|
| 44 |
+
"33": "nike_air_vapormax_flyknit",
|
| 45 |
+
"34": "nike_air_vapormax_plus",
|
| 46 |
+
"35": "nike_blazer_mid_77",
|
| 47 |
+
"36": "nike_cortez",
|
| 48 |
+
"37": "nike_dunk_high",
|
| 49 |
+
"38": "nike_dunk_low",
|
| 50 |
+
"39": "puma_suede_classic",
|
| 51 |
+
"40": "reebok_classic_leather",
|
| 52 |
+
"41": "reebok_club_c_85",
|
| 53 |
+
"42": "salomon_xt-6",
|
| 54 |
+
"43": "vans_authentic",
|
| 55 |
+
"44": "vans_old_skool",
|
| 56 |
+
"45": "vans_sk8-hi",
|
| 57 |
+
"46": "vans_slip-on_checkerboard",
|
| 58 |
+
"47": "yeezy_700_wave_runner",
|
| 59 |
+
"48": "yeezy_boost_350_v2",
|
| 60 |
+
"49": "yeezy_slide"
|
| 61 |
+
},
|
| 62 |
+
"image_size": 224,
|
| 63 |
+
"initializer_range": 0.02,
|
| 64 |
+
"intermediate_size": 3072,
|
| 65 |
+
"label2id": {
|
| 66 |
+
"adidas_forum_high": 0,
|
| 67 |
+
"adidas_forum_low": 1,
|
| 68 |
+
"adidas_gazelle": 2,
|
| 69 |
+
"adidas_nmd_r1": 3,
|
| 70 |
+
"adidas_samba": 4,
|
| 71 |
+
"adidas_stan_smith": 5,
|
| 72 |
+
"adidas_superstar": 6,
|
| 73 |
+
"adidas_ultraboost": 7,
|
| 74 |
+
"asics_gel-lyte_iii": 8,
|
| 75 |
+
"converse_chuck_70_high": 9,
|
| 76 |
+
"converse_chuck_70_low": 10,
|
| 77 |
+
"converse_chuck_taylor_all-star_high": 11,
|
| 78 |
+
"converse_chuck_taylor_all-star_low": 12,
|
| 79 |
+
"converse_one_star": 13,
|
| 80 |
+
"new_balance_327": 14,
|
| 81 |
+
"new_balance_550": 15,
|
| 82 |
+
"new_balance_574": 16,
|
| 83 |
+
"new_balance_990": 17,
|
| 84 |
+
"new_balance_992": 18,
|
| 85 |
+
"nike_air_force_1_high": 19,
|
| 86 |
+
"nike_air_force_1_low": 20,
|
| 87 |
+
"nike_air_force_1_mid": 21,
|
| 88 |
+
"nike_air_jordan_11": 22,
|
| 89 |
+
"nike_air_jordan_1_high": 23,
|
| 90 |
+
"nike_air_jordan_1_low": 24,
|
| 91 |
+
"nike_air_jordan_3": 25,
|
| 92 |
+
"nike_air_jordan_4": 26,
|
| 93 |
+
"nike_air_max_1": 27,
|
| 94 |
+
"nike_air_max_270": 28,
|
| 95 |
+
"nike_air_max_90": 29,
|
| 96 |
+
"nike_air_max_95": 30,
|
| 97 |
+
"nike_air_max_97": 31,
|
| 98 |
+
"nike_air_max_plus_(tn)": 32,
|
| 99 |
+
"nike_air_vapormax_flyknit": 33,
|
| 100 |
+
"nike_air_vapormax_plus": 34,
|
| 101 |
+
"nike_blazer_mid_77": 35,
|
| 102 |
+
"nike_cortez": 36,
|
| 103 |
+
"nike_dunk_high": 37,
|
| 104 |
+
"nike_dunk_low": 38,
|
| 105 |
+
"puma_suede_classic": 39,
|
| 106 |
+
"reebok_classic_leather": 40,
|
| 107 |
+
"reebok_club_c_85": 41,
|
| 108 |
+
"salomon_xt-6": 42,
|
| 109 |
+
"vans_authentic": 43,
|
| 110 |
+
"vans_old_skool": 44,
|
| 111 |
+
"vans_sk8-hi": 45,
|
| 112 |
+
"vans_slip-on_checkerboard": 46,
|
| 113 |
+
"yeezy_700_wave_runner": 47,
|
| 114 |
+
"yeezy_boost_350_v2": 48,
|
| 115 |
+
"yeezy_slide": 49
|
| 116 |
+
},
|
| 117 |
+
"layer_norm_eps": 1e-12,
|
| 118 |
+
"model_type": "vit",
|
| 119 |
+
"num_attention_heads": 12,
|
| 120 |
+
"num_channels": 3,
|
| 121 |
+
"num_hidden_layers": 12,
|
| 122 |
+
"patch_size": 16,
|
| 123 |
+
"pooler_act": "tanh",
|
| 124 |
+
"pooler_output_size": 768,
|
| 125 |
+
"problem_type": "single_label_classification",
|
| 126 |
+
"qkv_bias": true,
|
| 127 |
+
"torch_dtype": "float32",
|
| 128 |
+
"transformers_version": "4.51.3"
|
| 129 |
+
}
|
checkpoint-5640/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0c4745e3fdcd455e985f55ae003c88c87d197975d2307ad5dda833f8c30809d0
