]
SIMILARITIES = [
("cos", CosSimilarity, {"threshold": 0.3}),
]
DIRECTIONS = [
("edge", EdgeBased, {}),
("order", OrderBased, {}),
]
SCORERS = [
("add_f=0.0_b=1.0_s=1.0", AddScorer, {}),
]
SUMMARIZERS = [DefaultSummarizer(num_words=200)]
experiment_time = int(time.time())
results_path = Path(f"results/exp10")
for embedder_id, embedder, embedder_args in EMBEDDERS:
Embedder = embedder(**embedder_args)
for Summarizer, (dataset_id, dataset, dataset_args) in zip(SUMMARIZERS, DATASETS):
DataSet = dataset(**dataset_args)
docs = list(DataSet)
if DEBUG:
docs = docs[:5]
print(f"embedding dataset {dataset_id} with {embedder_id}")
embeds = [Embedder.get_embeddings(doc) for doc in tqdm(docs)]
for similarity_id, similarity, similarity_args in SIMILARITIES:
Similarity = similarity(**similarity_args)
# {"model": "roberta-large-nli-mean-tokens"}
# ),
]
SIMILARITIES = [
("cos", CosSimilarity, {}),
]
DIRECTIONS = [
("edge", EdgeBased, {}),
]
SCORERS = [
("add_f=0.0_b=1.0_s=1.0", AddScorer, {}),
]
Summarizer = DefaultSummarizer(num_words=220)
experiment_time = int(time.time())
# results_path = Path(f"results/{experiment_time}")
results_path = Path(f"results/exp4")
for embedder_id, embedder, embedder_args in EMBEDDERS:
Embedder = embedder(**embedder_args)
for dataset_id, dataset, dataset_args in DATASETS:
DataSet = dataset(**dataset_args)
docs = list(DataSet)
if DEBUG:
docs = docs[:5]
print(f"embedding dataset {dataset_id} with {embedder_id}")
embeds = [Embedder.get_embeddings(doc) for doc in tqdm(docs)]
for similarity_id, similarity, similarity_args in SIMILARITIES:
# {"model": "roberta-large-nli-mean-tokens"}
# ),
]
SIMILARITIES = [
("cos", CosSimilarity, {}),
]
DIRECTIONS = [
("order", OrderBased, {}),
]
SCORERS = [
("add_f=0.0_b=1.0_s=1.0", AddScorer, {}),
]
SUMMARIZERS = [
DefaultSummarizer(num_words=200),
DefaultSummarizer(num_words=220)
]
experiment_time = int(time.time())
results_path = Path(f"results/exp8")
for embedder_id, embedder, embedder_args in EMBEDDERS:
Embedder = embedder(**embedder_args)
for Summarizer, (dataset_id, dataset,
dataset_args) in zip(SUMMARIZERS, DATASETS):
DataSet = dataset(**dataset_args)
docs = list(DataSet)
if DEBUG:
docs = docs[:5]
print(f"embedding dataset {dataset_id} with {embedder_id}")
}),
]
SIMILARITIES = [
("cos", CosSimilarity, {}),
]
DIRECTIONS = [
("edge", EdgeBased, {}),
]
SCORERS = [
("add_f=0.0_b=1.0_s=0.5", AddScorer, {
"section_weight": 0.5
}),
]
Summarizer = DefaultSummarizer()
experiment_time = int(time.time())
results_path = Path(f"results/exp3")
for embedder_id, embedder, embedder_args in EMBEDDERS:
Embedder = embedder(**embedder_args)
for dataset_id, dataset, dataset_args in DATASETS:
DataSet = dataset(**dataset_args)
docs = list(DataSet)
if DEBUG:
docs = docs[:5]
print(f"embedding dataset {dataset_id} with {embedder_id}")
embeds = [Embedder.get_embeddings(doc) for doc in tqdm(docs)]
for similarity_id, similarity, similarity_args in SIMILARITIES:
Similarity = similarity(**similarity_args)