Transformers
Safetensors
English
scientific-claim-verification
scifact
evidence-selection
retrieved-evidence
Instructions to use rishhh/verisci-claim-verifier-constrained-selector-joint-seed123 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use rishhh/verisci-claim-verifier-constrained-selector-joint-seed123 with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("rishhh/verisci-claim-verifier-constrained-selector-joint-seed123", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Upload learned selector joint verifier candidate
Browse files- README.md +28 -0
- evaluation/constrained_selector_joint_eval_report.md +40 -0
- evaluation/constrained_selector_joint_eval_results.json +378 -0
- evaluation/constrained_selector_joint_predictions.jsonl +0 -0
- selector/added_tokens.json +3 -0
- selector/config.json +45 -0
- selector/model.safetensors +3 -0
- selector/special_tokens_map.json +15 -0
- selector/spm.model +3 -0
- selector/tokenizer.json +0 -0
- selector/tokenizer_config.json +59 -0
- selector/training_args.bin +3 -0
- verifier/added_tokens.json +3 -0
- verifier/config.json +47 -0
- verifier/model.safetensors +3 -0
- verifier/special_tokens_map.json +51 -0
- verifier/spm.model +3 -0
- verifier/tokenizer.json +0 -0
- verifier/tokenizer_config.json +63 -0
- verifier/training_args.bin +3 -0
README.md
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language: en
|
| 3 |
+
library_name: transformers
|
| 4 |
+
license: cc-by-nc-2.0
|
| 5 |
+
tags:
|
| 6 |
+
- scientific-claim-verification
|
| 7 |
+
- scifact
|
| 8 |
+
- evidence-selection
|
| 9 |
+
- retrieved-evidence
|
| 10 |
+
metrics:
|
| 11 |
+
- accuracy
|
| 12 |
+
- f1
|
| 13 |
+
---
|
| 14 |
+
|
| 15 |
+
# VeriSci Constrained Selector Joint Verifier
|
| 16 |
+
|
| 17 |
+
Experimental candidate with train-dev constrained decision calibration.
|
| 18 |
+
|
| 19 |
+
- Selector base: `microsoft/deberta-v3-small`
|
| 20 |
+
- Verifier base: `rishhh/verisci-claim-verifier-retrieval-adapted-seed123`
|
| 21 |
+
- Raw SciFact validation accuracy: 0.7000
|
| 22 |
+
- Raw SciFact validation macro F1: 0.6614
|
| 23 |
+
- Thresholded SciFact validation accuracy: 0.7000
|
| 24 |
+
- Thresholded SciFact validation macro F1: 0.6614
|
| 25 |
+
- Promotion gate passed: `False`
|
| 26 |
+
- Promotion gate reason: `candidate_accuracy_below_target: 0.700 < 0.910`
|
| 27 |
+
|
| 28 |
+
This artifact is experimental unless the repository documentation marks it as promoted.
|
evaluation/constrained_selector_joint_eval_report.md
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Constrained Selector Joint Verifier Evaluation
|
| 2 |
+
|
| 3 |
+
- Model repo: `rishhh/verisci-claim-verifier-constrained-selector-joint-seed123`
|
| 4 |
+
- Selector base: `microsoft/deberta-v3-small`
|
| 5 |
+
- Verifier base: `rishhh/verisci-claim-verifier-retrieval-adapted-seed123`
|
| 6 |
+
- Retriever: `rishhh/verisci-scifact-e5-retriever`
|
| 7 |
+
|
| 8 |
+
## SciFact Validation
|
| 9 |
+
|
| 10 |
+
- Raw original-label accuracy: 0.700
|
| 11 |
+
- Raw original-label macro F1: 0.661
|
| 12 |
+
- Thresholded original-label accuracy: 0.700
|
| 13 |
+
- Thresholded original-label macro F1: 0.661
|
| 14 |
+
- Raw retrieved-evidence-label accuracy: 0.736
|
| 15 |
+
- Raw retrieved-evidence-label macro F1: 0.709
|
| 16 |
+
- Thresholded retrieved-evidence-label accuracy: 0.736
|
| 17 |
+
- Thresholded retrieved-evidence-label macro F1: 0.709
|
| 18 |
+
- Retrieval oracle accuracy: 0.929
|
| 19 |
+
- Promotion gate passed: `False`
|
| 20 |
+
- Promotion gate reason: `candidate_accuracy_below_target: 0.700 < 0.910`
|
| 21 |
+
|
| 22 |
+
## Decision Policy
|
| 23 |
+
|
| 24 |
+
```json
|
| 25 |
+
{
|
| 26 |
+
"strategy": "argmax_fallback_no_policy_met_nei_recall_floor",
|
| 27 |
+
"min_nei_recall": 0.74
|
| 28 |
+
}
|
| 29 |
+
```
|
| 30 |
+
|
| 31 |
+
## Selection Protocol
|
| 32 |
+
|
| 33 |
+
- Scope: SciFact train only; train/dev split by claim id
|
| 34 |
+
- Validation used for model selection: `False`
|
| 35 |
+
|
| 36 |
+
## Limitations
|
| 37 |
+
|
| 38 |
+
- Selector supervision uses SciFact evidence sentence IDs when the evidence document is retrieved.