|
| 3 |
+
size 343371632
|
checkpoint-5640/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3310bf09978d639f503bac2f5400e0837b4ed61f1f3db0b2565bafa6ba106db1
|
| 3 |
+
size 686864186
|
checkpoint-5640/preprocessor_config.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_convert_rgb": null,
|
| 3 |
+
"do_normalize": true,
|
| 4 |
+
"do_rescale": true,
|
| 5 |
+
"do_resize": true,
|
| 6 |
+
"image_mean": [
|
| 7 |
+
0.5,
|
| 8 |
+
0.5,
|
| 9 |
+
0.5
|
| 10 |
+
],
|
| 11 |
+
"image_processor_type": "ViTImageProcessor",
|
| 12 |
+
"image_std": [
|
| 13 |
+
0.5,
|
| 14 |
+
0.5,
|
| 15 |
+
0.5
|
| 16 |
+
],
|
| 17 |
+
"resample": 2,
|
| 18 |
+
"rescale_factor": 0.00392156862745098,
|
| 19 |
+
"size": {
|
| 20 |
+
"height": 224,
|
| 21 |
+
"width": 224
|
| 22 |
+
}
|
| 23 |
+
}
|
checkpoint-5640/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5591368b5e645f6f84c8526df6dfd0c8c2904179c960a8a757d795a3b9c9edef
|
| 3 |
+
size 14244
|
checkpoint-5640/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:58a6de07b073daad8f68678fd6f5f692094c61f764b1579a3cc1b60ed9aa9687
|
| 3 |
+
size 1064
|
checkpoint-5640/trainer_state.json
ADDED
|
@@ -0,0 +1,511 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 5640,
|
| 3 |
+
"best_metric": 2.5463945865631104,
|
| 4 |
+
"best_model_checkpoint": "popular_sneakers_detection/checkpoint-5640",
|
| 5 |
+
"epoch": 40.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 5640,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 1.0,
|
| 14 |
+
"eval_accuracy": 0.042333333333333334,
|
| 15 |
+
"eval_loss": 3.8944478034973145,
|
| 16 |
+
"eval_model_preparation_time": 0.0041,
|
| 17 |
+
"eval_runtime": 27.4259,
|
| 18 |
+
"eval_samples_per_second": 109.386,
|
| 19 |
+
"eval_steps_per_second": 13.673,
|
| 20 |
+
"step": 141
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 2.0,
|
| 24 |
+
"eval_accuracy": 0.06366666666666666,
|
| 25 |
+
"eval_loss": 3.8702337741851807,
|
| 26 |
+
"eval_model_preparation_time": 0.0041,
|
| 27 |
+
"eval_runtime": 26.7233,
|
| 28 |
+
"eval_samples_per_second": 112.261,
|
| 29 |
+
"eval_steps_per_second": 14.033,
|
| 30 |
+
"step": 282
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 3.0,
|
| 34 |
+
"eval_accuracy": 0.094,
|
| 35 |
+
"eval_loss": 3.8375682830810547,
|
| 36 |
+
"eval_model_preparation_time": 0.0041,
|
| 37 |
+
"eval_runtime": 27.5229,
|
| 38 |
+
"eval_samples_per_second": 109.0,
|
| 39 |
+
"eval_steps_per_second": 13.625,
|
| 40 |
+
"step": 423
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 3.546099290780142,
|
| 44 |
+
"grad_norm": 1.5614980459213257,
|
| 45 |
+
"learning_rate": 4.598389982110913e-06,
|
| 46 |
+
"loss": 3.8547,
|
| 47 |
+
"step": 500
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 4.0,
|
| 51 |
+
"eval_accuracy": 0.13733333333333334,
|
| 52 |
+
"eval_loss": 3.79917573928833,
|
| 53 |
+
"eval_model_preparation_time": 0.0041,
|
| 54 |
+
"eval_runtime": 28.4527,
|
| 55 |
+
"eval_samples_per_second": 105.438,
|
| 56 |
+
"eval_steps_per_second": 13.18,
|
| 57 |
+
"step": 564
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 5.0,
|
| 61 |
+
"eval_accuracy": 0.19166666666666668,
|
| 62 |
+
"eval_loss": 3.7531919479370117,
|
| 63 |
+
