turkish-stt 0.1.0__tar.gz

turkish_stt-0.1.0/LICENSE

@@ -0,0 +1,79 @@
+mihu Community License
+Version 1.0
+
+These terms govern all use of the turkish-stt materials released by mihu
+("mihu", "Licensor"). By downloading, accessing, or otherwise using the Materials, you
+("Licensee") accept and agree to be bound by this License.
+
+1. DEFINITIONS
+   1.1 "Materials" — the turkish-stt model weights, model card, and any
+       accompanying code, documentation, or related artifacts released under this License.
+   1.2 "Outputs" — anything produced by running the Materials.
+   1.3 "Derivative" — any work created from, adapted from, or built upon the Materials.
+   1.4 "Annual Revenue" — the combined worldwide gross revenue of Licensee together with
+       its affiliates over the preceding twelve months.
+
+2. COMMUNITY GRANT
+   Provided Licensee's Annual Revenue stays below USD 500,000, and subject to every term
+   below, Licensor grants Licensee a worldwide, non-exclusive, non-transferable,
+   royalty-free, no-cost right to use, reproduce, and make Derivatives of the Materials.
+   This covers, without limitation:
+     (a) research and academic work;
+     (b) personal and other non-commercial use;
+     (c) commercial use within the revenue limit above, once any commercial-use
+         registration that Licensor requires has been completed.
+
+3. ABOVE THE THRESHOLD
+   Once Licensee's Annual Revenue reaches USD 500,000 or more, no use is permitted under
+   the Community Grant. Licensee must first secure a separate commercial agreement from
+   mihu. Write to support@mihu.ai to arrange terms.
+
+4. ATTRIBUTION
+   Licensee shall (a) keep this License intact, together with any copyright or attribution
+   notices, in a "NOTICE" file shipped alongside the Materials or any Derivative, and
+   (b) show "Powered by mihu" clearly in any product, service, interface, or documentation
+   that builds on the Materials or their Outputs.
+
+5. RESTRICTIONS
+   Licensee shall not:
+     (a) use the Materials or Outputs to train, build, or refine any competing speech,
+         foundational, or general-purpose AI model;
+     (b) strip, hide, or modify any attribution or proprietary notice;
+     (c) use the mihu name, logo, or marks beyond the attribution required in Section 4
+         (no trademark rights are granted here);
+     (d) use the Materials unlawfully or contrary to any Acceptable Use Policy that
+         Licensor may publish;
+     (e) sublicense, sell, or redistribute the Materials except where this License
+         expressly allows it.
+
+6. OWNERSHIP
+   Between the parties, Licensor keeps all right, title, and interest in the Materials.
+   Licensee holds the Derivatives it lawfully makes and the Outputs it generates, always
+   subject to Licensor's rights in the underlying Materials. Nothing beyond the rights
+   stated here is granted, by implication or otherwise.
+
+7. NO WARRANTY
+   THE MATERIALS ARE SUPPLIED "AS IS", WITH NO WARRANTY OF ANY KIND, WHETHER EXPRESS OR
+   IMPLIED, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
+   PURPOSE, OR NON-INFRINGEMENT.
+
+8. LIMITATION OF LIABILITY
+   TO THE FULLEST EXTENT THE LAW ALLOWS, LICENSOR IS NOT LIABLE FOR ANY INDIRECT,
+   INCIDENTAL, SPECIAL, CONSEQUENTIAL, OR EXEMPLARY DAMAGES CONNECTED TO THE MATERIALS OR
+   THIS LICENSE.
+
+9. TERMINATION
+   This License ends automatically if (a) Licensee's Annual Revenue reaches or passes
+   USD 500,000 without a commercial agreement in place, (b) Licensee breaches any term, or
+   (c) Licensee brings litigation or a patent claim against Licensor over the Materials.
+   On termination, Licensee must stop all use and delete every copy of the Materials. To
+   keep using them, arrange a commercial agreement via support@mihu.ai.
+
+10. GENERAL
+   This License is the complete agreement covering the Materials. It is governed by the
+   laws of Licensor's principal place of business, disregarding conflict-of-law rules.
+
+© mihu. All rights reserved.
+
+turkish-stt and the mihu Community License are provided by Hunters AI, operating
+under the mihu brand. "mihu" is a trademark of Hunters AI.

