Spaces:
Sleeping
Sleeping
Upload 11 files
Browse files- data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/span_classifier/Rxi8b70XJA/checkpoint.pt +3 -0
- data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/span_classifier/Rxi8b70XJA/config.json +29 -0
- data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/span_classifier/Rxi8b70XJA/config_train.json +51 -0
- data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/dR8xQB4ODU/checkpoint.pt +3 -0
- data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/dR8xQB4ODU/config.json +166 -0
- data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/dR8xQB4ODU/config_train.json +51 -0
- data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/pbK46jjAVx/checkpoint.pt +3 -0
- data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/pbK46jjAVx/config.json +166 -0
- data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/pbK46jjAVx/config_train.json +51 -0
- data/meta-llama/Llama-3.2-1B/STOKE_100/config.json +8 -0
- data/meta-llama/Llama-3.2-1B/STOKE_100/stoke_config.json +10 -0
data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/span_classifier/Rxi8b70XJA/checkpoint.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:25951d9b73437a7aa344f4c207cbda2f88d9bf5fa94d1a779617948b18a1c4ed
|
| 3 |
+
size 8439912
|
data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/span_classifier/Rxi8b70XJA/config.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": "meta-llama/Llama-3.2-1B",
|
| 3 |
+
"type": "span_classifier",
|
| 4 |
+
"label_map": [
|
| 5 |
+
"no_span",
|
| 6 |
+
"span"
|
| 7 |
+
],
|
| 8 |
+
"learning_rate": 0.0003,
|
| 9 |
+
"classifier_dim": 4096,
|
| 10 |
+
"loss_weights": [
|
| 11 |
+
1.0,
|
| 12 |
+
1.0
|
| 13 |
+
],
|
| 14 |
+
"identifier": "Rxi8b70XJA",
|
| 15 |
+
"best_f1_validation": 0.8677362203598022,
|
| 16 |
+
"best_f1_validation_classwise": {
|
| 17 |
+
"span": {
|
| 18 |
+
"p": 0.896858811378479,
|
| 19 |
+
"r": 0.8404456377029419,
|
| 20 |
+
"f": 0.867736279964447,
|
| 21 |
+
"s": 24324.0
|
| 22 |
+
},
|
| 23 |
+
"macro": {
|
| 24 |
+
"p": 0.896858811378479,
|
| 25 |
+
"r": 0.8404456377029419,
|
| 26 |
+
"f": 0.867736279964447
|
| 27 |
+
}
|
| 28 |
+
}
|
| 29 |
+
}
|
data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/span_classifier/Rxi8b70XJA/config_train.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"path": "data/meta-llama/Llama-3.2-1B/STOKE_100",
|
| 3 |
+
"splits": [
|
| 4 |
+
"train",
|
| 5 |
+
"validation"
|
| 6 |
+
],
|
| 7 |
+
"layers": [
|
| 8 |
+
8,
|
| 9 |
+
9,
|
| 10 |
+
10,
|
| 11 |
+
11,
|
| 12 |
+
12
|
| 13 |
+
],
|
| 14 |
+
"hfcache": "",
|
| 15 |
+
"classifier_dims": [
|
| 16 |
+
4096
|
| 17 |
+
],
|
| 18 |
+
"learning_rates": [
|
| 19 |
+
0.0001,