"eval_model_preparation_time": 0.0041,
|
| 64 |
+
"eval_runtime": 27.2395,
|
| 65 |
+
"eval_samples_per_second": 110.134,
|
| 66 |
+
"eval_steps_per_second": 13.767,
|
| 67 |
+
"step": 705
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 6.0,
|
| 71 |
+
"eval_accuracy": 0.25533333333333336,
|
| 72 |
+
"eval_loss": 3.6991868019104004,
|
| 73 |
+
"eval_model_preparation_time": 0.0041,
|
| 74 |
+
"eval_runtime": 27.7403,
|
| 75 |
+
"eval_samples_per_second": 108.146,
|
| 76 |
+
"eval_steps_per_second": 13.518,
|
| 77 |
+
"step": 846
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 7.0,
|
| 81 |
+
"eval_accuracy": 0.32566666666666666,
|
| 82 |
+
"eval_loss": 3.6358957290649414,
|
| 83 |
+
"eval_model_preparation_time": 0.0041,
|
| 84 |
+
"eval_runtime": 27.4022,
|
| 85 |
+
"eval_samples_per_second": 109.48,
|
| 86 |
+
"eval_steps_per_second": 13.685,
|
| 87 |
+
"step": 987
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 7.092198581560283,
|
| 91 |
+
"grad_norm": 1.7323589324951172,
|
| 92 |
+
"learning_rate": 4.151162790697675e-06,
|
| 93 |
+
"loss": 3.6677,
|
| 94 |
+
"step": 1000
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 8.0,
|
| 98 |
+
"eval_accuracy": 0.391,
|
| 99 |
+
"eval_loss": 3.561831474304199,
|
| 100 |
+
"eval_model_preparation_time": 0.0041,
|
| 101 |
+
"eval_runtime": 28.1913,
|
| 102 |
+
"eval_samples_per_second": 106.416,
|
| 103 |
+
"eval_steps_per_second": 13.302,
|
| 104 |
+
"step": 1128
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 9.0,
|
| 108 |
+
"eval_accuracy": 0.42866666666666664,
|
| 109 |
+
"eval_loss": 3.492708444595337,
|
| 110 |
+
"eval_model_preparation_time": 0.0041,
|
| 111 |
+
"eval_runtime": 27.5068,
|
| 112 |
+
"eval_samples_per_second": 109.064,
|
| 113 |
+
"eval_steps_per_second": 13.633,
|
| 114 |
+
"step": 1269
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 10.0,
|
| 118 |
+
"eval_accuracy": 0.4796666666666667,
|
| 119 |
+
"eval_loss": 3.4180657863616943,
|
| 120 |
+
"eval_model_preparation_time": 0.0041,
|
| 121 |
+
"eval_runtime": 27.5222,
|
| 122 |
+
"eval_samples_per_second": 109.003,
|
| 123 |
+
"eval_steps_per_second": 13.625,
|
| 124 |
+
"step": 1410
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 10.638297872340425,
|
| 128 |
+
"grad_norm": 1.7677080631256104,
|
| 129 |
+
"learning_rate": 3.7039355992844365e-06,
|
| 130 |
+
"loss": 3.4053,
|
| 131 |
+
"step": 1500
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 11.0,
|
| 135 |
+
"eval_accuracy": 0.49633333333333335,
|
| 136 |
+
"eval_loss": 3.353686571121216,
|
| 137 |
+
"eval_model_preparation_time": 0.0041,
|
| 138 |
+
"eval_runtime": 27.4731,
|
| 139 |
+
"eval_samples_per_second": 109.198,
|
| 140 |
+
"eval_steps_per_second": 13.65,
|
| 141 |
+
"step": 1551
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 12.0,
|
| 145 |
+
"eval_accuracy": 0.5183333333333333,
|
| 146 |
+
"eval_loss": 3.29174542427063,
|
| 147 |
+
"eval_model_preparation_time": 0.0041,
|
| 148 |
+
"eval_runtime": 27.3958,
|
| 149 |
+
"eval_samples_per_second": 109.506,
|
| 150 |
+
"eval_steps_per_second": 13.688,
|
| 151 |
+
"step": 1692
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 13.0,
|
| 155 |
+
"eval_accuracy": 0.5416666666666666,
|