|
| 39 |
+
- SciFact validation is public and is not a hidden benchmark.
|
| 40 |
+
- This is still limited by retrieval recall and by the small SciFact train set.
|
evaluation/constrained_selector_joint_eval_results.json
ADDED
|
@@ -0,0 +1,378 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"dataset": "allenai/scifact",
|
| 3 |
+
"decision_policy": {
|
| 4 |
+
"min_nei_recall": 0.74,
|
| 5 |
+
"strategy": "argmax_fallback_no_policy_met_nei_recall_floor"
|
| 6 |
+
},
|
| 7 |
+
"evaluation_scope": "constrained_selector_joint_verifier_validation",
|
| 8 |
+
"hub_model_id": "rishhh/verisci-claim-verifier-constrained-selector-joint-seed123",
|
| 9 |
+
"hybrid_alpha": 0.75,
|
| 10 |
+
"latency_seconds": 789.2572160660002,
|
| 11 |
+
"limitations": [
|
| 12 |
+
"Selector supervision uses SciFact evidence sentence IDs when the evidence document is retrieved.",
|
| 13 |
+
"SciFact validation is public and is not a hidden benchmark.",
|
| 14 |
+
"This is still limited by retrieval recall and by the small SciFact train set."
|
| 15 |
+
],
|
| 16 |
+
"max_selector_sentences_per_doc": 5,
|
| 17 |
+
"max_snippets": 6,
|
| 18 |
+
"max_snippets_per_doc": 2,
|
| 19 |
+
"min_dev_nei_recall": 0.74,
|
| 20 |
+
"promotion_gate": {
|
| 21 |
+
"baseline_accuracy": 0.691,
|
| 22 |
+
"candidate_accuracy": 0.7,
|
| 23 |
+
"candidate_macro_f1": 0.661419819837583,
|
| 24 |
+
"min_accuracy_delta": 0.01,
|
| 25 |
+
"passed": false,
|
| 26 |
+
"reason": "candidate_accuracy_below_target: 0.700 < 0.910",
|
| 27 |
+
"target_accuracy": 0.91
|
| 28 |
+
},
|
| 29 |
+
"retrieval_oracle": {
|
| 30 |
+
"accuracy": 0.9288888888888889,
|
| 31 |
+
"confusion_matrix": {
|
| 32 |
+
"NOT_ENOUGH_INFO": {
|
| 33 |
+
"NOT_ENOUGH_INFO": 112,
|
| 34 |
+
"REFUTES": 0,
|
| 35 |
+
"SUPPORTS": 0
|
| 36 |
+
},
|
| 37 |
+
"REFUTES": {
|
| 38 |
+
"NOT_ENOUGH_INFO": 20,
|
| 39 |
+
"REFUTES": 102,
|
| 40 |
+
"SUPPORTS": 0
|
| 41 |
+
},
|
| 42 |
+
"SUPPORTS": {
|
| 43 |
+
"NOT_ENOUGH_INFO": 12,
|
| 44 |
+
"REFUTES": 0,
|
| 45 |
+
"SUPPORTS": 204
|
| 46 |
+
}
|
| 47 |
+
},
|
| 48 |
+
"examples": 450,
|
| 49 |
+
"gold_label_distribution": {
|
| 50 |
+
"NOT_ENOUGH_INFO": 112,
|
| 51 |
+
"REFUTES": 122,
|
| 52 |
+
"SUPPORTS": 216
|
| 53 |
+
},
|
| 54 |
+
"macro_f1": 0.919047619047619,
|
| 55 |
+
"per_class": {
|
| 56 |
+
"NOT_ENOUGH_INFO": {
|
| 57 |
+
"f1-score": 0.875,
|
| 58 |
+
"precision": 0.7777777777777778,
|
| 59 |
+
"recall": 1.0,
|
| 60 |
+
"support": 112.0
|
| 61 |
+
},
|
| 62 |
+
"REFUTES": {
|
| 63 |
+
"f1-score": 0.9107142857142857,
|
| 64 |
+
"precision": 1.0,
|
| 65 |
+
"recall": 0.8360655737704918,
|
| 66 |
+
"support": 122.0
|
| 67 |
+
},
|
| 68 |
+
"SUPPORTS": {
|
| 69 |
+
"f1-score": 0.9714285714285714,
|
| 70 |
+
"precision": 1.0,
|
| 71 |
+
"recall": 0.9444444444444444,
|
