model update

README.md

@@ -0,0 +1,236 @@
+---
+datasets:
+- relbert/semeval2012_relational_similarity_v6
+model-index:
+- name: relbert/relbert-roberta-base-semeval2012-v6-mask-prompt-c-nce-2-child
+  results:
+  - task:
+      name: Relation Mapping
+      type: sorting-task
+    dataset:
+      name: Relation Mapping
+      args: relbert/relation_mapping
+      type: relation-mapping
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.6354563492063492
+  - task:
+      name: Analogy Questions (SAT full)
+      type: multiple-choice-qa
+    dataset:
+      name: SAT full
+      args: relbert/analogy_questions
+      type: analogy-questions
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.31016042780748665
+  - task:
+      name: Analogy Questions (SAT)
+      type: multiple-choice-qa
+    dataset:
+      name: SAT
+      args: relbert/analogy_questions
+      type: analogy-questions
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.32344213649851633
+  - task:
+      name: Analogy Questions (BATS)
+      type: multiple-choice-qa
+    dataset:
+      name: BATS
+      args: relbert/analogy_questions
+      type: analogy-questions
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.3085047248471373
+  - task:
+      name: Analogy Questions (Google)
+      type: multiple-choice-qa
+    dataset:
+      name: Google
+      args: relbert/analogy_questions
+      type: analogy-questions
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.328
+  - task:
+      name: Analogy Questions (U2)
+      type: multiple-choice-qa
+    dataset:
+      name: U2
+      args: relbert/analogy_questions
+      type: analogy-questions
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.3157894736842105
+  - task:
+      name: Analogy Questions (U4)
+      type: multiple-choice-qa
+    dataset:
+      name: U4
+      args: relbert/analogy_questions
+      type: analogy-questions
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.2986111111111111
+  - task:
+      name: Lexical Relation Classification (BLESS)
+      type: classification
+    dataset:
+      name: BLESS
+      args: relbert/lexical_relation_classification
+      type: relation-classification
+    metrics:
+    - name: F1
+      type: f1
+      value: 0.9019135151423836
+    - name: F1 (macro)
+      type: f1_macro
+      value: 0.8910141280520246
+  - task:
+      name: Lexical Relation Classification (CogALexV)
+      type: classification
+    dataset:
+      name: CogALexV
+      args: relbert/lexical_relation_classification
+      type: relation-classification
+    metrics:
+    - name: F1
+      type: f1
+      value: 0.7758215962441315
+    - name: F1 (macro)
+      type: f1_macro
+      value: 0.5011043573121808
+  - task:
+      name: Lexical Relation Classification (EVALution)
+      type: classification
+    dataset:
+      name: BLESS
+      args: relbert/lexical_relation_classification
+      type: relation-classification
+    metrics:
+    - name: F1
+      type: f1
+      value: 0.5677139761646804
+    - name: F1 (macro)
+      type: f1_macro
+      value: 0.5566709002001424
+  - task:
+      name: Lexical Relation Classification (K&H+N)
+      type: classification
+    dataset:
+      name: K&H+N
+      args: relbert/lexical_relation_classification
+      type: relation-classification
+    metrics:
+    - name: F1
+      type: f1
+      value: 0.9569451206788621
+    - name: F1 (macro)
+      type: f1_macro
+      value: 0.8723408257004264
+  - task:
+      name: Lexical Relation Classification (ROOT09)
+      type: classification
+    dataset:
+      name: ROOT09
+      args: relbert/lexical_relation_classification
+      type: relation-classification
+    metrics:
+    - name: F1
+      type: f1
+      value: 0.8621121905358822
+    - name: F1 (macro)
+      type: f1_macro
+      value: 0.8575846242198137
+
+---
+# relbert/relbert-roberta-base-semeval2012-v6-mask-prompt-c-nce-2-child
+
+RelBERT fine-tuned from [roberta-base](https://huggingface.co/roberta-base) on  
+[relbert/semeval2012_relational_similarity_v6](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v6).
+Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
+It achieves the following results on the relation understanding tasks:
+- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/relbert-roberta-base-semeval2012-v6-mask-prompt-c-nce-2-child/raw/main/analogy.json)):
+    - Accuracy on SAT (full): 0.31016042780748665 
+    - Accuracy on SAT: 0.32344213649851633
+    - Accuracy on BATS: 0.3085047248471373
+    - Accuracy on U2: 0.3157894736842105
+    - Accuracy on U4: 0.2986111111111111
+    - Accuracy on Google: 0.328
+- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/relbert-roberta-base-semeval2012-v6-mask-prompt-c-nce-2-child/raw/main/classification.json)):
+    - Micro F1 score on BLESS: 0.9019135151423836
+    - Micro F1 score on CogALexV: 0.7758215962441315
+    - Micro F1 score on EVALution: 0.5677139761646804
+    - Micro F1 score on K&H+N: 0.9569451206788621
+    - Micro F1 score on ROOT09: 0.8621121905358822
+- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/relbert-roberta-base-semeval2012-v6-mask-prompt-c-nce-2-child/raw/main/relation_mapping.json)):
+    - Accuracy on Relation Mapping: 0.6354563492063492 
+
+
+### Usage
+This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip   
+```shell
+pip install relbert
+```
+and activate model as below.
+```python
+from relbert import RelBERT
+model = RelBERT("relbert/relbert-roberta-base-semeval2012-v6-mask-prompt-c-nce-2-child")
+vector = model.get_embedding(['Tokyo', 'Japan'])  # shape of (1024, )
+```
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+ - model: roberta-base
+ - max_length: 64
+ - mode: mask
+ - data: relbert/semeval2012_relational_similarity_v6
+ - split: train
+ - split_eval: validation
+ - template_mode: manual
+ - loss_function: nce_logout
+ - classification_loss: False
+ - temperature_nce_constant: 0.05
+ - temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
+ - epoch: 5
+ - batch: 128
+ - lr: 5e-06
+ - lr_decay: False
+ - lr_warmup: 1
+ - weight_decay: 0
+ - random_seed: 2
+ - exclude_relation: None
+ - n_sample: 320
+ - gradient_accumulation: 8
+ - relation_level: None
+ - data_level: child
+
+The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/relbert-roberta-base-semeval2012-v6-mask-prompt-c-nce-2-child/raw/main/trainer_config.json).
+
+### Reference
+If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
+
+```
+
+@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
+    title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
+    author = "Ushio, Asahi  and
+      Schockaert, Steven  and
+      Camacho-Collados, Jose",
+    booktitle = "EMNLP 2021",
+    year = "2021",
+    address = "Online",
+    publisher = "Association for Computational Linguistics",
+}
+
+```