turkish_stt-0.1.0/PKG-INFO

@@ -0,0 +1,198 @@
+Metadata-Version: 2.4
+Name: turkish-stt
+Version: 0.1.0
+Summary: Real-time, low-latency Turkish speech-to-text (STT) that runs on the CPU.
+Author: mihu · Speech Systems
+Maintainer-email: Deniz Gökdünek <deniz@mihu.ai>
+License: mihu Community License
+        Version 1.0
+        
+        These terms govern all use of the turkish-stt materials released by mihu
+        ("mihu", "Licensor"). By downloading, accessing, or otherwise using the Materials, you
+        ("Licensee") accept and agree to be bound by this License.
+        
+        1. DEFINITIONS
+           1.1 "Materials" — the turkish-stt model weights, model card, and any
+               accompanying code, documentation, or related artifacts released under this License.
+           1.2 "Outputs" — anything produced by running the Materials.
+           1.3 "Derivative" — any work created from, adapted from, or built upon the Materials.
+           1.4 "Annual Revenue" — the combined worldwide gross revenue of Licensee together with
+               its affiliates over the preceding twelve months.
+        
+        2. COMMUNITY GRANT
+           Provided Licensee's Annual Revenue stays below USD 500,000, and subject to every term
+           below, Licensor grants Licensee a worldwide, non-exclusive, non-transferable,
+           royalty-free, no-cost right to use, reproduce, and make Derivatives of the Materials.
+           This covers, without limitation:
+             (a) research and academic work;
+             (b) personal and other non-commercial use;
+             (c) commercial use within the revenue limit above, once any commercial-use
+                 registration that Licensor requires has been completed.
+        
+        3. ABOVE THE THRESHOLD
+           Once Licensee's Annual Revenue reaches USD 500,000 or more, no use is permitted under
+           the Community Grant. Licensee must first secure a separate commercial agreement from
+           mihu. Write to support@mihu.ai to arrange terms.
+        
+        4. ATTRIBUTION
+           Licensee shall (a) keep this License intact, together with any copyright or attribution
+           notices, in a "NOTICE" file shipped alongside the Materials or any Derivative, and
+           (b) show "Powered by mihu" clearly in any product, service, interface, or documentation
+           that builds on the Materials or their Outputs.
+        
+        5. RESTRICTIONS
+           Licensee shall not:
+             (a) use the Materials or Outputs to train, build, or refine any competing speech,
+                 foundational, or general-purpose AI model;
+             (b) strip, hide, or modify any attribution or proprietary notice;
+             (c) use the mihu name, logo, or marks beyond the attribution required in Section 4
+                 (no trademark rights are granted here);
+             (d) use the Materials unlawfully or contrary to any Acceptable Use Policy that
+                 Licensor may publish;
+             (e) sublicense, sell, or redistribute the Materials except where this License
+                 expressly allows it.
+        
+        6. OWNERSHIP
+           Between the parties, Licensor keeps all right, title, and interest in the Materials.
+           Licensee holds the Derivatives it lawfully makes and the Outputs it generates, always
+           subject to Licensor's rights in the underlying Materials. Nothing beyond the rights
+           stated here is granted, by implication or otherwise.
+        
+        7. NO WARRANTY
+           THE MATERIALS ARE SUPPLIED "AS IS", WITH NO WARRANTY OF ANY KIND, WHETHER EXPRESS OR
+           IMPLIED, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
+           PURPOSE, OR NON-INFRINGEMENT.
+        
+        8. LIMITATION OF LIABILITY
+           TO THE FULLEST EXTENT THE LAW ALLOWS, LICENSOR IS NOT LIABLE FOR ANY INDIRECT,
+           INCIDENTAL, SPECIAL, CONSEQUENTIAL, OR EXEMPLARY DAMAGES CONNECTED TO THE MATERIALS OR
+           THIS LICENSE.
+        
+        9. TERMINATION
+           This License ends automatically if (a) Licensee's Annual Revenue reaches or passes
+           USD 500,000 without a commercial agreement in place, (b) Licensee breaches any term, or
+           (c) Licensee brings litigation or a patent claim against Licensor over the Materials.
+           On termination, Licensee must stop all use and delete every copy of the Materials. To
+           keep using them, arrange a commercial agreement via support@mihu.ai.
+        
+        10. GENERAL