| 72 |
+
"support": 216.0
|
| 73 |
+
},
|
| 74 |
+
"accuracy": 0.9288888888888889,
|
| 75 |
+
"macro avg": {
|
| 76 |
+
"f1-score": 0.919047619047619,
|
| 77 |
+
"precision": 0.9259259259259259,
|
| 78 |
+
"recall": 0.926836672738312,
|
| 79 |
+
"support": 450.0
|
| 80 |
+
},
|
| 81 |
+
"weighted avg": {
|
| 82 |
+
"f1-score": 0.930968253968254,
|
| 83 |
+
"precision": 0.9446913580246913,
|
| 84 |
+
"recall": 0.9288888888888889,
|
| 85 |
+
"support": 450.0
|
| 86 |
+
}
|
| 87 |
+
},
|
| 88 |
+
"predicted_label_distribution": {
|
| 89 |
+
"NOT_ENOUGH_INFO": 144,
|
| 90 |
+
"REFUTES": 102,
|
| 91 |
+
"SUPPORTS": 204
|
| 92 |
+
},
|
| 93 |
+
"strategy": "retrieval_oracle_nei_when_missing"
|
| 94 |
+
},
|
| 95 |
+
"retriever_id": "rishhh/verisci-scifact-e5-retriever",
|
| 96 |
+
"seed": 123,
|
| 97 |
+
"selection_protocol": {
|
| 98 |
+
"scope": "SciFact train only; train/dev split by claim id",
|
| 99 |
+
"selector_selection_metric": "best epoch by SciFact train-dev macro F1",
|
| 100 |
+
"threshold_selection_metric": "best train-dev accuracy with NEI recall floor",
|
| 101 |
+
"validation_used_for_model_selection": false,
|
| 102 |
+
"validation_used_for_training_examples": false,
|
| 103 |
+
"verifier_selection_metric": "best epoch by SciFact train-dev macro F1"
|
| 104 |
+
},
|
| 105 |
+
"selector_base_model_id": "microsoft/deberta-v3-small",
|
| 106 |
+
"selector_dev_examples": 13506,
|
| 107 |
+
"selector_train_examples": 17733,
|
| 108 |
+
"split": "validation",
|
| 109 |
+
"top_k": 5,
|
| 110 |
+
"validation": {
|
| 111 |
+
"accuracy": 0.7,
|
| 112 |
+
"confusion_matrix": {
|
| 113 |
+
"NOT_ENOUGH_INFO": {
|
| 114 |
+
"NOT_ENOUGH_INFO": 85,
|
| 115 |
+
"REFUTES": 7,
|
| 116 |
+
"SUPPORTS": 20
|
| 117 |
+
},
|
| 118 |
+
"REFUTES": {
|
| 119 |
+
"NOT_ENOUGH_INFO": 50,
|
| 120 |
+
"REFUTES": 52,
|
| 121 |
+
"SUPPORTS": 20
|
| 122 |
+
},
|
| 123 |
+
"SUPPORTS": {
|
| 124 |
+
"NOT_ENOUGH_INFO": 30,
|
| 125 |
+
"REFUTES": 8,
|
| 126 |
+
"SUPPORTS": 178
|
| 127 |
+
}
|
| 128 |
+
},
|
| 129 |
+
"examples": 450,
|
| 130 |
+
"gold_label_distribution": {
|
| 131 |
+
"NOT_ENOUGH_INFO": 112,
|
| 132 |
+
"REFUTES": 122,
|
| 133 |
+
"SUPPORTS": 216
|
| 134 |
+
},
|
| 135 |
+
"macro_f1": 0.661419819837583,
|
| 136 |
+
"per_class": {
|
| 137 |
+
"NOT_ENOUGH_INFO": {
|
| 138 |
+
"f1-score": 0.6137184115523465,
|
| 139 |
+
"precision": 0.5151515151515151,
|
| 140 |
+
"recall": 0.7589285714285714,
|
| 141 |
+
"support": 112.0
|
| 142 |
+
},
|
| 143 |
+
"REFUTES": {
|
| 144 |
+
"f1-score": 0.5502645502645502,
|
| 145 |
+
"precision": 0.7761194029850746,
|
| 146 |
+
"recall": 0.4262295081967213,
|
| 147 |
+
"support": 122.0
|
| 148 |
+
},
|
| 149 |
+
"SUPPORTS": {
|
| 150 |
+
"f1-score": 0.8202764976958525,
|
| 151 |
+
"precision": 0.8165137614678899,
|
| 152 |