analogy.json

@@ -0,0 +1 @@
+{"distance_function": "cosine_similarity", "sat/test": 0.32344213649851633, "sat/valid": 0.1891891891891892, "u2/test": 0.3157894736842105, "u2/valid": 0.20833333333333334, "u4/test": 0.2986111111111111, "u4/valid": 0.375, "google/test": 0.328, "google/valid": 0.36, "bats/test": 0.3085047248471373, "bats/valid": 0.3417085427135678, "sat_full": 0.31016042780748665}

classification.json

@@ -0,0 +1 @@
+{"lexical_relation_classification/BLESS": {"classifier_config": {"activation": "relu", "alpha": 0.0001, "batch_size": "auto", "beta_1": 0.9, "beta_2": 0.999, "early_stopping": false, "epsilon": 1e-08, "hidden_layer_sizes": [100], "learning_rate": "constant", "learning_rate_init": 0.001, "max_fun": 15000, "max_iter": 200, "momentum": 0.9, "n_iter_no_change": 10, "nesterovs_momentum": true, "power_t": 0.5, "random_state": 0, "shuffle": true, "solver": "adam", "tol": 0.0001, "validation_fraction": 0.1, "verbose": false, "warm_start": false}, "test/accuracy": 0.9019135151423836, "test/f1_macro": 0.8910141280520246, "test/f1_micro": 0.9019135151423836, "test/p_macro": 0.8804634780084405, "test/p_micro": 0.9019135151423836, "test/r_macro": 0.9044301304421047, "test/r_micro": 0.9019135151423836}, "lexical_relation_classification/CogALexV": {"classifier_config": {"activation": "relu", "alpha": 0.0001, "batch_size": "auto", "beta_1": 0.9, "beta_2": 0.999, "early_stopping": false, "epsilon": 1e-08, "hidden_layer_sizes": [100], "learning_rate": "constant", "learning_rate_init": 0.001, "max_fun": 15000, "max_iter": 200, "momentum": 0.9, "n_iter_no_change": 10, "nesterovs_momentum": true, "power_t": 0.5, "random_state": 0, "shuffle": true, "solver": "adam", "tol": 0.0001, "validation_fraction": 0.1, "verbose": false, "warm_start": false}, "test/accuracy": 0.7758215962441315, "test/f1_macro": 0.5011043573121808, "test/f1_micro": 0.7758215962441315, "test/p_macro": 0.5398316480499672, "test/p_micro": 0.7758215962441315, "test/r_macro": 0.47342479367754714, "test/r_micro": 0.7758215962441315}, "lexical_relation_classification/EVALution": {"classifier_config": {"activation": "relu", "alpha": 0.0001, "batch_size": "auto", "beta_1": 0.9, "beta_2": 0.999, "early_stopping": false, "epsilon": 1e-08, "hidden_layer_sizes": [100], "learning_rate": "constant", "learning_rate_init": 0.001, "max_fun": 15000, "max_iter": 200, "momentum": 0.9, "n_iter_no_change": 10, "nesterovs_momentum": true, "power_t": 0.5, "random_state": 0, "shuffle": true, "solver": "adam", "tol": 0.0001, "validation_fraction": 0.1, "verbose": false, "warm_start": false}, "test/accuracy": 0.5677139761646804, "test/f1_macro": 0.5566709002001424, "test/f1_micro": 0.5677139761646804, "test/p_macro": 0.555815839004099, "test/p_micro": 0.5677139761646804, "test/r_macro": 0.5588439691493308, "test/r_micro": 0.5677139761646804}, "lexical_relation_classification/K&H+N": {"classifier_config": {"activation": "relu", "alpha": 0.0001, "batch_size": "auto", "beta_1": 0.9, "beta_2": 0.999, "early_stopping": false, "epsilon": 1e-08, "hidden_layer_sizes": [100], "learning_rate": "constant", "learning_rate_init": 0.001, "max_fun": 15000, "max_iter": 200, "momentum": 0.9, "n_iter_no_change": 10, "nesterovs_momentum": true, "power_t": 0.5, "random_state": 0, "shuffle": true, "solver": "adam", "tol": 0.0001, "validation_fraction": 0.1, "verbose": false, "warm_start": false}, "test/accuracy": 0.9569451206788621, "test/f1_macro": 0.8723408257004264, "test/f1_micro": 0.9569451206788621, "test/p_macro": 0.8699093520583076, "test/p_micro": 0.9569451206788621, "test/r_macro": 0.8751212617262408, "test/r_micro": 0.9569451206788621}, "lexical_relation_classification/ROOT09": {"classifier_config": {"activation": "relu", "alpha": 0.0001, "batch_size": "auto", "beta_1": 0.9, "beta_2": 0.999, "early_stopping": false, "epsilon": 1e-08, "hidden_layer_sizes": [100], "learning_rate": "constant", "learning_rate_init": 0.001, "max_fun": 15000, "max_iter": 200, "momentum": 0.9, "n_iter_no_change": 10, "nesterovs_momentum": true, "power_t": 0.5, "random_state": 0, "shuffle": true, "solver": "adam", "tol": 0.0001, "validation_fraction": 0.1, "verbose": false, "warm_start": false}, "test/accuracy": 0.8621121905358822, "test/f1_macro": 0.8575846242198137, "test/f1_micro": 0.8621121905358822, "test/p_macro": 0.8623193682365183, "test/p_micro": 0.8621121905358822, "test/r_macro": 0.8536819839176232, "test/r_micro": 0.8621121905358822}}

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "relbert_output/models/semeval2012-v6-child/nce_logout.mask.c.2",
+  "_name_or_path": "roberta-base",
   "architectures": [
     "RobertaModel"
   ],

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0c7bbdc0e9361ae2975081c8cea81d26f98728346a585b4a25ecefdeb497cb43
-size 498649201
+oid sha256:1a1b2bdd1f700d2bc6c6b3ce78688535cdd73f2a968a15b38d811435684f0010
+size 498652017