|
| 20 |
+
5e-05,
|
| 21 |
+
0.0003
|
| 22 |
+
],
|
| 23 |
+
"cuda": true,
|
| 24 |
+
"n_steps_per_epoch": 10000,
|
| 25 |
+
"n_epochs": 30,
|
| 26 |
+
"batch_size": 8,
|
| 27 |
+
"balance_loss": false,
|
| 28 |
+
"loss_weights_span": [
|
| 29 |
+
[
|
| 30 |
+
1.0,
|
| 31 |
+
1.0
|
| 32 |
+
],
|
| 33 |
+
[
|
| 34 |
+
1.0,
|
| 35 |
+
50.0
|
| 36 |
+
],
|
| 37 |
+
[
|
| 38 |
+
1.0,
|
| 39 |
+
100.0
|
| 40 |
+
]
|
| 41 |
+
],
|
| 42 |
+
"time": 1727765390.5829365,
|
| 43 |
+
"config_dataset": {
|
| 44 |
+
"generation_kwargs": {
|
| 45 |
+
"max_new_tokens": 100,
|
| 46 |
+
"repetition_penalty": 1.2
|
| 47 |
+
},
|
| 48 |
+
"model_id": "meta-llama/Llama-3.2-1B",
|
| 49 |
+
"flair_model_name": "flair/ner-english-ontonotes-large"
|
| 50 |
+
}
|
| 51 |
+
}
|
data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/dR8xQB4ODU/checkpoint.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dce5b3038d8767430a8bba16af61ec6af67c9d1aedc75a9f34c01feebac09b6e
|
| 3 |
+
size 33884328
|
data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/dR8xQB4ODU/config.json
ADDED
|
@@ -0,0 +1,166 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"layer": 10,
|
| 3 |
+
"model": "meta-llama/Llama-3.2-1B",
|
| 4 |
+
"type": "token_classifier",
|
| 5 |
+
"label_map": [
|
| 6 |
+
"O",
|
| 7 |
+
"CARDINAL",
|
| 8 |
+
"DATE",
|
| 9 |
+
"EVENT",
|
| 10 |
+
"FAC",
|
| 11 |
+
"GPE",
|
| 12 |
+
"LANGUAGE",
|
| 13 |
+
"LAW",
|
| 14 |
+
"LOC",
|
| 15 |
+
"MONEY",
|
| 16 |
+
"NORP",
|
| 17 |
+
"ORDINAL",
|
| 18 |
+
"ORG",
|
| 19 |
+
"PERCENT",
|
| 20 |
+
"PERSON",
|
| 21 |
+
"PRODUCT",
|
| 22 |
+
"QUANTITY",
|
| 23 |
+
"TIME",
|
| 24 |
+
"WORK_OF_ART"
|
| 25 |
+
],
|
| 26 |
+
"learning_rate": 5e-05,
|
| 27 |
+
"classifier_dim": 4096,
|
| 28 |
+
"loss_weights": [
|
| 29 |
+
1.0,
|
| 30 |
+
1.0,
|
| 31 |
+
1.0,
|
| 32 |
+
1.0,
|
| 33 |
+
1.0,
|
| 34 |
+
1.0,
|
| 35 |
+
1.0,
|
| 36 |
+
1.0,
|
| 37 |
+
1.0,
|
| 38 |
+
1.0,
|
| 39 |
+
1.0,
|
| 40 |
+
1.0,
|
| 41 |
+
1.0,
|
| 42 |
+
1.0,
|
| 43 |
+
1.0,
|
| 44 |
+
1.0,
|
| 45 |
+
1.0,
|
| 46 |
+
1.0,
|
| 47 |
+
1.0
|
| 48 |
+
],
|
| 49 |
+
"identifier": "dR8xQB4ODU",
|
| 50 |
+
"best_f1_validation": 0.9056437015533447,
|
| 51 |
+
"best_f1_validation_classwise": {
|
| 52 |
+
"CARDINAL": {
|
| 53 |
+
"p": 0.8679801225662231,
|
| 54 |
+
"r": 0.8777581453323364,
|
| 55 |
+
"f": 0.8728417754173279,
|
| 56 |
+
"s": 10741.0
|
| 57 |
+
},
|
| 58 |
+
"DATE": {
|
| 59 |
+
"p": 0.9519810676574707,
|
| 60 |
+
"r": 0.9389873743057251,
|
| 61 |
+
"f": 0.9454395771026611,
|
| 62 |
+
"s": 8572.0
|
| 63 |
+
},
|
| 64 |
+
"EVENT": {
|
| 65 |
+
"p": 0.8587140440940857,
|
| 66 |
+
"r": 0.8319672346115112,
|
| 67 |
+
"f": 0.8451290726661682,
|
| 68 |
+
"s": 1220.0
|
| 69 |
+
},