+
"recall": 0.8240740740740741,
|
| 153 |
+
"support": 216.0
|
| 154 |
+
},
|
| 155 |
+
"accuracy": 0.7,
|
| 156 |
+
"macro avg": {
|
| 157 |
+
"f1-score": 0.661419819837583,
|
| 158 |
+
"precision": 0.7025948932014933,
|
| 159 |
+
"recall": 0.6697440512331223,
|
| 160 |
+
"support": 450.0
|
| 161 |
+
},
|
| 162 |
+
"weighted avg": {
|
| 163 |
+
"f1-score": 0.6956632460632046,
|
| 164 |
+
"precision": 0.7305566874182512,
|
| 165 |
+
"recall": 0.7,
|
| 166 |
+
"support": 450.0
|
| 167 |
+
}
|
| 168 |
+
},
|
| 169 |
+
"predicted_label_distribution": {
|
| 170 |
+
"NOT_ENOUGH_INFO": 165,
|
| 171 |
+
"REFUTES": 67,
|
| 172 |
+
"SUPPORTS": 218
|
| 173 |
+
},
|
| 174 |
+
"strategy": "scifact_validation_original_labels"
|
| 175 |
+
},
|
| 176 |
+
"validation_examples": 450,
|
| 177 |
+
"validation_retrieved_evidence_label_eval": {
|
| 178 |
+
"accuracy": 0.7355555555555555,
|
| 179 |
+
"confusion_matrix": {
|
| 180 |
+
"NOT_ENOUGH_INFO": {
|
| 181 |
+
"NOT_ENOUGH_INFO": 109,
|
| 182 |
+
"REFUTES": 7,
|
| 183 |
+
"SUPPORTS": 28
|
| 184 |
+
},
|
| 185 |
+
"REFUTES": {
|
| 186 |
+
"NOT_ENOUGH_INFO": 30,
|
| 187 |
+
"REFUTES": 52,
|
| 188 |
+
"SUPPORTS": 20
|
| 189 |
+
},
|
| 190 |
+
"SUPPORTS": {
|
| 191 |
+
"NOT_ENOUGH_INFO": 26,
|
| 192 |
+
"REFUTES": 8,
|
| 193 |
+
"SUPPORTS": 170
|
| 194 |
+
}
|
| 195 |
+
},
|
| 196 |
+
"examples": 450,
|
| 197 |
+
"gold_label_distribution": {
|
| 198 |
+
"NOT_ENOUGH_INFO": 144,
|
| 199 |
+
"REFUTES": 102,
|
| 200 |
+
"SUPPORTS": 204
|
| 201 |
+
},
|
| 202 |
+
"macro_f1": 0.7088578124330206,
|
| 203 |
+
"per_class": {
|
| 204 |
+
"NOT_ENOUGH_INFO": {
|
| 205 |
+
"f1-score": 0.7055016181229773,
|
| 206 |
+
"precision": 0.6606060606060606,
|
| 207 |
+
"recall": 0.7569444444444444,
|
| 208 |
+
"support": 144.0
|
| 209 |
+
},
|
| 210 |
+
"REFUTES": {
|
| 211 |
+
"f1-score": 0.6153846153846154,
|
| 212 |
+
"precision": 0.7761194029850746,
|
| 213 |
+
"recall": 0.5098039215686274,
|
| 214 |
+
"support": 102.0
|
| 215 |
+
},
|
| 216 |
+
"SUPPORTS": {
|
| 217 |
+
"f1-score": 0.8056872037914692,
|
| 218 |
+
"precision": 0.7798165137614679,
|
| 219 |
+
"recall": 0.8333333333333334,
|
| 220 |
+
"support": 204.0
|
| 221 |
+
},
|
| 222 |
+
"accuracy": 0.7355555555555555,
|
| 223 |
+
"macro avg": {
|
| 224 |
+
"f1-score": 0.7088578124330206,
|
| 225 |
+
"precision": 0.7388473257842011,
|
| 226 |
+
"recall": 0.7000272331154683,
|
| 227 |
+
"support": 450.0
|
| 228 |
+
},
|
| 229 |
+
"weighted avg": {
|
| 230 |
+
"f1-score": 0.7304925630053316,
|
| 231 |
+
"precision": 0.7408311569757551,
|
| 232 |
+
"recall": 0.7355555555555555,
|
| 233 |
+
"support": 450.0
|
| 234 |
+
}
|
| 235 |
+
},
|
| 236 |
+
"predicted_label_distribution": {
|
| 237 |
+
"NOT_ENOUGH_INFO": 165,
|
| 238 |
+
"REFUTES": 67,
|
| 239 |
+
"SUPPORTS": 218
|
