GigaAM v3: transcribe.cpp GGUF GGUF conversions of ai sage/GigaAM v3 for use with transcribe.cpp. Ported from upstream commit c7f128b, pinned 2026 05 12. Validated against the gigaam author package reference at transcribe.cpp commit 42b96d9 on 2026 05 12. Offline Russian speech to text with greedy RNN T decoding. Same 16 layer Conformer encoder as the e2e variant, fine tuned to emit lowercased Russian with no punctuation; 33 entry character vocabulary. Downloads Quantization Download Size WER (FLEURS ru) : : F32 gigaam v3 rnnt F32.gguf 846 MB 8.08% F16 gigaam v3 rnnt F16.gguf 430 MB 8.08% Q8 0 gigaam v3 rnnt Q8 0.gguf 260 MB 8.08% Q6 K gigaam v3 rnnt Q6 K.gguf 217 MB 8.07% Q5 K M gigaam v3 rnnt Q5 K M.gguf 196 MB 8.12% Q4 K M gigaam v3 rnnt Q4 K M.gguf 175 MB 8.12% WER measured on the full FLEURS ru test split (775 utterances) with greedy decoding and no external LM. F32 reference baseline: 8.08%. Upstream gigaam author package measured on the same manifest: 9.46%; the 1.4 pp gap is upstream rejecting 5 long ( 25 s) utterances with Too long wav file, use 'transcribe longform' method. (counted as 100% deletion errors). On the 770 utt subset both sides decode, transcribe.cpp matches…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy