Link to colab notebook via github

.pre-commit-config.yaml

@@ -0,0 +1,22 @@
+repos:
+    # Get rid of Jupyter Notebook output because we don't want to keep it in Git
+  - repo: https://github.com/kynan/nbstripout
+    rev: 0.8.1
+    hooks:
+      - id: nbstripout
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v5.0.0
+    hooks:
+      - id: check-added-large-files
+        args: ["--maxkb=2048"]
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    # Ruff version.
+    rev: v0.11.7
+    hooks:
+      # Run the linter.
+      - id: ruff
+        types_or: [python, pyi] # Don't run on `jupyter` files
+        args: [--fix]
+      # Run the formatter.
+      - id: ruff-format
+        types_or: [python, pyi] # Don't run on `jupyter` files

README.md

@@ -1,7 +1,7 @@
 <a href="https://huggingface.co/collections/kyutai/speech-to-text-685403682cf8a23ab9466886" target="_blank" style="margin: 2px;">
     <img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-KyutaiSTT-blue" style="display: inline-block; vertical-align: middle;"/>
 </a>
-<a target="_blank" href="https://colab.research.google.com/drive/1mc0Q-FoHxU2pEvId8rTdS4q1r1zorJhS?usp=sharing">
+<a target="_blank" href="https://colab.research.google.com/github/kyutai-labs/delayed-streams-modeling/blob/main/transcribe_via_pytorch.ipynb">
   <img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/>
 </a>
 
@@ -33,6 +33,21 @@ These speech-to-text models have several advantages:
   can be used to detect when the user is speaking. This is especially useful
   for building voice agents.
 
+### Implementations overview
+
+We provide different implementations of Kyutai STT for different use cases.
+Here is how to choose which one to use:
+
+- **PyTorch: for research and tinkering.**
+  If you want to call the model from Python for research or experimentation, use our PyTorch implementation.
+- **Rust: for production.**
+  If you want to serve Kyutai STT in a production setting, use our Rust server.
+  Our robust Rust server provides streaming access to the model over websockets.
+  We use this server to run [Unmute](https://unmute.sh/); on a L40S GPU, we can serve 64 simultaneous connections at a real-time factor of 3x.
+- **MLX: for on-device inference on iPhone and Mac.**
+  MLX is Apple's ML framework that allows you to use hardware acceleration on Apple silicon.
+  If you want to run the model on a Mac or an iPhone, choose the MLX implementation.
+
 You can retrieve the sample files used in the following snippets via:
 ```bash
 wget https://github.com/metavoiceio/metavoice-src/raw/main/assets/bria.mp3
@@ -43,10 +58,14 @@ wget https://github.com/kyutai-labs/moshi/raw/refs/heads/main/data/sample_fr_hib
 <a href="https://huggingface.co/kyutai/stt-2.6b-en" target="_blank" style="margin: 2px;">
     <img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Model-blue" style="display: inline-block; vertical-align: middle;"/>
 </a>
-<a target="_blank" href="https://colab.research.google.com/drive/1mc0Q-FoHxU2pEvId8rTdS4q1r1zorJhS?usp=sharing">
+<a target="_blank" href="https://colab.research.google.com/github/kyutai-labs/delayed-streams-modeling/blob/main/transcribe_via_pytorch.ipynb">
   <img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/>
 </a>
 
+For an example of how to use the model in a way where you can directly stream in PyTorch tensors,
+[see our Colab notebook](https://colab.research.google.com/github/kyutai-labs/delayed-streams-modeling/blob/main/transcribe_via_pytorch.ipynb).
+
+If you just want to run the model on a file, you can use `moshi.run_inference`.
 This requires the [moshi package](https://pypi.org/project/moshi/)
 with version 0.2.6 or later, which can be installed via pip.
 
@@ -58,25 +77,25 @@ If you have [uv](https://docs.astral.sh/uv/) installed, you can skip the install
 ```bash
 uvx --with moshi python -m moshi.run_inference --hf-repo kyutai/stt-2.6b-en bria.mp3
 ```
-It will install the moshi package in a temporary environment and run the speech-to-text.
 
 Additionally, we provide two scripts that highlight different usage scenarios. The first script illustrates how to extract word-level timestamps from the model's outputs:
 
 ```bash
 uv run \
-  scripts/streaming_stt_timestamps.py \
+  scripts/transcribe_from_file_via_pytorch.py \
   --hf-repo kyutai/stt-2.6b-en \
   --file bria.mp3
 ```
 
 The second script can be used to run a model on an existing Hugging Face dataset and calculate its performance metrics: 
 ```bash
-uv run scripts/streaming_stt.py  \
+uv run scripts/evaluate_on_dataset.py  \
   --dataset meanwhile  \
   --hf-repo kyutai/stt-2.6b-en
 ```
 
 ### Rust server
+
 <a href="https://huggingface.co/kyutai/stt-2.6b-en-candle" target="_blank" style="margin: 2px;">
     <img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Model-blue" style="display: inline-block; vertical-align: middle;"/>
 </a>
@@ -104,15 +123,19 @@ and for `kyutai/stt-2.6b-en`, use `configs/config-stt-en-hf.toml`,
 moshi-server worker --config configs/config-stt-en_fr-hf.toml
 ```
 
-Once the server has started you can run a streaming inference with the following
-script.
+Once the server has started you can transcribe audio from your microphone with the following script.
 ```bash
-uv run scripts/asr-streaming-query.py bria.mp3
+uv run scripts/transcribe_from_mic_via_rust_server.py
+```
+
+We also provide a script for transcribing from an audio file.
+```bash
+uv run scripts/transcribe_from_file_via_rust_server.py bria.mp3
 ```
 
 The script limits the decoding speed to simulates real-time processing of the audio. 
 Faster processing can be triggered by setting 
-the real-time factor, e.g. `--rtf 500` will process
+the real-time factor, e.g. `--rtf 1000` will process
 the data as fast as possible.
 
 ### Rust standalone
@@ -166,3 +189,14 @@ Note that parts of this code is based on [AudioCraft](https://github.com/faceboo
 the MIT license.
 
 The weights for the speech-to-text models are released under the CC-BY 4.0 license.
+
+## Developing
+
+Install the [pre-commit hooks](https://pre-commit.com/) by running:
+
+```bash
+pip install pre-commit
+pre-commit install
+```
+
+If you're using `uv`, you can replace the two commands with `uvx pre-commit install`.

scripts/evaluate_on_dataset.py

@@ -41,18 +41,17 @@ uv run scripts/streaming_stt.py \
 # Rev16 === cer: 6.57% wer: 10.08% corpus_wer: 11.43% RTF = 40.34
 # Earnings21 === cer: 5.73% wer: 9.84% corpus_wer: 10.38% RTF = 73.15
 
-import dataclasses
-import julius
-import jiwer
-from datasets import load_dataset, Dataset
-from whisper.normalizers import EnglishTextNormalizer
 import argparse
-
-import torch
-import moshi.models
-import tqdm
+import dataclasses
 import time
 
+import jiwer
+import julius
+import moshi.models
+import torch
+import tqdm
+from datasets import Dataset, load_dataset
+from whisper.normalizers import EnglishTextNormalizer
 
 _NORMALIZER = EnglishTextNormalizer()
 
@@ -120,9 +119,9 @@ class AsrMetrics:
         self.num_sequences += 1
 
     def compute(self) -> dict:
-        assert (
-            self.num_sequences > 0
-        ), "Unable to compute with total number of comparisons <= 0"  # type: ignore
+        assert self.num_sequences > 0, (
+            "Unable to compute with total number of comparisons <= 0"
+        )  # type: ignore
         return {
             "cer": (self.cer_sum / self.num_sequences),
             "wer": (self.wer_sum / self.num_sequences),

