Downloads · 30 days
45
28% of all-time downloads
RsGoksel/RsGoksel_ITU_Mainframe-ct2
RsGoksel_ITU_Mainframe-ct2 is a automatic speech recognition model from RsGoksel. Use it when you need speech turned into text. It is set up for ctranslate2. The card lists the license as mit.
Türkçe ASR — openai/whisper-large-v3-turbo üzerine LoRA fine-tune. Çok-domain + robustluk + medikal uyarlama; genel-unseen (ys-0) için optimize. CTranslate2 (float16) derlemesi — faster-whisper / WhisperX ile hızlı çı…
Downloads · 30 days
45
28% of all-time downloads
All-time downloads
163
Public
Repo size
6.5 GB
Likes
1
Public
Click a slice to open those files.
.bin1.6 GB · 100%
From the Hugging Face model README
Türkçe ASR — openai/whisper-large-v3-turbo üzerine LoRA fine-tune. Çok-domain + robustluk + medikal uyarlama; genel-unseen (ys-0) için optimize.
CTranslate2 (float16) derlemesi — faster-whisper / WhisperX ile hızlı çıkarım.
from faster_whisper import WhisperModel
model = WhisperModel("RsGoksel/RsGoksel_ITU_Mainframe-ct2", device="cuda", compute_type="float16")
segments, _ = model.transcribe("audio.wav", language="tr",
condition_on_previous_text=False, beam_size=5)
text = " ".join(s.text for s in segments).strip()
import whisperx
m = whisperx.load_model("RsGoksel/RsGoksel_ITU_Mainframe-ct2", device="cuda", compute_type="float16",
language="tr", asr_options={"condition_on_previous_text": False})
res = m.transcribe("audio.wav", language="tr")
text = " ".join(s["text"] for s in res["segments"])
Önerilen: language="tr", condition_on_previous_text=False.
Metrik: Türkçe normalizasyon sonrası mean-per-utterance WER.
Tam-hassasiyet (transformers) sürümü: RsGoksel/RsGoksel_ITU_Mainframe.