|
| 70 |
+
"FAC": {
|
| 71 |
+
"p": 0.8515185713768005,
|
| 72 |
+
"r": 0.8122317790985107,
|
| 73 |
+
"f": 0.8314113020896912,
|
| 74 |
+
"s": 932.0
|
| 75 |
+
},
|
| 76 |
+
"GPE": {
|
| 77 |
+
"p": 0.9000998735427856,
|
| 78 |
+
"r": 0.9094448685646057,
|
| 79 |
+
"f": 0.904748260974884,
|
| 80 |
+
"s": 6935.0
|
| 81 |
+
},
|
| 82 |
+
"LANGUAGE": {
|
| 83 |
+
"p": 0.75,
|
| 84 |
+
"r": 0.7200000286102295,
|
| 85 |
+
"f": 0.7346938848495483,
|
| 86 |
+
"s": 25.0
|
| 87 |
+
},
|
| 88 |
+
"LAW": {
|
| 89 |
+
"p": 0.8709677457809448,
|
| 90 |
+
"r": 0.73828125,
|
| 91 |
+
"f": 0.7991543412208557,
|
| 92 |
+
"s": 256.0
|
| 93 |
+
},
|
| 94 |
+
"LOC": {
|
| 95 |
+
"p": 0.8258426785469055,
|
| 96 |
+
"r": 0.7101449370384216,
|
| 97 |
+
"f": 0.7636363506317139,
|
| 98 |
+
"s": 414.0
|
| 99 |
+
},
|
| 100 |
+
"MONEY": {
|
| 101 |
+
"p": 0.876042902469635,
|
| 102 |
+
"r": 0.8626760840415955,
|
| 103 |
+
"f": 0.8693081140518188,
|
| 104 |
+
"s": 1704.0
|
| 105 |
+
},
|
| 106 |
+
"NORP": {
|
| 107 |
+
"p": 0.9160357713699341,
|
| 108 |
+
"r": 0.887333333492279,
|
| 109 |
+
"f": 0.9014561772346497,
|
| 110 |
+
"s": 1500.0
|
| 111 |
+
},
|
| 112 |
+
"ORDINAL": {
|
| 113 |
+
"p": 0.9303238391876221,
|
| 114 |
+
"r": 0.9498997926712036,
|
| 115 |
+
"f": 0.9400099515914917,
|
| 116 |
+
"s": 998.0
|
| 117 |
+
},
|
| 118 |
+
"ORG": {
|
| 119 |
+
"p": 0.8974575400352478,
|
| 120 |
+
"r": 0.8792765140533447,
|
| 121 |
+
"f": 0.8882739543914795,
|
| 122 |
+
"s": 9675.0
|
| 123 |
+
},
|
| 124 |
+
"PERCENT": {
|
| 125 |
+
"p": 0.8629592657089233,
|
| 126 |
+
"r": 0.8083720803260803,
|
| 127 |
+
"f": 0.8347742557525635,
|
| 128 |
+
"s": 1075.0
|
| 129 |
+
},
|
| 130 |
+
"PERSON": {
|
| 131 |
+
"p": 0.9707135558128357,
|
| 132 |
+
"r": 0.9713156223297119,
|
| 133 |
+
"f": 0.9710144996643066,
|
| 134 |
+
"s": 12899.0
|
| 135 |
+
},
|
| 136 |
+
"PRODUCT": {
|
| 137 |
+
"p": 0.7828418016433716,
|
| 138 |
+
"r": 0.7564767003059387,
|
| 139 |
+
"f": 0.7694334387779236,
|
| 140 |
+
"s": 386.0
|
| 141 |
+
},
|
| 142 |
+
"QUANTITY": {
|
| 143 |
+
"p": 0.8409090638160706,
|
| 144 |
+
"r": 0.7758846879005432,
|
| 145 |
+
"f": 0.8070893287658691,
|
| 146 |
+
"s": 763.0
|
| 147 |
+
},
|
| 148 |
+
"TIME": {
|
| 149 |
+
"p": 0.8710959553718567,
|
| 150 |
+
"r": 0.8373362421989441,
|
| 151 |
+
"f": 0.8538825511932373,
|
| 152 |
+
"s": 1832.0
|
| 153 |
+
},
|
| 154 |
+
"WORK_OF_ART": {
|
| 155 |
+
"p": 0.7803030014038086,
|
| 156 |
+
"r": 0.7152777910232544,
|
| 157 |
+
"f": 0.7463768124580383,
|
| 158 |
+
"s": 576.0
|
| 159 |
+
},
|
| 160 |
+
"macro": {
|
| 161 |
+
"p": 0.8669881820678711,
|
| 162 |
+
"r": 0.8323702216148376,
|
| 163 |
+
"f": 0.8488152027130127
|
| 164 |
+
}
|
| 165 |
+
}
|
| 166 |
+
}
|