| 240 |
+
},
|
| 241 |
+
"strategy": "scifact_validation_retrieved_evidence_labels"
|
| 242 |
+
},
|
| 243 |
+
"validation_retrieved_evidence_label_eval_thresholded": {
|
| 244 |
+
"accuracy": 0.7355555555555555,
|
| 245 |
+
"confusion_matrix": {
|
| 246 |
+
"NOT_ENOUGH_INFO": {
|
| 247 |
+
"NOT_ENOUGH_INFO": 109,
|
| 248 |
+
"REFUTES": 7,
|
| 249 |
+
"SUPPORTS": 28
|
| 250 |
+
},
|
| 251 |
+
"REFUTES": {
|
| 252 |
+
"NOT_ENOUGH_INFO": 30,
|
| 253 |
+
"REFUTES": 52,
|
| 254 |
+
"SUPPORTS": 20
|
| 255 |
+
},
|
| 256 |
+
"SUPPORTS": {
|
| 257 |
+
"NOT_ENOUGH_INFO": 26,
|
| 258 |
+
"REFUTES": 8,
|
| 259 |
+
"SUPPORTS": 170
|
| 260 |
+
}
|
| 261 |
+
},
|
| 262 |
+
"examples": 450,
|
| 263 |
+
"gold_label_distribution": {
|
| 264 |
+
"NOT_ENOUGH_INFO": 144,
|
| 265 |
+
"REFUTES": 102,
|
| 266 |
+
"SUPPORTS": 204
|
| 267 |
+
},
|
| 268 |
+
"macro_f1": 0.7088578124330206,
|
| 269 |
+
"per_class": {
|
| 270 |
+
"NOT_ENOUGH_INFO": {
|
| 271 |
+
"f1-score": 0.7055016181229773,
|
| 272 |
+
"precision": 0.6606060606060606,
|
| 273 |
+
"recall": 0.7569444444444444,
|
| 274 |
+
"support": 144.0
|
| 275 |
+
},
|
| 276 |
+
"REFUTES": {
|
| 277 |
+
"f1-score": 0.6153846153846154,
|
| 278 |
+
"precision": 0.7761194029850746,
|
| 279 |
+
"recall": 0.5098039215686274,
|
| 280 |
+
"support": 102.0
|
| 281 |
+
},
|
| 282 |
+
"SUPPORTS": {
|
| 283 |
+
"f1-score": 0.8056872037914692,
|
| 284 |
+
"precision": 0.7798165137614679,
|
| 285 |
+
"recall": 0.8333333333333334,
|
| 286 |
+
"support": 204.0
|
| 287 |
+
},
|
| 288 |
+
"accuracy": 0.7355555555555555,
|
| 289 |
+
"macro avg": {
|
| 290 |
+
"f1-score": 0.7088578124330206,
|
| 291 |
+
"precision": 0.7388473257842011,
|
| 292 |
+
"recall": 0.7000272331154683,
|
| 293 |
+
"support": 450.0
|
| 294 |
+
},
|
| 295 |
+
"weighted avg": {
|
| 296 |
+
"f1-score": 0.7304925630053316,
|
| 297 |
+
"precision": 0.7408311569757551,
|
| 298 |
+
"recall": 0.7355555555555555,
|
| 299 |
+
"support": 450.0
|
| 300 |
+
}
|
| 301 |
+
},
|
| 302 |
+
"predicted_label_distribution": {
|
| 303 |
+
"NOT_ENOUGH_INFO": 165,
|
| 304 |
+
"REFUTES": 67,
|
| 305 |
+
"SUPPORTS": 218
|
| 306 |
+
},
|
| 307 |
+
"strategy": "scifact_validation_retrieved_evidence_labels_thresholded"
|
| 308 |
+
},
|
| 309 |
+
"validation_thresholded": {
|
| 310 |
+
"accuracy": 0.7,
|
| 311 |
+
"confusion_matrix": {
|
| 312 |
+
"NOT_ENOUGH_INFO": {
|
| 313 |
+
"NOT_ENOUGH_INFO": 85,
|
| 314 |
+
"REFUTES": 7,
|
| 315 |
+
"SUPPORTS": 20
|
| 316 |
+
},
|
| 317 |
+
"REFUTES": {
|
| 318 |
+
"NOT_ENOUGH_INFO": 50,
|
| 319 |
+
"REFUTES": 52,
|
| 320 |
+
"SUPPORTS": 20
|
| 321 |
+
},
|
| 322 |
+
"SUPPORTS": {
|
| 323 |
+
"NOT_ENOUGH_INFO": 30,
|
| 324 |
+
"REFUTES": 8,
|
| 325 |
+
"SUPPORTS": 178
|
| 326 |
+
}
|
| 327 |
+
},
|
| 328 |
+
"examples": 450,
|
| 329 |
+