| 156 |
+
"eval_loss": 3.230329990386963,
|
| 157 |
+
"eval_model_preparation_time": 0.0041,
|
| 158 |
+
"eval_runtime": 27.2445,
|
| 159 |
+
"eval_samples_per_second": 110.114,
|
| 160 |
+
"eval_steps_per_second": 13.764,
|
| 161 |
+
"step": 1833
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 14.0,
|
| 165 |
+
"eval_accuracy": 0.5566666666666666,
|
| 166 |
+
"eval_loss": 3.1734297275543213,
|
| 167 |
+
"eval_model_preparation_time": 0.0041,
|
| 168 |
+
"eval_runtime": 27.0335,
|
| 169 |
+
"eval_samples_per_second": 110.973,
|
| 170 |
+
"eval_steps_per_second": 13.872,
|
| 171 |
+
"step": 1974
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 14.184397163120567,
|
| 175 |
+
"grad_norm": 1.914097785949707,
|
| 176 |
+
"learning_rate": 3.2567084078711986e-06,
|
| 177 |
+
"loss": 3.1467,
|
| 178 |
+
"step": 2000
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 15.0,
|
| 182 |
+
"eval_accuracy": 0.5753333333333334,
|
| 183 |
+
"eval_loss": 3.1160686016082764,
|
| 184 |
+
"eval_model_preparation_time": 0.0041,
|
| 185 |
+
"eval_runtime": 26.9914,
|
| 186 |
+
"eval_samples_per_second": 111.146,
|
| 187 |
+
"eval_steps_per_second": 13.893,
|
| 188 |
+
"step": 2115
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 16.0,
|
| 192 |
+
"eval_accuracy": 0.5836666666666667,
|
| 193 |
+
"eval_loss": 3.0692453384399414,
|
| 194 |
+
"eval_model_preparation_time": 0.0041,
|
| 195 |
+
"eval_runtime": 27.4615,
|
| 196 |
+
"eval_samples_per_second": 109.244,
|
| 197 |
+
"eval_steps_per_second": 13.655,
|
| 198 |
+
"step": 2256
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 17.0,
|
| 202 |
+
"eval_accuracy": 0.602,
|
| 203 |
+
"eval_loss": 3.0239269733428955,
|
| 204 |
+
"eval_model_preparation_time": 0.0041,
|
| 205 |
+
"eval_runtime": 27.511,
|
| 206 |
+
"eval_samples_per_second": 109.047,
|
| 207 |
+
"eval_steps_per_second": 13.631,
|
| 208 |
+
"step": 2397
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 17.73049645390071,
|
| 212 |
+
"grad_norm": 1.9646587371826172,
|
| 213 |
+
"learning_rate": 2.809481216457961e-06,
|
| 214 |
+
"loss": 2.9364,
|
| 215 |
+
"step": 2500
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 18.0,
|
| 219 |
+
"eval_accuracy": 0.604,
|
| 220 |
+
"eval_loss": 2.983224868774414,
|
| 221 |
+
"eval_model_preparation_time": 0.0041,
|
| 222 |
+
"eval_runtime": 27.5742,
|
| 223 |
+
"eval_samples_per_second": 108.797,
|
| 224 |
+
"eval_steps_per_second": 13.6,
|
| 225 |
+
"step": 2538
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 19.0,
|
| 229 |
+
"eval_accuracy": 0.6146666666666667,
|
| 230 |
+
"eval_loss": 2.9459805488586426,
|
| 231 |
+
"eval_model_preparation_time": 0.0041,
|
| 232 |
+
"eval_runtime": 27.3324,
|
| 233 |
+
"eval_samples_per_second": 109.76,
|
| 234 |
+
"eval_steps_per_second": 13.72,
|
| 235 |
+
"step": 2679
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 20.0,
|
| 239 |
+
"eval_accuracy": 0.6283333333333333,
|
| 240 |
+
"eval_loss": 2.9052772521972656,
|
| 241 |
+
"eval_model_preparation_time": 0.0041,
|
| 242 |
+
"eval_runtime": 27.2251,
|
| 243 |
+
"eval_samples_per_second": 110.192,
|
| 244 |
+
"eval_steps_per_second": 13.774,
|
| 245 |
+
"step": 2820
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 21.0,