data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/dR8xQB4ODU/config_train.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"path": "data/meta-llama/Llama-3.2-1B/STOKE_100",
|
| 3 |
+
"splits": [
|
| 4 |
+
"train",
|
| 5 |
+
"validation"
|
| 6 |
+
],
|
| 7 |
+
"layers": [
|
| 8 |
+
8,
|
| 9 |
+
9,
|
| 10 |
+
10,
|
| 11 |
+
11,
|
| 12 |
+
12
|
| 13 |
+
],
|
| 14 |
+
"hfcache": "",
|
| 15 |
+
"classifier_dims": [
|
| 16 |
+
4096
|
| 17 |
+
],
|
| 18 |
+
"learning_rates": [
|
| 19 |
+
0.0001,
|
| 20 |
+
5e-05,
|
| 21 |
+
0.0003
|
| 22 |
+
],
|
| 23 |
+
"cuda": true,
|
| 24 |
+
"n_steps_per_epoch": 10000,
|
| 25 |
+
"n_epochs": 30,
|
| 26 |
+
"batch_size": 8,
|
| 27 |
+
"balance_loss": false,
|
| 28 |
+
"loss_weights_span": [
|
| 29 |
+
[
|
| 30 |
+
1.0,
|
| 31 |
+
1.0
|
| 32 |
+
],
|
| 33 |
+
[
|
| 34 |
+
1.0,
|
| 35 |
+
50.0
|
| 36 |
+
],
|
| 37 |
+
[
|
| 38 |
+
1.0,
|
| 39 |
+
100.0
|
| 40 |
+
]
|
| 41 |
+
],
|
| 42 |
+
"time": 1727765390.5829365,
|
| 43 |
+
"config_dataset": {
|
| 44 |
+
"generation_kwargs": {
|
| 45 |
+
"max_new_tokens": 100,
|
| 46 |
+
"repetition_penalty": 1.2
|
| 47 |
+
},
|
| 48 |
+
"model_id": "meta-llama/Llama-3.2-1B",
|
| 49 |
+
"flair_model_name": "flair/ner-english-ontonotes-large"
|
| 50 |
+
}
|
| 51 |
+
}
|
data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/pbK46jjAVx/checkpoint.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f32816959f5fd27967c754a61b07d8ae6c92b7881e2fbb6a68b54b8c0c575122
|
| 3 |
+
size 33884328
|
data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/pbK46jjAVx/config.json
ADDED
|
@@ -0,0 +1,166 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"layer": 10,
|
| 3 |
+
"model": "meta-llama/Llama-3.2-1B",
|
| 4 |
+
"type": "token_classifier",
|
| 5 |
+
"label_map": [
|
| 6 |
+
"O",
|
| 7 |
+
"CARDINAL",
|
| 8 |
+
"DATE",
|
| 9 |
+
"EVENT",
|
| 10 |
+
"FAC",
|
| 11 |
+
"GPE",
|
| 12 |
+
"LANGUAGE",
|
| 13 |
+
"LAW",
|
| 14 |
+
"LOC",
|
| 15 |
+
"MONEY",
|
| 16 |
+
"NORP",
|
| 17 |
+
"ORDINAL",
|
| 18 |
+
"ORG",
|
| 19 |
+
"PERCENT",
|
| 20 |
+
"PERSON",
|
| 21 |
+
"PRODUCT",
|
| 22 |
+
"QUANTITY",
|
| 23 |
+
"TIME",
|
| 24 |
+
"WORK_OF_ART"
|
| 25 |
+
],
|
| 26 |
+
"learning_rate": 0.0003,
|
| 27 |
+
"classifier_dim": 4096,
|
| 28 |
+
"loss_weights": [
|
| 29 |
+
1.0,
|
| 30 |
+
1.0,
|
| 31 |
+
1.0,
|
| 32 |
+
1.0,
|
| 33 |
+
1.0,
|
| 34 |
+
1.0,
|
| 35 |
+
1.0,
|
| 36 |
+
1.0,
|
| 37 |
+
1.0,
|
| 38 |
+
1.0,
|
| 39 |
+
1.0,
|
| 40 |
+
1.0,
|
| 41 |
+
1.0,
|
| 42 |
+
1.0,
|
| 43 |
+
1.0,
|
| 44 |
+
1.0,
|
| 45 |
+
1.0,
|
| 46 |
+
1.0,
|
| 47 |
+
1.0
|
| 48 |
+
],
|
| 49 |
+
"identifier": "pbK46jjAVx",
|
| 50 |
+
"best_f1_validation": 0.9048610329627991,
|
| 51 |
+
"best_f1_validation_classwise": {
|
| 52 |