"gold_label_distribution": {
|
| 330 |
+
"NOT_ENOUGH_INFO": 112,
|
| 331 |
+
"REFUTES": 122,
|
| 332 |
+
"SUPPORTS": 216
|
| 333 |
+
},
|
| 334 |
+
"macro_f1": 0.661419819837583,
|
| 335 |
+
"per_class": {
|
| 336 |
+
"NOT_ENOUGH_INFO": {
|
| 337 |
+
"f1-score": 0.6137184115523465,
|
| 338 |
+
"precision": 0.5151515151515151,
|
| 339 |
+
"recall": 0.7589285714285714,
|
| 340 |
+
"support": 112.0
|
| 341 |
+
},
|
| 342 |
+
"REFUTES": {
|
| 343 |
+
"f1-score": 0.5502645502645502,
|
| 344 |
+
"precision": 0.7761194029850746,
|
| 345 |
+
"recall": 0.4262295081967213,
|
| 346 |
+
"support": 122.0
|
| 347 |
+
},
|
| 348 |
+
"SUPPORTS": {
|
| 349 |
+
"f1-score": 0.8202764976958525,
|
| 350 |
+
"precision": 0.8165137614678899,
|
| 351 |
+
"recall": 0.8240740740740741,
|
| 352 |
+
"support": 216.0
|
| 353 |
+
},
|
| 354 |
+
"accuracy": 0.7,
|
| 355 |
+
"macro avg": {
|
| 356 |
+
"f1-score": 0.661419819837583,
|
| 357 |
+
"precision": 0.7025948932014933,
|
| 358 |
+
"recall": 0.6697440512331223,
|
| 359 |
+
"support": 450.0
|
| 360 |
+
},
|
| 361 |
+
"weighted avg": {
|
| 362 |
+
"f1-score": 0.6956632460632046,
|
| 363 |
+
"precision": 0.7305566874182512,
|
| 364 |
+
"recall": 0.7,
|
| 365 |
+
"support": 450.0
|
| 366 |
+
}
|
| 367 |
+
},
|
| 368 |
+
"predicted_label_distribution": {
|
| 369 |
+
"NOT_ENOUGH_INFO": 165,
|
| 370 |
+
"REFUTES": 67,
|
| 371 |
+
"SUPPORTS": 218
|
| 372 |
+
},
|
| 373 |
+
"strategy": "scifact_validation_original_labels_thresholded"
|
| 374 |
+
},
|
| 375 |
+
"verifier_base_model_id": "rishhh/verisci-claim-verifier-retrieval-adapted-seed123",
|
| 376 |
+
"verifier_dev_examples": 545,
|
| 377 |
+
"verifier_train_examples": 716
|
| 378 |
+
}
|
evaluation/constrained_selector_joint_predictions.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
selector/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"[MASK]": 128000
|
| 3 |
+
}
|
selector/config.json
ADDED
|
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"DebertaV2ForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"bos_token_id": 1,
|
| 7 |
+
"dtype": "float32",
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "NOT_EVIDENCE",
|
| 14 |
+
"1": "EVIDENCE"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"EVIDENCE": 1,
|
| 20 |
+
"NOT_EVIDENCE": 0
|
| 21 |
+
},
|
| 22 |
+
"layer_norm_eps": 1e-07,
|
| 23 |
+
"legacy": true,
|
| 24 |
+
"max_position_embeddings": 512,
|
| 25 |
+
"max_relative_positions": -1,
|
| 26 |
+
"model_type": "deberta-v2",
|
| 27 |
+
"norm_rel_ebd": "layer_norm",
|
| 28 |
+
"num_attention_heads": 12,
|
| 29 |
+
"num_hidden_layers": 6,
|
| 30 |
+
"pad_token_id": 0,
|
| 31 |
+
"pooler_dropout": 0,
|
| 32 |
+
"pooler_hidden_act": "gelu",
|
| 33 |
+
"pooler_hidden_size": 768,
|
| 34 |
+
"pos_att_type": [
|
| 35 |
+
"p2c",
|
| 36 |
+
"c2p"
|
| 37 |