|
| 249 |
+
"eval_accuracy": 0.6386666666666667,
|
| 250 |
+
"eval_loss": 2.8673319816589355,
|
| 251 |
+
"eval_model_preparation_time": 0.0041,
|
| 252 |
+
"eval_runtime": 27.5302,
|
| 253 |
+
"eval_samples_per_second": 108.971,
|
| 254 |
+
"eval_steps_per_second": 13.621,
|
| 255 |
+
"step": 2961
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 21.27659574468085,
|
| 259 |
+
"grad_norm": 2.0768418312072754,
|
| 260 |
+
"learning_rate": 2.3622540250447227e-06,
|
| 261 |
+
"loss": 2.7701,
|
| 262 |
+
"step": 3000
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 22.0,
|
| 266 |
+
"eval_accuracy": 0.642,
|
| 267 |
+
"eval_loss": 2.8295445442199707,
|
| 268 |
+
"eval_model_preparation_time": 0.0041,
|
| 269 |
+
"eval_runtime": 27.3684,
|
| 270 |
+
"eval_samples_per_second": 109.616,
|
| 271 |
+
"eval_steps_per_second": 13.702,
|
| 272 |
+
"step": 3102
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 23.0,
|
| 276 |
+
"eval_accuracy": 0.658,
|
| 277 |
+
"eval_loss": 2.7955031394958496,
|
| 278 |
+
"eval_model_preparation_time": 0.0041,
|
| 279 |
+
"eval_runtime": 27.6846,
|
| 280 |
+
"eval_samples_per_second": 108.363,
|
| 281 |
+
"eval_steps_per_second": 13.545,
|
| 282 |
+
"step": 3243
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 24.0,
|
| 286 |
+
"eval_accuracy": 0.6533333333333333,
|
| 287 |
+
"eval_loss": 2.7718265056610107,
|
| 288 |
+
"eval_model_preparation_time": 0.0041,
|
| 289 |
+
"eval_runtime": 28.3048,
|
| 290 |
+
"eval_samples_per_second": 105.989,
|
| 291 |
+
"eval_steps_per_second": 13.249,
|
| 292 |
+
"step": 3384
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 24.822695035460992,
|
| 296 |
+
"grad_norm": 2.5180611610412598,
|
| 297 |
+
"learning_rate": 1.915026833631485e-06,
|
| 298 |
+
"loss": 2.6383,
|
| 299 |
+
"step": 3500
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 25.0,
|
| 303 |
+
"eval_accuracy": 0.6656666666666666,
|
| 304 |
+
"eval_loss": 2.7361390590667725,
|
| 305 |
+
"eval_model_preparation_time": 0.0041,
|
| 306 |
+
"eval_runtime": 27.5955,
|
| 307 |
+
"eval_samples_per_second": 108.714,
|
| 308 |
+
"eval_steps_per_second": 13.589,
|
| 309 |
+
"step": 3525
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 26.0,
|
| 313 |
+
"eval_accuracy": 0.6693333333333333,
|
| 314 |
+
"eval_loss": 2.712496519088745,
|
| 315 |
+
"eval_model_preparation_time": 0.0041,
|
| 316 |
+
"eval_runtime": 27.5376,
|
| 317 |
+
"eval_samples_per_second": 108.942,
|
| 318 |
+
"eval_steps_per_second": 13.618,
|
| 319 |
+
"step": 3666
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 27.0,
|
| 323 |
+
"eval_accuracy": 0.6656666666666666,
|
| 324 |
+
"eval_loss": 2.695204257965088,
|
| 325 |
+
"eval_model_preparation_time": 0.0041,
|
| 326 |
+
"eval_runtime": 27.822,
|
| 327 |
+
"eval_samples_per_second": 107.828,
|
| 328 |
+
"eval_steps_per_second": 13.479,
|
| 329 |
+
"step": 3807
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 28.0,
|
| 333 |
+
"eval_accuracy": 0.6756666666666666,
|
| 334 |
+
"eval_loss": 2.6688482761383057,
|
| 335 |
+
"eval_model_preparation_time": 0.0041,
|
| 336 |
+
"eval_runtime": 27.4009,
|
| 337 |
+
"eval_samples_per_second": 109.485,
|
| 338 |
+
"eval_steps_per_second": 13.686,