+           This License is the complete agreement covering the Materials. It is governed by the
+           laws of Licensor's principal place of business, disregarding conflict-of-law rules.
+        
+        © mihu. All rights reserved.
+        
+        turkish-stt and the mihu Community License are provided by Hunters AI, operating
+        under the mihu brand. "mihu" is a trademark of Hunters AI.
+        
+Project-URL: Model weights, https://huggingface.co/mihuai/turkish-stt
+Keywords: speech-recognition,asr,streaming,turkish,on-device,cpu,voice-agent
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: Other/Proprietary License
+Classifier: Topic :: Multimedia :: Sound/Audio :: Speech
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: sherpa-onnx
+Requires-Dist: numpy
+Requires-Dist: soundfile
+Requires-Dist: soxr
+Requires-Dist: huggingface_hub
+Provides-Extra: eval
+Requires-Dist: jiwer; extra == "eval"
+Requires-Dist: num2words; extra == "eval"
+Dynamic: license-file
+
+# turkish-stt
+
+**Real-time, low-latency speech-to-text (STT) for Turkish — on the CPU.**
+
+Turkish-only · streaming · CPU-native · runs on the edge. A compact **66 M-parameter** causal streaming model,
+built for voice agents, call centers, and on-device transcription where a GPU is not an option — no API key, no
+cloud, no GPU.
+
+- 🇹🇷 **Turkish-first** — every parameter serves one language, not ~100
+- ⚡ **Streaming** — partial results as you speak (~320 ms), no 30-second windows
+- 🖥️ **CPU-native** — **RTF 0.128** on a single CPU thread (~8× faster than real time)
+- 🔌 **Tiny** — ~68 MB; runs on a Raspberry-Pi-class device
+- 🏷️ **Entity/brand aware** — tuned for Turkish brands, models, and numbers
+
+## Installation
+
+```bash
+pip install turkish-stt
+```
+
+## Usage
+
+```python
+import mihu_stt
+
+# one-shot: transcribe a file (16 kHz mono, or anything soundfile reads)
+text = mihu_stt.transcribe("audio.wav")
+print(text)
+```
+
+Streaming — feed audio as it arrives and read an updating transcript, with turn-end detection built in:
+
+```python
+import mihu_stt
+
+stt = mihu_stt.StreamingSTT()
+# pcm16 = 16-bit PCM bytes from your mic / phone line
+text, is_final = stt.accept_pcm16(pcm16, sample_rate=16000)
+# `text` updates while speech continues; `is_final` turns True at end-of-utterance
+```
+
+Model weights download automatically from Hugging Face
+([`mihuai/turkish-stt`](https://huggingface.co/mihuai/turkish-stt)) the first time and are cached.
+
+## Integrations
+
+Drop-in adapters for the common voice-agent stacks (see [`examples/`](examples)):
+
+**LiveKit Agents** — [`examples/livekit_stt.py`](examples/livekit_stt.py):
+```python
+from livekit.agents import AgentSession
+from livekit_stt import MihuSTT
+session = AgentSession(stt=MihuSTT(), llm=..., tts=...)
+```
+
+**Pipecat** — [`examples/pipecat_stt.py`](examples/pipecat_stt.py):
+```python
+from pipecat_stt import MihuSTTService
+pipeline = Pipeline([transport.input(), vad, MihuSTTService(), llm, tts, transport.output()])
+```
+
+## Benchmarks
+
+Measured in-house under one matched Turkish normalization. Real-human held-out **FLEURS-TR** (743 clips) and a
+second real-human set **MediaSpeech-TR** (200-clip subset), both audited absent from training. Latency on a
+single CPU thread (AMD EPYC 7282).
+
+| System | Class | Params | FLEURS WER | MediaSpeech WER | CPU RTF |
+|---|---|---:|---:|---:|---:|
+| **turkish-stt** | **streaming · CPU** | **66 M** | **13.90%** | **15.89%** | **0.128** |
+| Vosk-TR | streaming · CPU | ~50 M | 34.09% | 31.02% | 0.138 |
+| Whisper small | offline · GPU | 244 M | 15.12% | — | 1.098 |
+| Whisper large-v3 | offline · GPU | 1.5 B | 7.08% | — | 6.095 |
+
+Among **streaming, CPU-deployable** systems turkish-stt more than halves Vosk's error on both real-human sets,
+matches the four-times-larger offline Whisper-small, and is the only option that is real-time on CPU — it also
+degrades more gracefully than Vosk under additive noise. A detailed technical report — full methodology, ablations,
+latency/concurrency study, and the leakage audits — is being prepared and will be shared.
+
+## Reproduce the numbers
+
+The evaluation set, references, and scripts are in [`eval/`](eval) — including the text (n-gram) and acoustic
+(chromaprint) contamination audits used to verify the held-out sets are absent from training.
+
+```bash
+pip install "turkish-stt[eval]"
+python eval/run_eval.py --data-dir eval/tts-synthetic
+```
+
+## License
+
+**mihu Community License** — free for research, personal, and limited commercial use. See [`LICENSE`](LICENSE)
+for the full terms. Weights are released under the same license on Hugging Face.