+
],
|
| 38 |
+
"position_biased_input": false,
|
| 39 |
+
"position_buckets": 256,
|
| 40 |
+
"relative_attention": true,
|
| 41 |
+
"share_att_key": true,
|
| 42 |
+
"transformers_version": "4.57.6",
|
| 43 |
+
"type_vocab_size": 0,
|
| 44 |
+
"vocab_size": 128100
|
| 45 |
+
}
|
selector/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:26c1661885f3c609ba8786f965f142fd8dab54e24f80feb35a3570e6cec46b9e
|
| 3 |
+
size 567598552
|
selector/special_tokens_map.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "[CLS]",
|
| 3 |
+
"cls_token": "[CLS]",
|
| 4 |
+
"eos_token": "[SEP]",
|
| 5 |
+
"mask_token": "[MASK]",
|
| 6 |
+
"pad_token": "[PAD]",
|
| 7 |
+
"sep_token": "[SEP]",
|
| 8 |
+
"unk_token": {
|
| 9 |
+
"content": "[UNK]",
|
| 10 |
+
"lstrip": false,
|
| 11 |
+
"normalized": true,
|
| 12 |
+
"rstrip": false,
|
| 13 |
+
"single_word": false
|
| 14 |
+
}
|
| 15 |
+
}
|
selector/spm.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
|
| 3 |
+
size 2464616
|
selector/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
selector/tokenizer_config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "[CLS]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[SEP]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[UNK]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": true,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"128000": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "[CLS]",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "[CLS]",
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"eos_token": "[SEP]",
|
| 49 |
+
"extra_special_tokens": {},
|
| 50 |
+
"mask_token": "[MASK]",
|
| 51 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 52 |
+
"pad_token": "[PAD]",
|
| 53 |
+
"sep_token": "[SEP]",
|
| 54 |
+
"sp_model_kwargs": {},
|
| 55 |
+
"split_by_punct": false,
|
| 56 |
+
"tokenizer_class": "DebertaV2Tokenizer",
|
| 57 |
+
"unk_token": "[UNK]",
|
| 58 |
+
"vocab_type": "spm"
|
| 59 |
+
}
|
selector/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:89ae57bc427bc50818c08c9c8a728d64cffdd7911fcd28513cae03b393d05101
|
| 3 |
+
size 5969
|
verifier/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"[MASK]": 128000
|
| 3 |
+
}
|
verifier/config.json
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"DebertaV2ForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"bos_token_id": 1,
|
| 7 |
+
"dtype": "float32",
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "NOT_ENOUGH_INFO",
|
| 14 |
+
"1": "REFUTES",
|
| 15 |
+
"2": "SUPPORTS"
|
| 16 |
+
},
|
| 17 |
+
"initializer_range": 0.02,
|
| 18 |
+
"intermediate_size": 3072,
|
| 19 |
+
"label2id": {
|
| 20 |
+
"NOT_ENOUGH_INFO": 0,
|
| 21 |
+
"REFUTES": 1,
|
| 22 |
+
"SUPPORTS": 2
|
| 23 |
+
},
|
| 24 |
+
"layer_norm_eps": 1e-07,
|
| 25 |
+
