CoolFace
Datasetpublic

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.

sourceHugging Faceupdated 1y agoView on Hugging Face
20likes89kdownloads
HSBC_100Pages_TextNeedles_page_66.txt144 linesDownload Raw Back to Text_TextNeedles
1Net zero in our own operations continued2average carbon intensities and spend data 3to determine their contribution to our supply 4chain emissions. As more of our suppliers 5report their emissions, we should be able to 6include more accurate data and fewer industry 7averages in the calculation. We have applied 8a data quality score to the sources of data we 9used to determine counterparty emissions. For 10further details, see our GHG (Greenhouse Gas) 11Reporting Guidance at www.hsbc.com/esg.12In 2022, we disclosed our supply chain 13emissions for the first time, using supplier 14emissions data and industry averages where 15actual data was not available. This approach 16is heavily dependent on external data 17sources to calculate estimates of our supply 18chain emissions.19In 2023, emissions from our supply chain 20reduced by 3% compared with 2022. This is 21due to a reduction in spend and an increase 22in the availability of actual emissions data 23from our suppliers. Emissions have increased 24by 13% compared with 2019, as industry 25averages remain significantly elevated. Due 26to volatility in industry average data, we 27will undertake a review of our data sources 28and methodology during 2024. As supplier 29emissions reporting matures, we will be able 30to include more actual data and fewer industry 31averages in the methodology. Our initial supply 32chain emission figures may require updating 33as data availability changes over time and 34methodologies and climate science evolve.35 For further details of our methodologies and 36relevant environmental key facts, see the ESG 37Data Pack at www.hsbc.com/esg.38Energy consumption in kWh in 000s 392023 2022402019 41baseline42Total 772,736 797,264 913,55643UK only 209,939 222,322 281,27144Emissions from our energy and travel45We report our emissions following the 46Greenhouse Gas Protocol, which incorporates 47the scope 2 market-based emissions 48methodology. We report greenhouse gas 49emissions resulting from the energy used in 50our buildings and employees’ business travel. 51Due to the nature of our primary business, 52carbon dioxide is the main type of greenhouse 53gas applicable to our operations. While the 54amount is immaterial, our current reporting 55also incorporates methane and nitrous oxide 56for completeness. Our environmental data for 57our own operations is based on a 12-month 58period to 30 September.59In 2023, we reduced emissions from our 60energy consumption and travel to 293,333 61tonnes CO2e, which represents a 57.3% 62reduction compared with our 2019 baseline. 63This was mainly attributed to:64 – tra65vel volumes remaining low compared  66with pre-pandemic levels;67 – an in68crease in our consumption of 69renewable electricity to 58.4%; and70 – the r71eduction of energy consumption as  72a result of strategic footprint reductions 73and the implementation of over 450 energy 74conservation measures, which amounted  75to an estimated energy avoidance in  76excess of 12 million kWh.77Emissions from business travel increased 78compared with 2022, due to the easing of 79pandemic-related travel restrictions which 80resulted in a return to travel. A decrease in 81scope 1 emissions was partly attributed to a 82correction in the classification of road-based 83business travel in the UK and India from scope 841 to scope 3.85In 2023, we collected data on energy use 86and business travel for our operations in 34 87countries and territories, which accounted 88for approximately 96.0% of our full-time 89employees (‘FTEs’). To estimate the emissions 90of our operations in entities where we have 91operational control and a small presence, we 92scale up the emissions data from 96.0% to 93100%. We then apply emission uplift rates 94to reflect uncertainty concerning the quality 95and coverage of emission measurement and 96estimation. This is consistent with both the 97Intergovernmental Panel on Climate Change’s 98Good Practice Guidance and Uncertainty 99Management in National Greenhouse Gas 100Inventories and our internal analysis of data 101coverage and quality. 102Emissions from our supply chain103Our calculation methodology uses supplier 104emissions data where we have it from 105suppliers, through CDP. Where we do not 106have actual emissions data, we use industry 107Greenhouse gas emissions in tonnes 108CO2e per FTE1092023 20221102019 111baseline112Scope 1, 2 and 1133 (Category 6) 1.3 1.3 2.9114Scope 1, 2 and 1153 (Category 1, 1162 and 6)5.8 5.9 6.6117Energy and travel greenhouse gas emissions in tonnes CO2e1182023 20221192019 120baseline121Scope 11 16,918 19,329 22,066122Scope 2 (market-based)1 167,174 223,334 392,270123Scope 3 1,090,280 1,052,264 1,139,260124Category 1: Purchased goods and services1,2 859,256 865,747 829,635125Category 2: Capital goods1,2 121,783 144,232 37,617126Category 6: Business travel1 109,241 42,285 272,008127Total 1,274,372 1,294,927 1,553,596128Included energy UK 5,909 9,264 10,432129 1301  Our data is now presented on an absolute value basis and not rounded values. Data in 2023 is subject to an 131ind132ependent limited assurance by PwC in accordance with International Standard on Assurance engagements 1333410 (Assurance Engagements on Greenhouse Gas Statements). For further details, see GHG Reporting 134Guidance 2023 and third-party limited assurance report at www.hsbc.com/our-approach/esg-information/135esg-reporting-and-policies. In respect of data in 2019 and 2022, see our relevant Annual Report and Accounts.1362  Sup137ply chain emissions calculated using a combination of supplier emissions data and industry averages. 138A data quality score is applied to this calculation where 1 is high and 5 is low, based on the quality of 139emissions data. This is a weighted average score based on HSBC supplier spend and is in line with HSBC’s 140financed emissions reporting methodology. Data quality scores can be found in the ESG Data Pack.141 For further details of our methodologies, our PwC limited assurance reports and relevant environment key 142facts, see our ESG Data Pack at www.hsbc.com/esg.14364 HSBC Holdings plc Annual Report and Accounts 2023144ESG review | Environmental