turkish_stt-0.1.0/README.md

@@ -0,0 +1,93 @@
+# turkish-stt
+
+**Real-time, low-latency speech-to-text (STT) for Turkish — on the CPU.**
+
+Turkish-only · streaming · CPU-native · runs on the edge. A compact **66 M-parameter** causal streaming model,
+built for voice agents, call centers, and on-device transcription where a GPU is not an option — no API key, no
+cloud, no GPU.
+
+- 🇹🇷 **Turkish-first** — every parameter serves one language, not ~100
+- ⚡ **Streaming** — partial results as you speak (~320 ms), no 30-second windows
+- 🖥️ **CPU-native** — **RTF 0.128** on a single CPU thread (~8× faster than real time)
+- 🔌 **Tiny** — ~68 MB; runs on a Raspberry-Pi-class device
+- 🏷️ **Entity/brand aware** — tuned for Turkish brands, models, and numbers
+
+## Installation
+
+```bash
+pip install turkish-stt
+```
+
+## Usage
+
+```python
+import mihu_stt
+
+# one-shot: transcribe a file (16 kHz mono, or anything soundfile reads)
+text = mihu_stt.transcribe("audio.wav")
+print(text)
+```
+
+Streaming — feed audio as it arrives and read an updating transcript, with turn-end detection built in:
+
+```python
+import mihu_stt
+
+stt = mihu_stt.StreamingSTT()
+# pcm16 = 16-bit PCM bytes from your mic / phone line
+text, is_final = stt.accept_pcm16(pcm16, sample_rate=16000)
+# `text` updates while speech continues; `is_final` turns True at end-of-utterance
+```
+
+Model weights download automatically from Hugging Face
+([`mihuai/turkish-stt`](https://huggingface.co/mihuai/turkish-stt)) the first time and are cached.
+
+## Integrations
+
+Drop-in adapters for the common voice-agent stacks (see [`examples/`](examples)):
+
+**LiveKit Agents** — [`examples/livekit_stt.py`](examples/livekit_stt.py):
+```python
+from livekit.agents import AgentSession
+from livekit_stt import MihuSTT
+session = AgentSession(stt=MihuSTT(), llm=..., tts=...)
+```
+
+**Pipecat** — [`examples/pipecat_stt.py`](examples/pipecat_stt.py):
+```python
+from pipecat_stt import MihuSTTService
+pipeline = Pipeline([transport.input(), vad, MihuSTTService(), llm, tts, transport.output()])
+```
+
+## Benchmarks
+
+Measured in-house under one matched Turkish normalization. Real-human held-out **FLEURS-TR** (743 clips) and a
+second real-human set **MediaSpeech-TR** (200-clip subset), both audited absent from training. Latency on a
+single CPU thread (AMD EPYC 7282).
+
+| System | Class | Params | FLEURS WER | MediaSpeech WER | CPU RTF |
+|---|---|---:|---:|---:|---:|
+| **turkish-stt** | **streaming · CPU** | **66 M** | **13.90%** | **15.89%** | **0.128** |
+| Vosk-TR | streaming · CPU | ~50 M | 34.09% | 31.02% | 0.138 |
+| Whisper small | offline · GPU | 244 M | 15.12% | — | 1.098 |
+| Whisper large-v3 | offline · GPU | 1.5 B | 7.08% | — | 6.095 |
+
+Among **streaming, CPU-deployable** systems turkish-stt more than halves Vosk's error on both real-human sets,
+matches the four-times-larger offline Whisper-small, and is the only option that is real-time on CPU — it also
+degrades more gracefully than Vosk under additive noise. A detailed technical report — full methodology, ablations,
+latency/concurrency study, and the leakage audits — is being prepared and will be shared.
+
+## Reproduce the numbers
+
+The evaluation set, references, and scripts are in [`eval/`](eval) — including the text (n-gram) and acoustic
+(chromaprint) contamination audits used to verify the held-out sets are absent from training.
+
+```bash
+pip install "turkish-stt[eval]"
+python eval/run_eval.py --data-dir eval/tts-synthetic
+```
+
+## License
+
+**mihu Community License** — free for research, personal, and limited commercial use. See [`LICENSE`](LICENSE)
+for the full terms. Weights are released under the same license on Hugging Face.