relation_mapping.json

@@ -0,0 +1 @@
+{"accuracy": 0.6354563492063492, "prediction": [{"source": ["solar system", "sun", "planet", "mass", "attracts", "revolves", "gravity"], "true": ["atom", "nucleus", "electron", "charge", "attracts", "revolves", "electromagnetism"], "pred": ["electromagnetism", "atom", "nucleus", "charge", "attracts", "revolves", "electron"], "alignment_match": false, "accuracy": 0.42857142857142855, "similarity": 0.9378965035890966, "similarity_true": 0.9352844673651102}, {"source": ["water", "flows", "pressure", "water tower", "bucket", "filling", "emptying", "hydrodynamics"], "true": ["heat", "transfers", "temperature", "burner", "kettle", "heating", "cooling", "thermodynamics"], "pred": ["thermodynamics", "transfers", "burner", "temperature", "kettle", "cooling", "heating", "heat"], "alignment_match": false, "accuracy": 0.25, "similarity": 0.8655079415601442, "similarity_true": 0.8600756381357129}, {"source": ["waves", "shore", "reflects", "water", "breakwater", "rough", "calm", "crashing"], "true": ["sounds", "wall", "echoes", "air", "insulation", "loud", "quiet", "vibrating"], "pred": ["echoes", "wall", "vibrating", "air", "insulation", "sounds", "quiet", "loud"], "alignment_match": false, "accuracy": 0.5, "similarity": 0.8432825538521282, "similarity_true": 0.8432224427461498}, {"source": ["combustion", "fire", "fuel", "burning", "hot", "intense", "oxygen", "carbon dioxide"], "true": ["respiration", "animal", "food", "breathing", "living", "vigorous", "oxygen", "carbon dioxide"], "pred": ["respiration", "living", "animal", "breathing", "food", "vigorous", "oxygen", "carbon dioxide"], "alignment_match": false, "accuracy": 0.625, "similarity": 0.9551453797336843, "similarity_true": 0.9534160499919163}, {"source": ["sound", "low", "high", "echoes", "loud", "quiet", "horn"], "true": ["light", "red", "violet", "reflects", "bright", "dim", "lens"], "pred": ["light", "violet", "red", "reflects", "bright", "dim", "lens"], "alignment_match": false, "accuracy": 0.7142857142857143, "similarity": 0.8360415794603099, "similarity_true": 0.8351748445211076}, {"source": ["projectile", "trajectory", "earth", "parabolic", "air", "gravity", "attracts"], "true": ["planet", "orbit", "sun", "elliptical", "space", "gravity", "attracts"], "pred": ["sun", "orbit", "planet", "elliptical", "space", "gravity", "attracts"], "alignment_match": false, "accuracy": 0.7142857142857143, "similarity": 0.9517913452807311, "similarity_true": 0.9489012034568781}, {"source": ["breeds", "selection", "conformance", "artificial", "popularity", "breeding", "domesticated"], "true": ["species", "competition", "adaptation", "natural", "fitness", "mating", "wild"], "pred": ["species", "adaptation", "competition", "natural", "fitness", "mating", "wild"], "alignment_match": false, "accuracy": 0.7142857142857143, "similarity": 0.8988470091402622, "similarity_true": 0.8972794187255603}, {"source": ["ball", "billiards", "speed", "table", "bouncing", "moving", "slow", "fast"], "true": ["molecules", "gas", "temperature", "container", "pressing", "moving", "cold", "hot"], "pred": ["molecules", "gas", "temperature", "container", "pressing", "moving", "cold", "hot"], "alignment_match": true, "accuracy": 1, "similarity": 0.9071817913148514, "similarity_true": 0.9071817913148514}, {"source": ["computer", "processing", "erasing", "write", "read", "memory", "outputs", "inputs", "bug"], "true": ["mind", "thinking", "forgetting", "memorize", "remember", "memory", "muscles", "senses", "mistake"], "pred": ["mind", "thinking", "memorize", "forgetting", "remember", "memory", "senses", "muscles", "mistake"], "alignment_match": false, "accuracy": 0.5555555555555556, "similarity": 0.8825666371557345, "similarity_true": 0.8799575226900792}, {"source": ["slot machines", "reels", "spinning", "winning", "losing"], "true": ["bacteria", "genes", "mutating", "reproducing", "dying"], "pred": ["bacteria", "genes", "reproducing", "mutating", "dying"], "alignment_match": false, "accuracy": 