|
| 339 |
+
"step": 3948
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 28.368794326241133,
|
| 343 |
+
"grad_norm": 2.0685389041900635,
|
| 344 |
+
"learning_rate": 1.4677996422182469e-06,
|
| 345 |
+
"loss": 2.5324,
|
| 346 |
+
"step": 4000
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 29.0,
|
| 350 |
+
"eval_accuracy": 0.6776666666666666,
|
| 351 |
+
"eval_loss": 2.651216745376587,
|
| 352 |
+
"eval_model_preparation_time": 0.0041,
|
| 353 |
+
"eval_runtime": 27.4184,
|
| 354 |
+
"eval_samples_per_second": 109.415,
|
| 355 |
+
"eval_steps_per_second": 13.677,
|
| 356 |
+
"step": 4089
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 30.0,
|
| 360 |
+
"eval_accuracy": 0.684,
|
| 361 |
+
"eval_loss": 2.631049633026123,
|
| 362 |
+
"eval_model_preparation_time": 0.0041,
|
| 363 |
+
"eval_runtime": 27.4391,
|
| 364 |
+
"eval_samples_per_second": 109.333,
|
| 365 |
+
"eval_steps_per_second": 13.667,
|
| 366 |
+
"step": 4230
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 31.0,
|
| 370 |
+
"eval_accuracy": 0.6876666666666666,
|
| 371 |
+
"eval_loss": 2.61474347114563,
|
| 372 |
+
"eval_model_preparation_time": 0.0041,
|
| 373 |
+
"eval_runtime": 27.3793,
|
| 374 |
+
"eval_samples_per_second": 109.572,
|
| 375 |
+
"eval_steps_per_second": 13.697,
|
| 376 |
+
"step": 4371
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 31.914893617021278,
|
| 380 |
+
"grad_norm": 2.6519553661346436,
|
| 381 |
+
"learning_rate": 1.020572450805009e-06,
|
| 382 |
+
"loss": 2.4528,
|
| 383 |
+
"step": 4500
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 32.0,
|
| 387 |
+
"eval_accuracy": 0.6866666666666666,
|
| 388 |
+
"eval_loss": 2.5972399711608887,
|
| 389 |
+
"eval_model_preparation_time": 0.0041,
|
| 390 |
+
"eval_runtime": 26.9694,
|
| 391 |
+
"eval_samples_per_second": 111.237,
|
| 392 |
+
"eval_steps_per_second": 13.905,
|
| 393 |
+
"step": 4512
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 33.0,
|
| 397 |
+
"eval_accuracy": 0.6883333333333334,
|
| 398 |
+
"eval_loss": 2.586773157119751,
|
| 399 |
+
"eval_model_preparation_time": 0.0041,
|
| 400 |
+
"eval_runtime": 27.608,
|
| 401 |
+
"eval_samples_per_second": 108.664,
|
| 402 |
+
"eval_steps_per_second": 13.583,
|
| 403 |
+
"step": 4653
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 34.0,
|
| 407 |
+
"eval_accuracy": 0.6876666666666666,
|
| 408 |
+
"eval_loss": 2.576591730117798,
|
| 409 |
+
"eval_model_preparation_time": 0.0041,
|
| 410 |
+
"eval_runtime": 27.5729,
|
| 411 |
+
"eval_samples_per_second": 108.802,
|
| 412 |
+
"eval_steps_per_second": 13.6,
|
| 413 |
+
"step": 4794
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 35.0,
|
| 417 |
+
"eval_accuracy": 0.6926666666666667,
|
| 418 |
+
"eval_loss": 2.569082736968994,
|
| 419 |
+
"eval_model_preparation_time": 0.0041,
|
| 420 |
+
"eval_runtime": 27.9391,
|
| 421 |
+
"eval_samples_per_second": 107.376,
|
| 422 |
+
"eval_steps_per_second": 13.422,
|
| 423 |
+
"step": 4935
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 35.46099290780142,
|
| 427 |
+
"grad_norm": 2.5418479442596436,
|
| 428 |
+
"learning_rate": 5.733452593917711e-07,
|
| 429 |
+
"loss": 2.4022,
|
| 430 |
+
"step": 5000