scripts/transcribe_from_file_via_pytorch.py

@@ -19,15 +19,15 @@ uv run scripts/streaming_stt_timestamps.py \
 ```
 """
 
-import itertools
-import dataclasses
-import julius
-import sphn
 import argparse
+import dataclasses
+import itertools
 import math
 
-import torch
+import julius
 import moshi.models
+import sphn
+import torch
 import tqdm
 
 

scripts/transcribe_from_file_via_rust_server.py

@@ -10,17 +10,16 @@
 import argparse
 import asyncio
 import json
-import msgpack
-import sphn
-import struct
 import time
 
-import numpy as np
+import msgpack
+import sphn
 import websockets
 
 # Desired audio properties
 TARGET_SAMPLE_RATE = 24000
 TARGET_CHANNELS = 1  # Mono
+HEADERS = {"kyutai-api-key": "open_token"}
 all_text = []
 transcript = []
 finished = False
@@ -44,11 +43,13 @@ async def receive_messages(websocket):
             print("received:", data)
             if data["type"] == "Word":
                 all_text.append(data["text"])
-                transcript.append({
+                transcript.append(
+                    {
                         "speaker": "SPEAKER_00",
                         "text": data["text"],
                         "timestamp": [data["start_time"], data["start_time"]],
-                })
+                    }
+                )
             if data["type"] == "EndWord":
                 if len(transcript) > 0:
                     transcript[-1]["timestamp"][1] = data["stop_time"]
@@ -64,15 +65,19 @@ async def send_messages(websocket, rtf: float):
     global finished
     audio_data = load_and_process_audio(args.in_file)
     try:
-        # Start with a second of silence
-        chunk = { "type": "Audio", "pcm": [0.0] * 24000 }
+        # Start with a second of silence.
+        # This is needed for the 2.6B model for technical reasons.
+        chunk = {"type": "Audio", "pcm": [0.0] * 24000}
         msg = msgpack.packb(chunk, use_bin_type=True, use_single_float=True)
         await websocket.send(msg)
 
         chunk_size = 1920  # Send data in chunks
         start_time = time.time()
         for i in range(0, len(audio_data), chunk_size):
-            chunk = { "type": "Audio", "pcm": [float(x) for x in audio_data[i : i + chunk_size]] }
+            chunk = {
+                "type": "Audio",
+                "pcm": [float(x) for x in audio_data[i : i + chunk_size]],
+            }
             msg = msgpack.packb(chunk, use_bin_type=True, use_single_float=True)
             await websocket.send(msg)
             expected_send_time = start_time + (i + 1) / 24000 / rtf
@@ -81,13 +86,15 @@ async def send_messages(websocket, rtf: float):
                 await asyncio.sleep(expected_send_time - current_time)
             else:
                 await asyncio.sleep(0.001)
-        chunk = { "type": "Audio", "pcm": [0.0] * 1920 * 5 }
+        chunk = {"type": "Audio", "pcm": [0.0] * 1920 * 5}
         msg = msgpack.packb(chunk, use_bin_type=True, use_single_float=True)
         await websocket.send(msg)
-        msg = msgpack.packb({"type": "Marker", "id": 0}, use_bin_type=True, use_single_float=True)
+        msg = msgpack.packb(
+            {"type": "Marker", "id": 0}, use_bin_type=True, use_single_float=True
+        )
         await websocket.send(msg)
         for _ in range(35):
-            chunk = { "type": "Audio", "pcm": [0.0] * 1920 }
+            chunk = {"type": "Audio", "pcm": [0.0] * 1920}
             msg = msgpack.packb(chunk, use_bin_type=True, use_single_float=True)
             await websocket.send(msg)
         while True:
@@ -100,11 +107,10 @@ async def send_messages(websocket, rtf: float):
         print("Connection closed while sending messages.")
 
 
-async def stream_audio(url: str, rtf: float, api_key: str):
+async def stream_audio(url: str, rtf: float):
     """Stream audio data to a WebSocket server."""
 
-    headers = {"kyutai-api-key": api_key}
-    async with websockets.connect(url, additional_headers=headers) as websocket:
+    async with websockets.connect(url, additional_headers=HEADERS) as websocket:
         send_task = asyncio.create_task(send_messages(websocket, rtf))
         receive_task = asyncio.create_task(receive_messages(websocket))
         await asyncio.gather(send_task, receive_task)
@@ -115,7 +121,6 @@ if __name__ == "__main__":
     parser = argparse.ArgumentParser()
     parser.add_argument("in_file")
     parser.add_argument("--transcript")
-    parser.add_argument("--api-key", default="open_token")
     parser.add_argument(
         "--url",
         help="The url of the server to which to send the audio",
@@ -125,7 +130,7 @@ if __name__ == "__main__":
     args = parser.parse_args()
 
     url = f"{args.url}/api/asr-streaming"
-    asyncio.run(stream_audio(url, args.rtf, args.api_key))
+    asyncio.run(stream_audio(url, args.rtf))
     print(" ".join(all_text))
     if args.transcript is not None:
         with open(args.transcript, "w") as fobj:

scripts/transcribe_from_mic_via_mlx.py

@@ -11,19 +11,16 @@
 # ///
 
 import argparse
-from dataclasses import dataclass
 import json
-import numpy as np
 import queue
-import sounddevice as sd
 
-from huggingface_hub import hf_hub_download
 import mlx.core as mx
 import mlx.nn as nn
-from moshi_mlx import models, utils
 import rustymimi
 import sentencepiece
-
+import sounddevice as sd
+from huggingface_hub import hf_hub_download
+from moshi_mlx import models, utils
 
 if __name__ == "__main__":
     parser = argparse.ArgumentParser()
@@ -69,6 +66,7 @@ if __name__ == "__main__":
     )
 
     block_queue = queue.Queue()
+
     def audio_callback(indata, _frames, _time, _status):
         block_queue.put(indata.copy())
 
@@ -84,7 +82,9 @@ if __name__ == "__main__":
             block = block_queue.get()
             block = block[None, :, 0]
             other_audio_tokens = audio_tokenizer.encode_step(block[None, 0:1])
-            other_audio_tokens = mx.array(other_audio_tokens).transpose(0, 2, 1)[:, :, :other_codebooks]
+            other_audio_tokens = mx.array(other_audio_tokens).transpose(0, 2, 1)[
+                :, :, :other_codebooks
+            ]
             text_token = gen.step(other_audio_tokens[0])
             text_token = text_token[0].item()
             audio_tokens = gen.last_audio_tokens()
@@ -93,4 +93,3 @@ if __name__ == "__main__":
                 _text = text_tokenizer.id_to_piece(text_token)  # type: ignore
                 _text = _text.replace("▁", " ")
                 print(_text, end="", flush=True)
- 

scripts/transcribe_from_mic_via_rust_server.py

@@ -9,9 +9,9 @@
 # ///
 import argparse
 import asyncio
-import msgpack
 import signal
 
+import msgpack
 import numpy as np
 import sounddevice as sd
 import websockets
@@ -21,6 +21,7 @@ TARGET_SAMPLE_RATE = 24000
 TARGET_CHANNELS = 1  # Mono
 audio_queue = asyncio.Queue()
 
+
 async def receive_messages(websocket):
     """Receive and process messages from the WebSocket server."""
     try:
@@ -47,22 +48,26 @@ async def send_messages(websocket):
     except websockets.ConnectionClosed:
         print("Connection closed while sending messages.")
 
+
 async def stream_audio(url: str, api_key: str):
     """Stream audio data to a WebSocket server."""
     print("Starting microphone recording...")
     print("Press Ctrl+C to stop recording")
 
     loop = asyncio.get_event_loop()
+
     def audio_callback(indata, frames, time, status):
-        loop.call_soon_threadsafe(audio_queue.put_nowait, indata[:, 0].astype(np.float32).copy())
+        loop.call_soon_threadsafe(
+            audio_queue.put_nowait, indata[:, 0].astype(np.float32).copy()
+        )
 
     # Start audio stream
     with sd.InputStream(
         samplerate=TARGET_SAMPLE_RATE,
         channels=TARGET_CHANNELS,
-        dtype='float32',
+        dtype="float32",
         callback=audio_callback,
-        blocksize=1920  # 80ms blocks
+        blocksize=1920,  # 80ms blocks
     ):
         headers = {"kyutai-api-key": api_key}
         async with websockets.connect(url, additional_headers=headers) as websocket:
@@ -79,8 +84,12 @@ if __name__ == "__main__":