0.6, "similarity": 0.8350862489531784, "similarity_true": 0.8244601554986888}, {"source": ["war", "soldier", "destroy", "fighting", "defeat", "attacks", "weapon"], "true": ["argument", "debater", "refute", "arguing", "acceptance", "criticizes", "logic"], "pred": ["argument", "debater", "refute", "arguing", "acceptance", "criticizes", "logic"], "alignment_match": true, "accuracy": 1, "similarity": 0.8559109745393079, "similarity_true": 0.8559109745393079}, {"source": ["buyer", "merchandise", "buying", "selling", "returning", "valuable", "worthless"], "true": ["believer", "belief", "accepting", "advocating", "rejecting", "true", "false"], "pred": ["believer", "belief", "advocating", "accepting", "rejecting", "false", "true"], "alignment_match": false, "accuracy": 0.42857142857142855, "similarity": 0.851340276404699, "similarity_true": 0.8490575377278505}, {"source": ["foundations", "buildings", "supporting", "solid", "weak", "crack"], "true": ["reasons", "theories", "confirming", "rational", "dubious", "flaw"], "pred": ["theories", "reasons", "confirming", "rational", "dubious", "flaw"], "alignment_match": false, "accuracy": 0.6666666666666666, "similarity": 0.8671772033268667, "similarity_true": 0.8585839274370454}, {"source": ["obstructions", "destination", "route", "traveller", "traveling", "companion", "arriving"], "true": ["difficulties", "goal", "plan", "person", "problem solving", "partner", "succeeding"], "pred": ["difficulties", "goal", "partner", "person", "problem solving", "plan", "succeeding"], "alignment_match": false, "accuracy": 0.7142857142857143, "similarity": 0.8549339156102247, "similarity_true": 0.8546113750137663}, {"source": ["money", "allocate", "budget", "effective", "cheap", "expansive"], "true": ["time", "invest", "schedule", "efficient", "quick", "slow"], "pred": ["schedule", "invest", "time", "efficient", "quick", "slow"], "alignment_match": false, "accuracy": 0.6666666666666666, "similarity": 0.8616641539438034, "similarity_true": 0.8574365465889275}, {"source": ["seeds", "planted", "fruitful", "fruit", "grow", "wither", "blossom"], "true": ["ideas", "inspired", "productive", "product", "develop", "fail", "succeed"], "pred": ["ideas", "inspired", "productive", "product", "succeed", "fail", "develop"], "alignment_match": false, "accuracy": 0.7142857142857143, "similarity": 0.8585868309231562, "similarity_true": 0.8575748194727562}, {"source": ["machine", "working", "turned on", "turned off", "broken", "power", "repair"], "true": ["mind", "thinking", "awake", "asleep", "confused", "intelligence", "therapy"], "pred": ["mind", "thinking", "awake", "asleep", "confused", "intelligence", "therapy"], "alignment_match": true, "accuracy": 1, "similarity": 0.8395361656947484, "similarity_true": 0.8395361656947484}, {"source": ["object", "hold", "weight", "heavy", "light"], "true": ["idea", "understand", "analyze", "important", "trivial"], "pred": ["analyze", "important", "idea", "trivial", "understand"], "alignment_match": false, "accuracy": 0, "similarity": 0.8279733632955424, "similarity_true": 0.8239986689444972}, {"source": ["follow", "leader", "path", "follower", "lost", "wanders", "twisted", "straight"], "true": ["understand", "speaker", "argument", "listener", "misunderstood", "digresses", "complicated", "simple"], "pred": ["understand", "listener", "argument", "speaker", "misunderstood", "digresses", "complicated", "simple"], "alignment_match": false, "accuracy": 0.75, "similarity": 0.8643716675893652, "similarity_true": 0.8636411508653344}, {"source": ["seeing", "light", "illuminating", "darkness", "view", "hidden"], "true": ["understanding", "knowledge", "explaining", "confusion", "interpretation", "secret"], "pred": ["understanding", "confusion", "explaining", "knowledge", "interpretation", "secret"], "alignment_match": false, "accuracy": 0.6666666666666666, "similarity": 0.8750490612326192, "similarity_true": 0.8734588562466823}]}

