# reload pre-trained embeddings
parser.add_argument("--src_emb", type=str, default="", help="Reload source embeddings")
parser.add_argument("--tgt_emb", type=str, default="", help="Reload target embeddings")
parser.add_argument("--max_vocab", type=int, default=200000, help="Maximum vocabulary size")
parser.add_argument("--emb_dim", type=int, default=300, help="Embedding dimension")
parser.add_argument("--normalize_embeddings", type=str, default="", help="Normalize embeddings before training")
# parse parameters
params = parser.parse_args()
# check parameters
assert params.src_lang, "source language undefined"
assert os.path.isfile(params.src_emb)
assert not params.tgt_lang or os.path.isfile(params.tgt_emb)
# build logger / model / trainer / evaluator
logger = initialize_exp(params)
src_emb, tgt_emb, mapping, _ = build_model(params, False)
trainer = Trainer(src_emb, tgt_emb, mapping, None, params)
evaluator = Evaluator(trainer)
# run evaluations
to_log = OrderedDict({'n_iter': 0})
evaluator.monolingual_wordsim(to_log)
if params.tgt_lang:
evaluator.crosslingual_wordsim(to_log)
evaluator.word_translation(to_log)
evaluator.sent_translation(to_log)
# evaluator.dist_mean_cosine(to_log)
help="Embedding dimension")
parser.add_argument("--normalize_embeddings",
type=str,
default="",
help="Normalize embeddings before training")
# parse parameters
params = parser.parse_args()
# check parameters
assert params.src_lang, "source language undefined"
assert os.path.isfile(params.src_emb)
assert not params.tgt_lang or os.path.isfile(params.tgt_emb)
assert params.dico_eval == 'default' or os.path.isfile(params.dico_eval)
# build logger / model / trainer / evaluator
logger = initialize_exp(params)
src_emb, tgt_emb, mapping, _ = build_model(params, False)
trainer = Trainer(src_emb, tgt_emb, mapping, None, params)
evaluator = Evaluator(trainer)
# run evaluations
to_log = OrderedDict({'n_iter': 0})
evaluator.monolingual_wordsim(to_log)
# evaluator.monolingual_wordanalogy(to_log)
if params.tgt_lang:
evaluator.crosslingual_wordsim(to_log)
evaluator.word_translation(to_log)
evaluator.sent_translation(to_log)
# evaluator.dist_mean_cosine(to_log)
assert os.path.isfile(params.src_emb)
assert not params.tgt_lang or os.path.isfile(params.tgt_emb)
assert params.dico_eval == 'default' or os.path.isfile(params.dico_eval)
# build logger / model / trainer / evaluator
logger = initialize_exp(params)
src_emb, tgt_emb, mapping, _ = build_model(params, False)
trainer = Trainer(src_emb, tgt_emb, mapping, None, params)
evaluator = Evaluator(trainer)
# run evaluations
to_log = OrderedDict({'n_iter': 0})
log_export = {
"name" : params.model_name
}
evaluator.monolingual_wordsim(to_log,log_export)
evaluator.monolingual_wordanalogy(to_log,log_export)
if params.tgt_lang:
# evaluator.crosslingual_wordsim(to_log)
evaluator.word_translation(to_log)
# evaluator.sent_translation(to_log)
evaluator.dist_mean_cosine(to_log)
log_export['nn-1'] = to_log['precision_at_1-nn']
log_export['nn-5'] = to_log['precision_at_5-nn']
log_export['nn-10'] = to_log['precision_at_10-nn']
log_export['csls-1'] = to_log['precision_at_1-csls_knn_10']
log_export['csls-5'] = to_log['precision_at_5-csls_knn_10']
log_export['csls-10'] = to_log['precision_at_10-csls_knn_10']
log_export['mean-cosine-nn']= to_log['mean_cosine-nn-S2T-10000']
log_export['mean-cosine-csls']= to_log['mean_cosine-csls_knn_10-S2T-10000']