turkish_stt-0.1.0/mihu_stt/__init__.py

@@ -0,0 +1,120 @@
+# -*- coding: utf-8 -*-
+"""
+turkish-stt — real-time Turkish speech-to-text, on the CPU.
+
+    import mihu_stt
+
+    # one-shot
+    text = mihu_stt.transcribe("audio.wav")
+
+    # streaming (voice agents): feed audio as it arrives
+    stt = mihu_stt.StreamingSTT()
+    text, is_final = stt.accept_pcm16(pcm16_bytes, sample_rate=16000)
+
+Weights download once from Hugging Face (mihuai/turkish-stt) and are cached locally. No API key,
+no GPU, no cloud — Turkish-only, runs faster than real time on a single CPU thread.
+"""
+import numpy as np
+
+__version__ = "0.1.0"
+_HF_REPO = "mihuai/turkish-stt"
+_cache = {}
+
+
+def _model_files(model_dir=None, quantized=True):
+    # int8 encoder is the default (smaller/faster). The fp32 encoder is a byte-exact
+    # twin of the same model — quality is effectively identical (~0.1% WER), it just
+    # trades ~260 MB for no quantization rounding. Kept as an opt-in for anyone who wants it.
+    encoder = "encoder.int8.onnx" if quantized else "encoder.onnx"
+    if model_dir:
+        import os
+        j = lambda f: os.path.join(model_dir, f)
+        return j("tokens.txt"), j(encoder), j("decoder.onnx"), j("joiner.int8.onnx")
+    from huggingface_hub import hf_hub_download
+    dl = lambda f: hf_hub_download(_HF_REPO, f)
+    return dl("tokens.txt"), dl(encoder), dl("decoder.onnx"), dl("joiner.int8.onnx")
+
+
+def _build(model_dir=None, num_threads=2, endpoint=True, quantized=True):
+    import sherpa_onnx
+    tokens, encoder, decoder, joiner = _model_files(model_dir, quantized)
+    kw = dict(
+        tokens=tokens, encoder=encoder, decoder=decoder, joiner=joiner,
+        num_threads=num_threads, decoding_method="modified_beam_search", provider="cpu",
+    )
+    if endpoint:
+        kw.update(enable_endpoint_detection=True, rule1_min_trailing_silence=2.4,
+                  rule2_min_trailing_silence=1.2, rule3_min_utterance_length=300)
+    return sherpa_onnx.OnlineRecognizer.from_transducer(**kw)
+
+
+def _to_16k_float(audio, sample_rate):
+    if isinstance(audio, str):
+        import soundfile as sf
+        audio, sample_rate = sf.read(audio, dtype="float32")
+    audio = np.asarray(audio, dtype=np.float32)
+    if audio.ndim > 1:
+        audio = audio.mean(axis=1)
+    if audio.dtype.kind == "i":
+        audio = audio.astype(np.float32) / 32768.0
+    if sample_rate != 16000:
+        import soxr
+        audio = soxr.resample(audio, sample_rate, 16000)
+    return audio
+
+
+def transcribe(audio, sample_rate=16000, model_dir=None, quantized=True):
+    """Transcribe a wav path or a mono numpy array (float32 [-1,1] or int16). Returns Turkish text.
+
+    quantized=True  -> int8 encoder (default): smaller, faster.
+    quantized=False -> fp32 encoder: same accuracy, larger download.
+    """
+    rec = _cache.get(("batch", model_dir, quantized))
+    if rec is None:
+        rec = _cache[("batch", model_dir, quantized)] = _build(model_dir, endpoint=False, quantized=quantized)
+    x = _to_16k_float(audio, sample_rate)
+    s = rec.create_stream()
+    s.accept_waveform(16000, x)
+    s.accept_waveform(16000, np.zeros(8000, dtype=np.float32))
+    s.input_finished()
+    while rec.is_ready(s):
+        rec.decode_stream(s)
+    r = rec.get_result(s)
+    return (r if isinstance(r, str) else r.text).strip()
+
+
+class StreamingSTT:
+    """
+    Streaming recognizer with turn-end (endpoint) detection.
+
+        stt = StreamingSTT()
+        text, is_final = stt.accept_pcm16(pcm16_bytes, sample_rate)
+
+    `is_final` becomes True when the model detects end-of-utterance; the stream then resets so the
+    next turn starts clean. Ideal for LiveKit / Pipecat and other real-time voice pipelines.
+    """
+    def __init__(self, model_dir=None, num_threads=2, quantized=True):
+        self._rec = _build(model_dir, num_threads=num_threads, endpoint=True, quantized=quantized)
+        self._stream = self._rec.create_stream()
+
+    def _text(self):
+        r = self._rec.get_result(self._stream)
+        return (r if isinstance(r, str) else r.text).strip()
+
+    def accept_pcm16(self, pcm_bytes: bytes, sample_rate: int = 16000):
+        x = np.frombuffer(pcm_bytes, dtype=np.int16).astype(np.float32) / 32768.0
+        if sample_rate != 16000:
+            import soxr
+            x = soxr.resample(x, sample_rate, 16000)
+        self._stream.accept_waveform(16000, x)
+        while self._rec.is_ready(self._stream):
+            self._rec.decode_stream(self._stream)
+        text = self._text()
+        is_final = self._rec.is_endpoint(self._stream)
+        if is_final:
+            self._rec.reset(self._stream)
+        return text, is_final
+
+    def accept_float32(self, samples, sample_rate: int = 16000):
+        pcm = (np.clip(np.asarray(samples, dtype=np.float32), -1, 1) * 32768).astype(np.int16).tobytes()
+        return self.accept_pcm16(pcm, sample_rate)