tokenizer_config.json

@@ -6,7 +6,7 @@
   "errors": "replace",
   "mask_token": "<mask>",
   "model_max_length": 512,
-  "name_or_path": "relbert_output/models/semeval2012-v6-child/nce_logout.mask.c.2",
+  "name_or_path": "roberta-base",
   "pad_token": "<pad>",
   "sep_token": "</s>",
   "special_tokens_map_file": null,

trainer_config.json

@@ -0,0 +1 @@
+{"model": "roberta-base", "max_length": 64, "mode": "mask", "data": "relbert/semeval2012_relational_similarity_v6", "split": "train", "split_eval": "validation", "template_mode": "manual", "template": "Today, I finally discovered the relation between <subj> and <obj> : <mask>", "loss_function": "nce_logout", "classification_loss": false, "temperature_nce_constant": 0.05, "temperature_nce_rank": {"min": 0.01, "max": 0.05, "type": "linear"}, "epoch": 5, "batch": 128, "lr": 5e-06, "lr_decay": false, "lr_warmup": 1, "weight_decay": 0, "random_seed": 2, "exclude_relation": null, "n_sample": 320, "gradient_accumulation": 8, "relation_level": null, "data_level": "child"}

validation_loss.json

@@ -0,0 +1 @@
+{"split": "validation", "loss": 13.57873693297181, "data": "relbert/semeval2012_relational_similarity_v6", "exclude_relation": null, "relation_level": null, "level": "child"}