+
"CARDINAL": {
|
| 53 |
+
"p": 0.8730558156967163,
|
| 54 |
+
"r": 0.8727306723594666,
|
| 55 |
+
"f": 0.872893214225769,
|
| 56 |
+
"s": 10741.0
|
| 57 |
+
},
|
| 58 |
+
"DATE": {
|
| 59 |
+
"p": 0.9534441828727722,
|
| 60 |
+
"r": 0.9365375638008118,
|
| 61 |
+
"f": 0.944915235042572,
|
| 62 |
+
"s": 8572.0
|
| 63 |
+
},
|
| 64 |
+
"EVENT": {
|
| 65 |
+
"p": 0.8540268540382385,
|
| 66 |
+
"r": 0.83442622423172,
|
| 67 |
+
"f": 0.844112753868103,
|
| 68 |
+
"s": 1220.0
|
| 69 |
+
},
|
| 70 |
+
"FAC": {
|
| 71 |
+
"p": 0.8227027058601379,
|
| 72 |
+
"r": 0.8165236115455627,
|
| 73 |
+
"f": 0.8196015357971191,
|
| 74 |
+
"s": 932.0
|
| 75 |
+
},
|
| 76 |
+
"GPE": {
|
| 77 |
+
"p": 0.9014912247657776,
|
| 78 |
+
"r": 0.9065608978271484,
|
| 79 |
+
"f": 0.9040189981460571,
|
| 80 |
+
"s": 6935.0
|
| 81 |
+
},
|
| 82 |
+
"LANGUAGE": {
|
| 83 |
+
"p": 0.7272727489471436,
|
| 84 |
+
"r": 0.6399999856948853,
|
| 85 |
+
"f": 0.6808510422706604,
|
| 86 |
+
"s": 25.0
|
| 87 |
+
},
|
| 88 |
+
"LAW": {
|
| 89 |
+
"p": 0.8500000238418579,
|
| 90 |
+
"r": 0.73046875,
|
| 91 |
+
"f": 0.7857142686843872,
|
| 92 |
+
"s": 256.0
|
| 93 |
+
},
|
| 94 |
+
"LOC": {
|
| 95 |
+
"p": 0.8867924809455872,
|
| 96 |
+
"r": 0.6811594367027283,
|
| 97 |
+
"f": 0.7704918384552002,
|
| 98 |
+
"s": 414.0
|
| 99 |
+
},
|
| 100 |
+
"MONEY": {
|
| 101 |
+
"p": 0.873665452003479,
|
| 102 |
+
"r": 0.8644366264343262,
|
| 103 |
+
"f": 0.8690265417098999,
|
| 104 |
+
"s": 1704.0
|
| 105 |
+
},
|
| 106 |
+
"NORP": {
|
| 107 |
+
"p": 0.9220505356788635,
|
| 108 |
+
"r": 0.875333309173584,
|
| 109 |
+
"f": 0.898084819316864,
|
| 110 |
+
"s": 1500.0
|
| 111 |
+
},
|
| 112 |
+
"ORDINAL": {
|
| 113 |
+
"p": 0.9244186282157898,
|
| 114 |
+
"r": 0.9559118151664734,
|
| 115 |
+
"f": 0.9399014711380005,
|
| 116 |
+
"s": 998.0
|
| 117 |
+
},
|
| 118 |
+
"ORG": {
|
| 119 |
+
"p": 0.8920637965202332,
|
| 120 |
+
"r": 0.8841343522071838,
|
| 121 |
+
"f": 0.888081431388855,
|
| 122 |
+
"s": 9675.0
|
| 123 |
+
},
|
| 124 |
+
"PERCENT": {
|
| 125 |
+
"p": 0.8530852198600769,
|
| 126 |
+
"r": 0.8102325797080994,
|
| 127 |
+
"f": 0.8311069011688232,
|
| 128 |
+
"s": 1075.0
|
| 129 |
+
},
|
| 130 |
+
"PERSON": {
|
| 131 |
+
"p": 0.9692212343215942,
|
| 132 |
+
"r": 0.9716256856918335,
|
| 133 |
+
"f": 0.9704219698905945,
|
| 134 |
+
"s": 12899.0
|
| 135 |
+
},
|
| 136 |
+
"PRODUCT": {
|
| 137 |
+
"p": 0.7886179089546204,
|
| 138 |
+
"r": 0.7538859844207764,
|
| 139 |
+
"f": 0.7708609104156494,
|
| 140 |
+
"s": 386.0
|
| 141 |
+
},
|
| 142 |
+
"QUANTITY": {
|
| 143 |
+
"p": 0.8215258717536926,
|
| 144 |
+
"r": 0.7903014421463013,
|
| 145 |
+
"f": 0.8056111931800842,
|
| 146 |
+
"s": 763.0
|
| 147 |
+
},
|
| 148 |
+
"TIME": {
|
