{"type":"post","id":"glas-reports-fine-tuning-gigaam-v3-for-russian-a418b7","slug":"glas-reports-fine-tuning-gigaam-v3-for-russian-a418b7","title":"Glas reports fine-tuning GigaAM-v3 for Russian IT dictation","description":"**Thesis:** Domain-specific speech adaptation can sharply improve technical-term transcription without improving every dimension of general transcription quality.","retrieval_nugget":"**Thesis:** Domain-specific speech adaptation can sharply improve technical-term transcription without improving every dimension of general transcription quality.","published_at":"2026-10-20","updated_at":"2026-10-20","record_date":"2026-10-20","date_kind":"published_at","topics":["models","multimodal"],"source_urls":["https://habr.com/ru/articles/1095524?amp%3Butm_medium=rss&amp%3Butm_source=habrahabr"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:glas-reports-fine-tuning-gigaam-v3-for-russian-a418b7","status":"published","visuals":[],"editorial_provenance":{"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:1aa5b528e45c2f9d6a834b643ae00bece389e204b77be15b1dd89fbf7e4461de","reviewed_at":"2026-10-20T20:08:49.240Z","source_evidence_count":1,"verified_claim_count":3,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":3,"implication_count":1,"watch_condition_count":1,"related_record_count":0},"analysis":{"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"Domain-specific speech adaptation can sharply improve technical-term transcription without improving every dimension of general transcription quality.","observed_facts":[{"text":"Glas fine-tuned GigaAM-v3 on 25.1 hours of live speech.","source_urls":["https://habr.com/ru/articles/1095524?amp%3Butm_medium=rss&amp%3Butm_source=habrahabr"]},{"text":"On 525 held-out short recordings, WER fell from 8.81% to 5.76%, and correctly written terminology rose from 43.8% to 86.7%.","source_urls":["https://habr.com/ru/articles/1095524?amp%3Butm_medium=rss&amp%3Butm_source=habrahabr"]},{"text":"The authors say processing speed changed little, but word omissions increased and capitalization and punctuation worsened on ordinary speech.","source_urls":["https://habr.com/ru/articles/1095524?amp%3Butm_medium=rss&amp%3Butm_source=habrahabr"]}],"mechanism":"The team retained the deployed RNNT model architecture and focused adaptation on upper encoder layers and the RNNT head, using a terminology-aware dataset and explicit transcription conventions. The optimization target was a domain-specific failure mode—English names, loanwords, and abbreviations—not speech recognition in the abstract. This is why terminology accuracy can rise while other output qualities regress.","why_now":"The case gives builders a concrete example of a relatively bounded adaptation path for a production speech interface, alongside metrics that make trade-offs visible.","implications":["A specialized recognizer should be evaluated against both domain terminology and ordinary-language quality, rather than a single aggregate score."],"evidence_boundary":"The reported test set contains 525 short recordings and was labeled through the same automated pipeline used for training, with selective manual review. The authors explicitly note that annotation-style overlap may account for part of the measured gain. The result is version 0.01 beta, not evidence of broad independent replication.","watch_conditions":["Independent evaluation with human-verified labels should establish whether terminology gains persist without increased omissions or regressions in punctuation and capitalization."],"related_records":[],"new_branch_reason":"No prior New Runtime record is supplied; this is a distinct case about adapting a deployed speech-recognition model to technical vocabulary and the evaluation trade-offs that follow."},"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/glas-reports-fine-tuning-gigaam-v3-for-russian-a418b7/","markdown":"https://newruntime.com/posts/glas-reports-fine-tuning-gigaam-v3-for-russian-a418b7.md","json":"https://newruntime.com/posts/glas-reports-fine-tuning-gigaam-v3-for-russian-a418b7.json"},"next_reads":[{"type":"topic","path":"/topics/models/","url":"https://newruntime.com/topics/models/","title":"Models","reason":"Explore the models topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/models.md","json_url":"https://newruntime.com/topics/models.json"},{"type":"topic","path":"/topics/multimodal/","url":"https://newruntime.com/topics/multimodal/","title":"Multimodal","reason":"Explore the multimodal topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/multimodal.md","json_url":"https://newruntime.com/topics/multimodal.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/source-ledger.json","url":"https://newruntime.com/source-ledger.json","title":"Source Ledger","reason":"Inspect the public source evidence graph.","media_type":"application/json"}]}