{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/flash-vstream-efficient-real-time","title":"Flash-VStream: Efficient Real-Time Understanding for Long Video Streams","arxiv_id":"2506.23825","date":"2025-06-30","proceeding":null,"authors":["Haoji Zhang","Yiqin Wang","Yansong Tang","Yong liu","Jiashi Feng","Xiaojie Jin"],"abstract":"Benefiting from the advances in large language models and cross-modal alignment, existing multimodal large language models have achieved prominent performance in image and short video understanding. However, the understanding of long videos is still challenging, as their long-context nature results in significant computational and memory overhead. Most existing work treats long videos in the same way as short videos, which is inefficient for real-world applications and hard to generalize to even longer videos. To address these issues, we propose Flash-VStream, an efficient video language model capable of processing extremely long videos and responding to user queries in real time. Particularly, we design a Flash Memory module, containing a low-capacity context memory to aggregate long-context temporal information and model the distribution of information density, and a high-capacity augmentation memory to retrieve detailed spatial information based on this distribution. Compared to existing models, Flash-VStream achieves significant reductions in inference latency. Extensive experiments on long video benchmarks and comprehensive video benchmarks, i.e., EgoSchema, MLVU, LVBench, MVBench and Video-MME, demonstrate the state-of-the-art performance and outstanding efficiency of our method. Code is available at https://github.com/IVGSZ/Flash-VStream.","url_abs":"https://arxiv.org/abs/2506.23825v1","url_pdf":"https://arxiv.org/pdf/2506.23825v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"flash-vstream-efficient-real-time","repo_url":"https://github.com/IVGSZ/Flash-VStream","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":null,"task_name":"EgoSchema"},{"task_slug":"mme","task_name":"MME"},{"task_slug":null,"task_name":"MVBench"},{"task_slug":null,"task_name":"Video MME"},{"task_slug":"video-understanding","task_name":"Video Understanding"},{"task_slug":"cross-modal-alignment","task_name":"cross-modal alignment"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2506.23825","atlas_url":"https://app.syntology.ai/?focus=2506.23825","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2506.23825"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"deterministic:regex_extraction","url":"https://github.com/IVGSZ/Flash-VStream","reach":null}],"summary":{"ran_fixture":4,"ran_draft_wrong":3,"unverified":4},"by_repo_kind":{"official":{"samples":11,"ran":7,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"c522c620944fcbb9","entry":"attention_feature","repo":"IVGSZ/Flash-VStream","repo_kind":"official","path":"Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","file_url":"https://github.com/IVGSZ/Flash-VStream/blob/HEAD/Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"c522c620944fcbb9"}},{"code_sha256_prefix":"88e007251e2c289d","entry":"drop_feature","repo":"IVGSZ/Flash-VStream","repo_kind":"official","path":"Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","file_url":"https://github.com/IVGSZ/Flash-VStream/blob/HEAD/Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"88e007251e2c289d"}},{"code_sha256_prefix":"aeab8c53bd7cc503","entry":"fast_weighted_kmeans_ordered_feature","repo":"IVGSZ/Flash-VStream","repo_kind":"official","path":"Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","file_url":"https://github.com/IVGSZ/Flash-VStream/blob/HEAD/Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"aeab8c53bd7cc503"}},{"code_sha256_prefix":"19b51c02fb81ef10","entry":"merge_feature","repo":"IVGSZ/Flash-VStream","repo_kind":"official","path":"Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","file_url":"https://github.com/IVGSZ/Flash-VStream/blob/HEAD/Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"19b51c02fb81ef10"}},{"code_sha256_prefix":"005eab8e2724916d","entry":"torchpca_weighted_kmeans_ordered_feature","repo":"IVGSZ/Flash-VStream","repo_kind":"official","path":"Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","file_url":"https://github.com/IVGSZ/Flash-VStream/blob/HEAD/Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"005eab8e2724916d"}},{"code_sha256_prefix":"ae3ae26a8f618992","entry":"weighted_kmeans_feature","repo":"IVGSZ/Flash-VStream","repo_kind":"official","path":"Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","file_url":"https://github.com/IVGSZ/Flash-VStream/blob/HEAD/Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"ae3ae26a8f618992"}},{"code_sha256_prefix":"c8be36921dc6305a","entry":"weighted_kmeans_ordered_feature","repo":"IVGSZ/Flash-VStream","repo_kind":"official","path":"Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","file_url":"https://github.com/IVGSZ/Flash-VStream/blob/HEAD/Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"c8be36921dc6305a"}},{"code_sha256_prefix":"5f35f922c5cc5480","entry":"FlashMemory","repo":"IVGSZ/Flash-VStream","repo_kind":"official","path":"Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","file_url":"https://github.com/IVGSZ/Flash-VStream/blob/HEAD/Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"5f35f922c5cc5480"}},{"code_sha256_prefix":"9c19640a32bf3839","entry":"dbscan_feature","repo":"IVGSZ/Flash-VStream","repo_kind":"official","path":"Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","file_url":"https://github.com/IVGSZ/Flash-VStream/blob/HEAD/Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"9c19640a32bf3839"}},{"code_sha256_prefix":"d1517b2bc3b5f27d","entry":"gmm_feature","repo":"IVGSZ/Flash-VStream","repo_kind":"official","path":"Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","file_url":"https://github.com/IVGSZ/Flash-VStream/blob/HEAD/Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"d1517b2bc3b5f27d"}},{"code_sha256_prefix":"9ce5acc98b5b9e0b","entry":"pca_weighted_kmeans_ordered_feature","repo":"IVGSZ/Flash-VStream","repo_kind":"official","path":"Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","file_url":"https://github.com/IVGSZ/Flash-VStream/blob/HEAD/Flash-VStream-Qwen/models/vstream_qwen2vl_realtime.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"9ce5acc98b5b9e0b"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}