sentence-transformers How to use jebish7/mpnet-base-all-obliqa_NMR_3 with sentence-transformers:
from sentence_transformers import SentenceTransformer
model = SentenceTransformer("jebish7/mpnet-base-all-obliqa_NMR_3")
sentences = [
"Are there any ADGM-specific guidelines or best practices for integrating anti-money laundering (AML) compliance into our technology and financial systems to manage operational risks effectively?",
"REGULATORY REQUIREMENTS FOR AUTHORISED PERSONS ENGAGED IN REGULATED ACTIVITIES IN RELATION TO VIRTUAL ASSETS\nAnti-Money Laundering and Countering Financing of Terrorism\nIn order to develop a robust and sustainable regulatory framework for Virtual Assets, FSRA is of the view that a comprehensive application of its AML/CFT framework should be in place, including full compliance with, among other things, the:\n\na)\tUAE AML/CFT Federal Laws, including the UAE Cabinet Resolution No. (10) of 2019 Concerning the Executive Regulation of the Federal Law No. 20 of 2018 concerning Anti-Money Laundering and Combating Terrorism Financing;\n\nb)\tUAE Cabinet Resolution 20 of 2019 concerning the procedures of dealing with those listed under the UN sanctions list and UAE/local terrorist lists issued by the Cabinet, including the FSRA AML and Sanctions Rules and Guidance (“AML Rules”) or such other AML rules as may be applicable in ADGM from time to time; and\n\nc)\tadoption of international best practices (including the FATF Recommendations).\n",
"DIGITAL SECURITIES SETTLEMENT\nDigital Settlement Facilities (DSFs)\nFor the purposes of this Guidance and distinct from RCHs, the FSRA will consider DSFs suitable for the purposes of settlement (MIR Rule 3.8) and custody (MIR Rule 2.10) of Digital Securities. A DSF, holding an FSP for Providing Custody, may provide custody and settlement services in Digital Securities for RIEs and MTFs (as applicable). Therefore, for the purposes of custody and settlement of Digital Securities, the arrangements that a RIE or MTF would normally have in place with a RCH can be replaced with arrangements provided by a DSF, provided that certain requirements, as described in this section, are met.\n",
"REGULATORY REQUIREMENTS FOR AUTHORISED PERSONS ENGAGED IN REGULATED ACTIVITIES IN RELATION TO VIRTUAL ASSETS\nSecurity measures and procedures\nIT infrastructures should be strong enough to resist, without significant loss to Clients, a number of scenarios, including but not limited to: accidental destruction or breach of data, collusion or leakage of information by employees/former employees, successful hack of a cryptographic and hardware security module or server, or access by hackers of any single set of encryption/decryption keys that could result in a complete system breach.\n"
]
embeddings = model.encode(sentences)
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [4, 4] 