turkish_stt-0.1.0/pyproject.toml

@@ -0,0 +1,36 @@
+[build-system]
+requires = ["setuptools>=61"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "turkish-stt"
+version = "0.1.0"
+description = "Real-time, low-latency Turkish speech-to-text (STT) that runs on the CPU."
+readme = "README.md"
+requires-python = ">=3.9"
+license = { file = "LICENSE" }
+authors = [{ name = "mihu · Speech Systems" }]
+maintainers = [{ name = "Deniz Gökdünek", email = "deniz@mihu.ai" }]
+keywords = ["speech-recognition", "asr", "streaming", "turkish", "on-device", "cpu", "voice-agent"]
+classifiers = [
+  "Programming Language :: Python :: 3",
+  "License :: Other/Proprietary License",
+  "Topic :: Multimedia :: Sound/Audio :: Speech",
+  "Operating System :: OS Independent",
+]
+dependencies = [
+  "sherpa-onnx",
+  "numpy",
+  "soundfile",
+  "soxr",
+  "huggingface_hub",
+]
+
+[project.optional-dependencies]
+eval = ["jiwer", "num2words"]
+
+[project.urls]
+"Model weights" = "https://huggingface.co/mihuai/turkish-stt"
+
+[tool.setuptools.packages.find]
+include = ["mihu_stt*"]