| 149 |
+
"p": 0.8752886652946472,
|
| 150 |
+
"r": 0.8275108933448792,
|
| 151 |
+
"f": 0.8507295250892639,
|
| 152 |
+
"s": 1832.0
|
| 153 |
+
},
|
| 154 |
+
"WORK_OF_ART": {
|
| 155 |
+
"p": 0.7937743067741394,
|
| 156 |
+
"r": 0.7083333134651184,
|
| 157 |
+
"f": 0.7486238479614258,
|
| 158 |
+
"s": 576.0
|
| 159 |
+
},
|
| 160 |
+
"macro": {
|
| 161 |
+
"p": 0.8656943440437317,
|
| 162 |
+
"r": 0.8255618214607239,
|
| 163 |
+
"f": 0.8441693186759949
|
| 164 |
+
}
|
| 165 |
+
}
|
| 166 |
+
}
|
data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/pbK46jjAVx/config_train.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"path": "data/meta-llama/Llama-3.2-1B/STOKE_100",
|
| 3 |
+
"splits": [
|
| 4 |
+
"train",
|
| 5 |
+
"validation"
|
| 6 |
+
],
|
| 7 |
+
"layers": [
|
| 8 |
+
8,
|
| 9 |
+
9,
|
| 10 |
+
10,
|
| 11 |
+
11,
|
| 12 |
+
12
|
| 13 |
+
],
|
| 14 |
+
"hfcache": "",
|
| 15 |
+
"classifier_dims": [
|
| 16 |
+
4096
|
| 17 |
+
],
|
| 18 |
+
"learning_rates": [
|
| 19 |
+
0.0001,
|
| 20 |
+
5e-05,
|
| 21 |
+
0.0003
|
| 22 |
+
],
|
| 23 |
+
"cuda": true,
|
| 24 |
+
"n_steps_per_epoch": 10000,
|
| 25 |
+
"n_epochs": 30,
|
| 26 |
+
"batch_size": 8,
|
| 27 |
+
"balance_loss": false,
|
| 28 |
+
"loss_weights_span": [
|
| 29 |
+
[
|
| 30 |
+
1.0,
|
| 31 |
+
1.0
|
| 32 |
+
],
|
| 33 |
+
[
|
| 34 |
+
1.0,
|
| 35 |
+
50.0
|
| 36 |
+
],
|
| 37 |
+
[
|
| 38 |
+
1.0,
|
| 39 |
+
100.0
|
| 40 |
+
]
|
| 41 |
+
],
|
| 42 |
+
"time": 1727765390.5829365,
|
| 43 |
+
"config_dataset": {
|
| 44 |
+
"generation_kwargs": {
|
| 45 |
+
"max_new_tokens": 100,
|
| 46 |
+
"repetition_penalty": 1.2
|
| 47 |
+
},
|
| 48 |
+
"model_id": "meta-llama/Llama-3.2-1B",
|
| 49 |
+
"flair_model_name": "flair/ner-english-ontonotes-large"
|
| 50 |
+
}
|
| 51 |
+
}
|
data/meta-llama/Llama-3.2-1B/STOKE_100/config.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"generation_kwargs": {
|
| 3 |
+
"max_new_tokens": 100,
|
| 4 |
+
"repetition_penalty": 1.2
|
| 5 |
+
},
|
| 6 |
+
"model_id": "meta-llama/Llama-3.2-1B",
|
| 7 |
+
"flair_model_name": "flair/ner-english-ontonotes-large"
|
| 8 |
+
}
|
data/meta-llama/Llama-3.2-1B/STOKE_100/stoke_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"default": {
|
| 3 |
+
"classifier_token": "data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/pbK46jjAVx",
|
| 4 |
+
"classifier_span": "data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/span_classifier/Rxi8b70XJA"
|
| 5 |
+
},
|
| 6 |
+
"basic": {
|
| 7 |
+
"classifier_token": "data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/token_classifier/dR8xQB4ODU",
|
| 8 |
+
"classifier_span": "data/meta-llama/Llama-3.2-1B/STOKE_100/checkpoints/span_classifier/Rxi8b70XJA"
|
| 9 |
+
}
|
| 10 |
+
}
|