"legacy": true,
|
| 26 |
+
"max_position_embeddings": 512,
|
| 27 |
+
"max_relative_positions": -1,
|
| 28 |
+
"model_type": "deberta-v2",
|
| 29 |
+
"norm_rel_ebd": "layer_norm",
|
| 30 |
+
"num_attention_heads": 12,
|
| 31 |
+
"num_hidden_layers": 12,
|
| 32 |
+
"pad_token_id": 0,
|
| 33 |
+
"pooler_dropout": 0,
|
| 34 |
+
"pooler_hidden_act": "gelu",
|
| 35 |
+
"pooler_hidden_size": 768,
|
| 36 |
+
"pos_att_type": [
|
| 37 |
+
"p2c",
|
| 38 |
+
"c2p"
|
| 39 |
+
],
|
| 40 |
+
"position_biased_input": false,
|
| 41 |
+
"position_buckets": 256,
|
| 42 |
+
"relative_attention": true,
|
| 43 |
+
"share_att_key": true,
|
| 44 |
+
"transformers_version": "4.57.6",
|
| 45 |
+
"type_vocab_size": 0,
|
| 46 |
+
"vocab_size": 128100
|
| 47 |
+
}
|
verifier/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:335a7d41970aef66af6669cf5a0e2f62a213254bb1e97f98b2397a4ad846d501
|
| 3 |
+
size 737722356
|
verifier/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "[CLS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "[CLS]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "[SEP]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "[MASK]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "[PAD]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "[SEP]",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "[UNK]",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": true,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
verifier/spm.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
|
| 3 |
+
size 2464616
|
verifier/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
verifier/tokenizer_config.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "[CLS]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[SEP]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[UNK]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": true,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"128000": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "[CLS]",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "[CLS]",
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"eos_token": "[SEP]",
|
| 49 |
+
"extra_special_tokens": {},
|
| 50 |
+
"mask_token": "[MASK]",
|
| 51 |
+
"max_length": 512,
|
| 52 |
+
"model_max_length": 512,
|
| 53 |
+
"pad_token": "[PAD]",
|
| 54 |
+
"sep_token": "[SEP]",
|
| 55 |
+
"sp_model_kwargs": {},
|
| 56 |
+
"split_by_punct": false,
|
| 57 |
+
"stride": 0,
|
| 58 |
+
"tokenizer_class": "DebertaV2Tokenizer",
|
| 59 |
+
"truncation_side": "right",
|
| 60 |
+
"truncation_strategy": "longest_first",
|
| 61 |
+
"unk_token": "[UNK]",
|
| 62 |
+
"vocab_type": "spm"
|
| 63 |
+
}
|
verifier/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:613d97c9ad4ad4c58ca231f8b35060bee5d77c8521c0f3df4bf6203874d21226
|
| 3 |
+
size 5969
|