{"api_version":"codesota.v1","data":{"archived_results":[],"declared_task_reviews":[],"evidence_status":"CodeSOTA measurements only; model-card claims are not benchmark results","measured_parameters":[{"basis":"Unique parameters counted in the loaded evaluator model on White","model":"Qwen/Qwen2.5-VL-3B-Instruct","parameters":3754622976,"revision":"66285546d2b821cf421d4f5eb2576359d3770cd3","source":"https://huggingface.co/Qwen/Qwen2.5-VL-3B-Instruct/tree/66285546d2b821cf421d4f5eb2576359d3770cd3","updated":1792348351}],"model":"Qwen/Qwen2.5-VL-3B-Instruct","page":"/models/Qwen/Qwen2.5-VL-3B-Instruct","pending":[],"results":[{"applicability":{"intended_task":null,"interface":"unknown","model":"Qwen/Qwen2.5-VL-3B-Instruct","operational_evidence":"exact-scope run available","profile_id":"ocrbench-v1-full","relevance":"Do not use as a reference ranking; singleton checks changed predictions and scores.","requirements":["Use the separate serial full OCRBench reference protocol.","This batch4, bounded-resolution result is retained as diagnostic evidence."],"revision":"66285546d2b821cf421d4f5eb2576359d3770cd3","scientific_validity":"not established by a model card or a successful run","scope_id":"d5b62220da83f9ebb4c271f820fa35b964ad4a0400651e30e94f5727f4a14007","sources":[],"status":"experimental_batch_sensitive","warnings":[]},"benchmark":"OCRBench v1 / full 1,000 questions \u2014 experimental batch-4","coverage":"full","direction":"max","evaluation_seconds":369.03743726201355,"finished":1792347434,"gpu_slot_seconds":372.21522576897405,"manifest_url":"/artifacts/b25ac4100084fad49c9f3938/manifest.json","metric":"accuracy","n":1000,"protocol_version":"codesota-ocrbench-v1-full-bf16-qwen-batch4-v1","revision":"66285546d2b821cf421d4f5eb2576359d3770cd3","run_id":"b25ac4100084fad49c9f3938","run_url":"/runs/b25ac4100084fad49c9f3938","scope":"d5b62220da83f9ebb4c271f820fa35b964ad4a0400651e30e94f5727f4a14007","score":0.607,"task":"ocrbench-v1-full"},{"applicability":{"intended_task":null,"interface":"unknown","model":"Qwen/Qwen2.5-VL-3B-Instruct","operational_evidence":"exact-scope run available","profile_id":"ocrbench-v1-full-serial","relevance":"unreviewed","requirements":["Review this exact revision, intended task and model interface."],"revision":"66285546d2b821cf421d4f5eb2576359d3770cd3","scientific_validity":"not established by a model card or a successful run","scope_id":"dcccfe2d1c65816985fdcc2bb76bfad20fcc76719fc61e7a4fd8a82df9ec1d29","sources":[],"status":"unreviewed","warnings":[]},"benchmark":"OCRBench v1 / full serial reference","coverage":"full","direction":"max","evaluation_seconds":680.3199762400473,"finished":1792348351,"gpu_slot_seconds":683.3304858480114,"manifest_url":"/artifacts/32dde2b675d08afa871097ec/manifest.json","metric":"accuracy","n":1000,"protocol_version":"codesota-ocrbench-v1-full-native-resolution-serial-v1","revision":"66285546d2b821cf421d4f5eb2576359d3770cd3","run_id":"32dde2b675d08afa871097ec","run_url":"/runs/32dde2b675d08afa871097ec","scope":"dcccfe2d1c65816985fdcc2bb76bfad20fcc76719fc61e7a4fd8a82df9ec1d29","score":0.813,"task":"ocrbench-v1-full-serial"}],"revisions":["66285546d2b821cf421d4f5eb2576359d3770cd3"],"source":"https://huggingface.co/Qwen/Qwen2.5-VL-3B-Instruct"}}