         default="ws://127.0.0.1:8080",
     )
     parser.add_argument("--api-key", default="open_token")
-    parser.add_argument("--list-devices", action="store_true", help="List available audio devices")
-    parser.add_argument("--device", type=int, help="Input device ID (use --list-devices to see options)")
+    parser.add_argument(
+        "--list-devices", action="store_true", help="List available audio devices"
+    )
+    parser.add_argument(
+        "--device", type=int, help="Input device ID (use --list-devices to see options)"
+    )
 
     args = parser.parse_args()
 

transcribe_via_pytorch.ipynb

@@ -0,0 +1,240 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "gJEMjPgeI-rw",
+    "outputId": "7491c067-b1be-4505-b3f5-19ba4c00a593"
+   },
+   "outputs": [],
+   "source": [
+    "!pip install moshi"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "CA4K5iDFJcqJ",
+    "outputId": "b609843a-a193-4729-b099-5f8780532333"
+   },
+   "outputs": [],
+   "source": [
+    "!wget https://github.com/kyutai-labs/moshi/raw/refs/heads/main/data/sample_fr_hibiki_crepes.mp3"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "id": "VA3Haix3IZ8Q"
+   },
+   "outputs": [],
+   "source": [
+    "from dataclasses import dataclass\n",
+    "import time\n",
+    "import sentencepiece\n",
+    "import sphn\n",
+    "import textwrap\n",
+    "import torch\n",
+    "\n",
+    "from moshi.models import loaders, MimiModel, LMModel, LMGen"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "id": "9AK5zBMTI9bw"
+   },
+   "outputs": [],
+   "source": [
+    "@dataclass\n",
+    "class InferenceState:\n",
+    "    mimi: MimiModel\n",
+    "    text_tokenizer: sentencepiece.SentencePieceProcessor\n",
+    "    lm_gen: LMGen\n",
+    "\n",
+    "    def __init__(\n",
+    "        self,\n",
+    "        mimi: MimiModel,\n",
+    "        text_tokenizer: sentencepiece.SentencePieceProcessor,\n",
+    "        lm: LMModel,\n",
+    "        batch_size: int,\n",
+    "        device: str | torch.device,\n",
+    "    ):\n",
+    "        self.mimi = mimi\n",
+    "        self.text_tokenizer = text_tokenizer\n",
+    "        self.lm_gen = LMGen(lm, temp=0, temp_text=0, use_sampling=False)\n",
+    "        self.device = device\n",
+    "        self.frame_size = int(self.mimi.sample_rate / self.mimi.frame_rate)\n",
+    "        self.batch_size = batch_size\n",
+    "        self.mimi.streaming_forever(batch_size)\n",
+    "        self.lm_gen.streaming_forever(batch_size)\n",
+    "\n",
+    "    def run(self, in_pcms: torch.Tensor):\n",
+    "        device = self.lm_gen.lm_model.device\n",
+    "        ntokens = 0\n",
+    "        first_frame = True\n",
+    "        chunks = [\n",
+    "            c\n",
+    "            for c in in_pcms.split(self.frame_size, dim=2)\n",
+    "            if c.shape[-1] == self.frame_size\n",
+    "        ]\n",
+    "        start_time = time.time()\n",
+    "        all_text = []\n",
+    "        for chunk in chunks:\n",
+    "            codes = self.mimi.encode(chunk)\n",
+    "            if first_frame:\n",
+    "                # Ensure that the first slice of codes is properly seen by the transformer\n",
+    "                # as otherwise the first slice is replaced by the initial tokens.\n",
+    "                tokens = self.lm_gen.step(codes)\n",
+    "                first_frame = False\n",
+    "            tokens = self.lm_gen.step(codes)\n",
+    "            if tokens is None:\n",
+    "                continue\n",
+    "            assert tokens.shape[1] == 1\n",
+    "            one_text = tokens[0, 0].cpu()\n",
+    "            if one_text.item() not in [0, 3]:\n",
+    "                text = self.text_tokenizer.id_to_piece(one_text.item())\n",
+    "                text = text.replace(\"▁\", \" \")\n",
+    "                all_text.append(text)\n",
+    "            ntokens += 1\n",
+    "        dt = time.time() - start_time\n",
+    "        print(\n",
+    "            f\"processed {ntokens} steps in {dt:.0f}s, {1000 * dt / ntokens:.2f}ms/step\"\n",
