AmazonScience/document-haystack
Document Haystack Dataset This repository contains the dataset for the paper “Document Haystack: A Long Context Multimodal Image/Document Understanding Vision LLM Benchmark”. 📑 Abstract Paper The proliferation of multimodal Large Language Models has significantly advanced the ability to analyze and understand complex data inputs from different modalities. However, the processing of long documents remains under-explored, largely due to a lack of suitable… See the full description on the dataset page: https://huggingface.co/datasets/AmazonScience/document-haystack.
2090k
1modified due to pandemic-related hardships to not be reported as past due based on the contractual terms of the loan, even when 2borrowers may not have made payments on their loans during the modification period. 3The following table presents a summary of accruing loans past due by delinquency status:4Table 23: Accruing Loans Past Due (a)5 Amount % of Total Loans Outstanding6 7December 31, 2023 December 31, 20228Change9December 31, 2023 December 31, 2022Dollars in millions $ %10Early stage loan delinquencies11Accruing loans past due 30 to 59 days $ 685 $ 747 $ (62) (8) % 0.21 % 0.23 %12Accruing loans past due 60 to 89 days 270 261 9 3 % 0.08 % 0.08 %13Total early stage loan delinquencies 955 1,008 (53) (5) % 0.30 % 0.31 %14Late stage loan delinquencies15Accruing loans past due 90 days or more 429 482 (53) (11) % 0.13 % 0.15 %16Total accruing loans past due $ 1,384 $ 1,490 $ (106) (7) % 0.43 % 0.46 %17(a) Past due loan amounts include government insured or guaranteed loans of $0.4 billion at both December 31, 2023 and 2022.18Accruing loans past due 90 days or more continue to accrue interest because they are (i) well secured by collateral and are in the 19process of collection, (ii) managed in homogeneous portfolios with specified charge-off timeframes adhering to regulatory guidelines, 20or (iii) certain government insured or guaranteed loans. As such, they are excluded from nonperforming loans.21Loan Modifications22We provide relief to our customers experiencing financial hardships through a variety of solutions. Commercial loan and lease 23modifications are based on each individual borrower’s situation, while consumer loan modifications are evaluated under our hardship 24relief programs.25On January 1, 2023, we adopted ASU 2022-02 Financial Instruments - Credit Losses (Topic 326): Troubled Debt Restructurings and 26Vintage Disclosures, which eliminates the accounting guidance for TDRs and enhances the disclosure requirements for certain loan 27modifications when a borrower is experiencing financial difficulty. Refer to Note 1 Accounting Policies and Note 3 Loans and Related 28Allowance for Credit Losses for additional information on our adoption of this ASU. 29Allowance for Credit Losses 30Our determination of the ACL is based on historical loss and performance experience, current economic conditions, the reasonable and31supportable forecasts of future economic conditions and other relevant factors, including current borrower and/or transaction 32characteristics and assessments of the remaining estimated contractual term as of the balance sheet date. We maintain the ACL at an 33appropriate level for expected losses on our existing investment securities, loans, equipment finance leases, other financial assets and 34unfunded lending related commitments. 35Expected losses are estimated primarily using a combination of (i) the expected losses over a reasonable and supportable forecast 36period, (ii) a period of reversion to long run average expected losses, where applicable and (iii) long run average expected losses for 37the remaining estimated contractual term. 38We use forward-looking information in estimating expected credit losses for our reasonable and supportable forecast period. For this 39purpose, we have established a framework which includes a three-year forecast period and the use of four economic scenarios and 40associated probability weights, which in combination create a forecast of expected economic outcomes. Forward-looking information, 41such as forecasted relevant macroeconomic variables, is incorporated into the expected credit loss estimates using quantitative 42macroeconomic models, as well as through analysis from PNC’s economists and management’s judgment. 43The reversion period is used to bridge our three-year reasonable and supportable forecast period and the long-run average expected 44credit losses. We consider a number of factors in determining the duration of the reversion period, such as contractual maturity of the 45asset, observed historical patterns and the estimated credit loss rates at the end of the forecast period relative to the beginning of the 46long run average period. The reversion period is typically 1-3 years, if not immediate. 47The long-run average expected credit losses are derived from available historical credit information. We use long-run average 48expected losses for the portfolio over the estimated remaining contractual term beyond our reasonable and supportable forecast period 49and the reversion period.50The following discussion provides additional information on our reserves for loans and leases as well as unfunded lending related 51commitments. See Note 1 Accounting Policies for further discussion on our ACL, including details of our methodologies and 52 5366 The PNC Financial Services Group, Inc. – 2023 Form 10-K