turkish_stt-0.1.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

turkish_stt-0.1.0/turkish_stt.egg-info/PKG-INFO

@@ -0,0 +1,198 @@
+Metadata-Version: 2.4
+Name: turkish-stt
+Version: 0.1.0
+Summary: Real-time, low-latency Turkish speech-to-text (STT) that runs on the CPU.
+Author: mihu · Speech Systems
+Maintainer-email: Deniz Gökdünek <deniz@mihu.ai>
+License: mihu Community License
+        Version 1.0
+        
+        These terms govern all use of the turkish-stt materials released by mihu
+        ("mihu", "Licensor"). By downloading, accessing, or otherwise using the Materials, you
+        ("Licensee") accept and agree to be bound by this License.
+        
+        1. DEFINITIONS
+           1.1 "Materials" — the turkish-stt model weights, model card, and any
+               accompanying code, documentation, or related artifacts released under this License.
+           1.2 "Outputs" — anything produced by running the Materials.
+           1.3 "Derivative" — any work created from, adapted from, or built upon the Materials.
+           1.4 "Annual Revenue" — the combined worldwide gross revenue of Licensee together with
+               its affiliates over the preceding twelve months.
+        
+        2. COMMUNITY GRANT
+           Provided Licensee's Annual Revenue stays below USD 500,000, and subject to every term
+           below, Licensor grants Licensee a worldwide, non-exclusive, non-transferable,
+           royalty-free, no-cost right to use, reproduce, and make Derivatives of the Materials.
+           This covers, without limitation:
+             (a) research and academic work;
+             (b) personal and other non-commercial use;
+             (c) commercial use within the revenue limit above, once any commercial-use
+                 registration that Licensor requires has been completed.
+        
+        3. ABOVE THE THRESHOLD
+           Once Licensee's Annual Revenue reaches USD 500,000 or more, no use is permitted under
+           the Community Grant. Licensee must first secure a separate commercial agreement from
+           mihu. Write to support@mihu.ai to arrange terms.
+        
+        4. ATTRIBUTION
+           Licensee shall (a) keep this License intact, together with any copyright or attribution
+           notices, in a "NOTICE" file shipped alongside the Materials or any Derivative, and
+           (b) show "Powered by mihu" clearly in any product, service, interface, or documentation
+           that builds on the Materials or their Outputs.
+        
+        5. RESTRICTIONS
+           Licensee shall not:
+             (a) use the Materials or Outputs to train, build, or refine any competing speech,
+                 foundational, or general-purpose AI model;
+             (b) strip, hide, or modify any attribution or proprietary notice;
+             (c) use the mihu name, logo, or marks beyond the attribution required in Section 4
+                 (no trademark rights are granted here);
+             (d) use the Materials unlawfully or contrary to any Acceptable Use Policy that
+                 Licensor may publish;
+             (e) sublicense, sell, or redistribute the Materials except where this License
+                 expressly allows it.
+        
+        6. OWNERSHIP
+           Between the parties, Licensor keeps all right, title, and interest in the Materials.
+           Licensee holds the Derivatives it lawfully makes and the Outputs it generates, always
+           subject to Licensor's rights in the underlying Materials. Nothing beyond the rights
+           stated here is granted, by implication or otherwise.
+        
+        7. NO WARRANTY
+           THE MATERIALS ARE SUPPLIED "AS IS", WITH NO WARRANTY OF ANY KIND, WHETHER EXPRESS OR
+           IMPLIED, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
+           PURPOSE, OR NON-INFRINGEMENT.
+        
+        8. LIMITATION OF LIABILITY
+           TO THE FULLEST EXTENT THE LAW ALLOWS, LICENSOR IS NOT LIABLE FOR ANY INDIRECT,
+           INCIDENTAL, SPECIAL, CONSEQUENTIAL, OR EXEMPLARY DAMAGES CONNECTED TO THE MATERIALS OR
+           THIS LICENSE.
+        
+        9. TERMINATION
+           This License ends automatically if (a) Licensee's Annual Revenue reaches or passes
+           USD 500,000 without a commercial agreement in place, (b) Licensee breaches any term, or
+           (c) Licensee brings litigation or a patent claim against Licensor over the Materials.
+           On termination, Licensee must stop all use and delete every copy of the Materials. To
+           keep using them, arrange a commercial agreement via support@mihu.ai.
+        
+        10. GENERAL
+           This License is the complete agreement covering the Materials. It is governed by the
+           laws of Licensor's principal place of business, disregarding conflict-of-law rules.
+        
+        © mihu. All rights reserved.
+        
+        turkish-stt and the mihu Community License are provided by Hunters AI, operating