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 36.0,
|
| 434 |
+
"eval_accuracy": 0.693,
|
| 435 |
+
"eval_loss": 2.5580801963806152,
|
| 436 |
+
"eval_model_preparation_time": 0.0041,
|
| 437 |
+
"eval_runtime": 27.1847,
|
| 438 |
+
"eval_samples_per_second": 110.356,
|
| 439 |
+
"eval_steps_per_second": 13.795,
|
| 440 |
+
"step": 5076
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 37.0,
|
| 444 |
+
"eval_accuracy": 0.694,
|
| 445 |
+
"eval_loss": 2.5533597469329834,
|
| 446 |
+
"eval_model_preparation_time": 0.0041,
|
| 447 |
+
"eval_runtime": 27.3678,
|
| 448 |
+
"eval_samples_per_second": 109.618,
|
| 449 |
+
"eval_steps_per_second": 13.702,
|
| 450 |
+
"step": 5217
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 38.0,
|
| 454 |
+
"eval_accuracy": 0.695,
|
| 455 |
+
"eval_loss": 2.5490283966064453,
|
| 456 |
+
"eval_model_preparation_time": 0.0041,
|
| 457 |
+
"eval_runtime": 27.0235,
|
| 458 |
+
"eval_samples_per_second": 111.014,
|
| 459 |
+
"eval_steps_per_second": 13.877,
|
| 460 |
+
"step": 5358
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 39.0,
|
| 464 |
+
"eval_accuracy": 0.694,
|
| 465 |
+
"eval_loss": 2.547304153442383,
|
| 466 |
+
"eval_model_preparation_time": 0.0041,
|
| 467 |
+
"eval_runtime": 27.9193,
|
| 468 |
+
"eval_samples_per_second": 107.453,
|
| 469 |
+
"eval_steps_per_second": 13.432,
|
| 470 |
+
"step": 5499
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 39.00709219858156,
|
| 474 |
+
"grad_norm": 2.5994229316711426,
|
| 475 |
+
"learning_rate": 1.2611806797853311e-07,
|
| 476 |
+
"loss": 2.3657,
|
| 477 |
+
"step": 5500
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 40.0,
|
| 481 |
+
"eval_accuracy": 0.6943333333333334,
|
| 482 |
+
"eval_loss": 2.5463945865631104,
|
| 483 |
+
"eval_model_preparation_time": 0.0041,
|
| 484 |
+
"eval_runtime": 27.6641,
|
| 485 |
+
"eval_samples_per_second": 108.444,
|
| 486 |
+
"eval_steps_per_second": 13.555,
|
| 487 |
+
"step": 5640
|
| 488 |
+
}
|
| 489 |
+
],
|
| 490 |
+
"logging_steps": 500,
|
| 491 |
+
"max_steps": 5640,
|
| 492 |
+
"num_input_tokens_seen": 0,
|
| 493 |
+
"num_train_epochs": 40,
|
| 494 |
+
"save_steps": 500,
|
| 495 |
+
"stateful_callbacks": {
|
| 496 |
+
"TrainerControl": {
|
| 497 |
+
"args": {
|
| 498 |
+
"should_epoch_stop": false,
|
| 499 |
+
"should_evaluate": false,
|
| 500 |
+
"should_log": false,
|
| 501 |
+
"should_save": true,
|
| 502 |
+
"should_training_stop": true
|
| 503 |
+
},
|
| 504 |
+
"attributes": {}
|
| 505 |
+
}
|
| 506 |
+
},
|
| 507 |
+
"total_flos": 1.395455892332544e+19,
|
| 508 |
+
"train_batch_size": 32,
|
| 509 |
+
"trial_name": null,
|
| 510 |
+
"trial_params": null
|
| 511 |
+
}
|
checkpoint-5640/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a85b11601d829e022858ca77f9c10bbf712a1c51fb1d678c6dc467ebd3fa2edc
|
| 3 |
+
size 5240
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 343371632
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0c4745e3fdcd455e985f55ae003c88c87d197975d2307ad5dda833f8c30809d0
|
| 3 |
size 343371632
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5240
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a85b11601d829e022858ca77f9c10bbf712a1c51fb1d678c6dc467ebd3fa2edc
|
| 3 |
size 5240
|