+    "        )\n",
+    "        return \"\".join(all_text)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/",
+     "height": 353,
+     "referenced_widgets": [
+      "0a5f6f887e2b4cd1990a0e9ec0153ed9",
+      "f7893826fcba4bdc87539589d669249b",
+      "8805afb12c484781be85082ff02dad13",
+      "97679c0d9ab44bed9a3456f2fcb541fd",
+      "d73c0321bed54a52b5e1da0a7788e32a",
+      "d67be13a920d4fc89e5570b5b29fc1d2",
+      "6b377c2d7bf945fb89e46c39d246a332",
+      "b82ff365c78e41ad8094b46daf79449d",
+      "477aa7fa82dc42d5bce6f1743c45d626",
+      "cbd288510c474430beb66f346f382c45",
+      "aafc347cdf28428ea6a7abe5b46b726f",
+      "fca09acd5d0d45468c8b04bfb2de7646",
+      "79e35214b51b4a9e9b3f7144b0b34f7b",
+      "89e9a37f69904bd48b954d627bff6687",
+      "57028789c78248a7b0ad4f031c9545c9",
+      "1150fcb427994c2984d4d0f4e4745fe5",
+      "e24b1fc52f294f849019c9b3befb613f",
+      "8724878682cf4c3ca992667c45009398",
+      "36a22c977d5242008871310133b7d2af",
+      "5b3683cad5cb4877b43fadd003edf97f",
+      "703f98272e4d469d8f27f5a465715dd8",
+      "9dbe02ef5fac41cfaee3d02946e65c88",
+      "37faa87ad03a4271992c21ce6a629e18",
+      "570c547e48cd421b814b2c5e028e4c0b",
+      "b173768580fc4c0a8e3abf272e4c363a",
+      "e57d1620f0a9427b85d8b4885ef4e8e3",
+      "5dd4474df70743498b616608182714dd",
+      "cc907676a65f4ad1bf68a77b4a00e89b",
+      "a34abc3b118e4305951a466919c28ff6",
+      "a77ccfcdb90146c7a63b4b2d232bc494",
+      "f7313e6e3a27475993cab3961d6ae363",
+      "39b47fad9c554839868fe9e4bbf7def2",
+      "14e9511ea0bd44c49f0cf3abf1a6d40e",
+      "a4ea8e0c4cac4d5e88b7e3f527e4fe90",
+      "571afc0f4b2840c9830d6b5a307ed1f9",
+      "6ec593cab5b64f0ea638bb175b9daa5c",
+      "77a52aed00ae408bb24524880e19ec8a",
+      "0b2de4b29b4b44fe9d96361a40c793d0",
+      "3c5b5fb1a5ac468a89c1058bd90cfb58",
+      "e53e0a2a240e43cfa562c89b3d703dea",
+      "35966343cf9249ef8bc028a0d5c5f97d",
+      "e36a37e0d41c47ccb8bc6d56c19fb17c",
+      "279ccf7de43847a1a6579c9182a46cc8",
+      "41b5d6ab0b7d43c790a55f125c0e7494"
+     ]
+    },
+    "id": "UsQJdAgkLp9n",
+    "outputId": "9b7131c3-69c5-4323-8312-2ce7621d8869"
+   },
+   "outputs": [],
+   "source": [
+    "device = \"cuda\"\n",
+    "# Use the en+fr low latency model, an alternative is kyutai/stt-2.6b-en\n",
+    "checkpoint_info = loaders.CheckpointInfo.from_hf_repo(\"kyutai/stt-1b-en_fr\")\n",
+    "mimi = checkpoint_info.get_mimi(device=device)\n",
+    "text_tokenizer = checkpoint_info.get_text_tokenizer()\n",
+    "lm = checkpoint_info.get_moshi(device=device)\n",
+    "in_pcms, _ = sphn.read(\"sample_fr_hibiki_crepes.mp3\", sample_rate=mimi.sample_rate)\n",
+    "in_pcms = torch.from_numpy(in_pcms).to(device=device)\n",
+    "\n",
+    "stt_config = checkpoint_info.stt_config\n",
+    "pad_left = int(stt_config.get(\"audio_silence_prefix_seconds\", 0.0) * 24000)\n",
+    "pad_right = int((stt_config.get(\"audio_delay_seconds\", 0.0) + 1.0) * 24000)\n",
+    "in_pcms = torch.nn.functional.pad(in_pcms, (pad_left, pad_right), mode=\"constant\")\n",
+    "in_pcms = in_pcms[None, 0:1].expand(1, -1, -1)\n",
+    "\n",
+    "state = InferenceState(mimi, text_tokenizer, lm, batch_size=1, device=device)\n",
+    "text = state.run(in_pcms)\n",
+    "print(textwrap.fill(text, width=100))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/",
+     "height": 75
+    },
+    "id": "CIAXs9oaPrtj",
+    "outputId": "94cc208c-2454-4dd4-a64e-d79025144af5"
+   },
+   "outputs": [],
+   "source": [
+    "from IPython.display import Audio\n",
+    "\n",
+    "Audio(\"sample_fr_hibiki_crepes.mp3\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "id": "qkUZ6CBKOdTa"
+   },
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "accelerator": "GPU",
+  "colab": {
+   "gpuType": "L4",
+   "provenance": []
+  },
+  "kernelspec": {
+   "display_name": "Python 3",
+   "name": "python3"
+  },
+  "language_info": {
+   "name": "python"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 0
+}