+        under the mihu brand. "mihu" is a trademark of Hunters AI.
+        
+Project-URL: Model weights, https://huggingface.co/mihuai/turkish-stt
+Keywords: speech-recognition,asr,streaming,turkish,on-device,cpu,voice-agent
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: Other/Proprietary License
+Classifier: Topic :: Multimedia :: Sound/Audio :: Speech
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: sherpa-onnx
+Requires-Dist: numpy
+Requires-Dist: soundfile
+Requires-Dist: soxr
+Requires-Dist: huggingface_hub
+Provides-Extra: eval
+Requires-Dist: jiwer; extra == "eval"
+Requires-Dist: num2words; extra == "eval"
+Dynamic: license-file
+
+# turkish-stt
+
+**Real-time, low-latency speech-to-text (STT) for Turkish — on the CPU.**
+
+Turkish-only · streaming · CPU-native · runs on the edge. A compact **66 M-parameter** causal streaming model,
+built for voice agents, call centers, and on-device transcription where a GPU is not an option — no API key, no
+cloud, no GPU.
+
+- 🇹🇷 **Turkish-first** — every parameter serves one language, not ~100
+- ⚡ **Streaming** — partial results as you speak (~320 ms), no 30-second windows
+- 🖥️ **CPU-native** — **RTF 0.128** on a single CPU thread (~8× faster than real time)
+- 🔌 **Tiny** — ~68 MB; runs on a Raspberry-Pi-class device
+- 🏷️ **Entity/brand aware** — tuned for Turkish brands, models, and numbers
+
+## Installation
+
+```bash
+pip install turkish-stt
+```
+
+## Usage
+
+```python
+import mihu_stt
+
+# one-shot: transcribe a file (16 kHz mono, or anything soundfile reads)
+text = mihu_stt.transcribe("audio.wav")
+print(text)
+```
+
+Streaming — feed audio as it arrives and read an updating transcript, with turn-end detection built in:
+
+```python
+import mihu_stt
+
+stt = mihu_stt.StreamingSTT()
+# pcm16 = 16-bit PCM bytes from your mic / phone line
+text, is_final = stt.accept_pcm16(pcm16, sample_rate=16000)
+# `text` updates while speech continues; `is_final` turns True at end-of-utterance
+```
+
+Model weights download automatically from Hugging Face
+([`mihuai/turkish-stt`](https://huggingface.co/mihuai/turkish-stt)) the first time and are cached.
+
+## Integrations
+
+Drop-in adapters for the common voice-agent stacks (see [`examples/`](examples)):
+
+**LiveKit Agents** — [`examples/livekit_stt.py`](examples/livekit_stt.py):
+```python
+from livekit.agents import AgentSession
+from livekit_stt import MihuSTT
+session = AgentSession(stt=MihuSTT(), llm=..., tts=...)
+```
+
+**Pipecat** — [`examples/pipecat_stt.py`](examples/pipecat_stt.py):
+```python
+from pipecat_stt import MihuSTTService
+pipeline = Pipeline([transport.input(), vad, MihuSTTService(), llm, tts, transport.output()])
+```
+
+## Benchmarks
+
+Measured in-house under one matched Turkish normalization. Real-human held-out **FLEURS-TR** (743 clips) and a
+second real-human set **MediaSpeech-TR** (200-clip subset), both audited absent from training. Latency on a
+single CPU thread (AMD EPYC 7282).
+
+| System | Class | Params | FLEURS WER | MediaSpeech WER | CPU RTF |
+|---|---|---:|---:|---:|---:|
+| **turkish-stt** | **streaming · CPU** | **66 M** | **13.90%** | **15.89%** | **0.128** |
+| Vosk-TR | streaming · CPU | ~50 M | 34.09% | 31.02% | 0.138 |
+| Whisper small | offline · GPU | 244 M | 15.12% | — | 1.098 |
+| Whisper large-v3 | offline · GPU | 1.5 B | 7.08% | — | 6.095 |
+
+Among **streaming, CPU-deployable** systems turkish-stt more than halves Vosk's error on both real-human sets,
+matches the four-times-larger offline Whisper-small, and is the only option that is real-time on CPU — it also
+degrades more gracefully than Vosk under additive noise. A detailed technical report — full methodology, ablations,
+latency/concurrency study, and the leakage audits — is being prepared and will be shared.
+
+## Reproduce the numbers
+
+The evaluation set, references, and scripts are in [`eval/`](eval) — including the text (n-gram) and acoustic
+(chromaprint) contamination audits used to verify the held-out sets are absent from training.
+
+```bash
+pip install "turkish-stt[eval]"
+python eval/run_eval.py --data-dir eval/tts-synthetic
+```
+
+## License
+
+**mihu Community License** — free for research, personal, and limited commercial use. See [`LICENSE`](LICENSE)
+for the full terms. Weights are released under the same license on Hugging Face.

turkish_stt-0.1.0/turkish_stt.egg-info/SOURCES.txt

@@ -0,0 +1,9 @@
+LICENSE
+README.md
+pyproject.toml
+mihu_stt/__init__.py
+turkish_stt.egg-info/PKG-INFO
+turkish_stt.egg-info/SOURCES.txt
+turkish_stt.egg-info/dependency_links.txt
+turkish_stt.egg-info/requires.txt
+turkish_stt.egg-info/top_level.txt

turkish_stt-0.1.0/turkish_stt.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

turkish_stt-0.1.0/turkish_stt.egg-info/requires.txt

@@ -0,0 +1,9 @@
+sherpa-onnx
+numpy
+soundfile
+soxr
+huggingface_hub
+
+[eval]
+jiwer
+num2words

turkish_stt-0.1.0/turkish_stt.egg-info/top_level.txt

@@ -0,0 +1 @@
+mihu_stt