GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
opencouncil
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
opencouncil
Model Tree
openai/whisper-large-v3
Input Modalities
Output Modalities
Supported Functionality
openai/whisper-large-v3. The adapter does nothing on its own.q_proj and v_proj only, PEFT 0.19.1.python
from transformers import WhisperForConditionalGeneration, WhisperProcessorfrom peft import PeftModelbase = WhisperForConditionalGeneration.from_pretrained("openai/whisper-large-v3")model = PeftModel.from_pretrained(base, "opencouncil/whisper-large-v3-el-council-lora")model = model.merge_and_unload()processor = WhisperProcessor.from_pretrained("openai/whisper-large-v3",language="greek", task="transcribe")
For v1, pass revision="v1" to from_pretrained.
Every number below comes from this configuration through faster-whisper on a
CTranslate2 float16 conversion of the merged model. A different beam size or a different
temperature ladder gives a different number, so quote these figures only alongside this
configuration.
python
dict(language="el", beam_size=5, condition_on_previous_text=False,temperature=[0.0, 0.2, 0.4, 0.6, 0.8, 1.0],no_speech_threshold=0.6, log_prob_threshold=-1.0,compression_ratio_threshold=2.4, vad_filter=False)
The training data changed. The recipe did not.
v1 treated each corrected utterance as one training example. Those average 3.55 seconds, and Whisper always reads a 30-second window, so most of every example was padding the model never meets at inference.
For v2 I dropped every clip where two people speak at once, then packed continuous spans of a single speaker until each window carried about 22 seconds of speech.
markdown
v1 28,967 clips [ utterance ]............................... 3.55 s speechv2 2,475 packs [ one speaker, one continuous span ] ~22 s speech|<------- 30-second Whisper window ------->|
Two things moved at the same time, overlap filtering and window occupancy, so this repository cannot tell you which of them did the work.
Measured on 39 frozen validation windows from Argos and Orestiada, two councils absent from all training data, and on 391 held-out windows from 117 meetings scored inside OpenCouncil's own benchmark. Every row below was decoded on one machine with the configuration above.
| validation WER | held-out test WER | deletion rate | |
|---|---|---|---|
| v2 (this revision) | 0.1390 | 0.1795 | 0.0313 |
v1 (revision="v1") | 0.1600 | 0.1867 | 0.0525 |
whisper-large-v3, not fine-tuned | 0.1988 | 0.0335 |
v2 beats v1 on validation across all three random seeds I trained. On the held-out test set the WER difference is -0.0040 with a 95% meeting-clustered interval of [-0.0078, +0.0002], which crosses zero.
The difference you can rely on is the deletion rate. v2 drops 0.0313 of the reference against v1's 0.0525, interval [-0.0251, -0.0174]. It leaves out about 40% less of what was said. For a workflow where a person reviews the transcript, a wrong word is easier to catch than a missing sentence.
On 250 occurrences of councillor surnames and place names in the same 39 windows:
| DS-WER | |
|---|---|
| Soniox | 0.3280 |
| ElevenLabs Scribe v2 | 0.3720 |
| v2, seed 47 (this revision) | 0.4360 |
| v2, seed 29 | 0.4640 |
| v1 | 0.4800 |
| v2, seed 13 | 0.4840 |
whisper-large-v3, not fine-tuned | 0.5400 |
The three v2 rows differ only in the random seed and span 0.0480, which is wider than the 0.0187 separating the v2 average from v1. Treat the domain-term ordering between v1 and v2 as unresolved.
It loses to ElevenLabs Scribe v2 and to Soniox on our own benchmark, on overall WER and on domain terms both. If you need the best available Greek council transcription and can pay for an API, use one of those. This adapter is useful when you want a self-hosted model with no per-minute cost, and it is a clear improvement on the base model it starts from.
Greek municipal council meetings: long-form, multi-speaker, procedural vocabulary, variable room acoustics. It was fine-tuned on that and measured on that.
This repository contains the adapter weights and configuration. No dataset, no audio, no transcripts.
Apache 2.0, matching the base model.
Trained and evaluated in the OpenCouncil GSoC 2026 project. Method, experiment records and every number above: github.com/eellak/gsoc2026-opencouncil-stt, report at FINAL_REPORT.md.
Use v2, the default. It leaves out far less of the meeting and matches or beats v1
everywhere else. Load revision="v1" if you need to reproduce a result published before
2026-08-23.