Merge 6dda1a4f32 into 4a3a8f11a7

.dockerignore

@@ -0,0 +1,17 @@
+# The .dockerignore file excludes files from the container build process.
+#
+# https://docs.docker.com/engine/reference/builder/#dockerignore-file
+
+# Exclude Git files
+.git
+.github
+.gitignore
+
+# Exclude Python cache files
+__pycache__
+.mypy_cache
+.pytest_cache
+.ruff_cache
+
+# Exclude Python virtual environment
+/venv

.gitignore

@@ -29,4 +29,5 @@ src/audiocraft
 !/demo/
 !/demo/*
 /demo/temp/*.txt
-!/demo/temp/84_121550_000074_000000.txt
+!/demo/temp/84_121550_000074_000000.txt
+.cog/tmp/*

README.md

@@ -1,5 +1,6 @@
 # VoiceCraft: Zero-Shot Speech Editing and Text-to-Speech in the Wild
-[![Paper](https://img.shields.io/badge/arXiv-2301.12503-brightgreen.svg?style=flat-square)](https://jasonppy.github.io/assets/pdfs/VoiceCraft.pdf)  [![githubio](https://img.shields.io/badge/GitHub.io-Audio_Samples-blue?logo=Github&style=flat-square)](https://jasonppy.github.io/VoiceCraft_web/)  [![Hugging Face Spaces](https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Spaces-blue)](https://huggingface.co/spaces/pyp1/VoiceCraft_gradio)  [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1IOjpglQyMTO2C3Y94LD9FY0Ocn-RJRg6?usp=sharing)
+[![Paper](https://img.shields.io/badge/arXiv-2403.16973-brightgreen.svg?style=flat-square)](https://arxiv.org/pdf/2403.16973.pdf)  [![HuggingFace](https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Spaces-blue)](https://huggingface.co/spaces/pyp1/VoiceCraft_gradio)  [![Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1IOjpglQyMTO2C3Y94LD9FY0Ocn-RJRg6?usp=sharing)  [![Replicate](https://replicate.com/cjwbw/voicecraft/badge)](https://replicate.com/cjwbw/voicecraft)  [![YouTube demo](https://img.shields.io/youtube/views/eikybOi8iwU)](https://youtu.be/eikybOi8iwU)  [![Demo page](https://img.shields.io/badge/Audio_Samples-blue?logo=Github&style=flat-square)](https://jasonppy.github.io/VoiceCraft_web/)
+
 
 ### TL;DR
 VoiceCraft is a token infilling neural codec language model, that achieves state-of-the-art performance on both **speech editing** and **zero-shot text-to-speech (TTS)** on in-the-wild data including audiobooks, internet videos, and podcasts.
@@ -18,6 +19,8 @@ When you are inside the docker image or you have installed all dependencies, Che
 If you want to do model development such as training/finetuning, I recommend following [envrionment setup](#environment-setup) and [training](#training).
 
 ## News
+:star: 04/22/2024: 330M/830M TTS Enhanced Models are up [here](https://huggingface.co/pyp1), load them through [`gradio_app.py`](./gradio_app.py) or [`inference_tts.ipynb`](./inference_tts.ipynb)! Replicate demo is up, major thanks to [@chenxwh](https://github.com/chenxwh)!
+
 :star: 04/11/2024: VoiceCraft Gradio is now available on HuggingFace Spaces [here](https://huggingface.co/spaces/pyp1/VoiceCraft_gradio)! Major thanks to [@zuev-stepan](https://github.com/zuev-stepan), [@Sewlell](https://github.com/Sewlell), [@pgsoar](https://github.com/pgosar) [@Ph0rk0z](https://github.com/Ph0rk0z). 
 
 :star: 04/05/2024: I finetuned giga330M with the TTS objective on gigaspeech and 1/5 of librilight. Weights are [here](https://huggingface.co/pyp1/VoiceCraft/tree/main). Make sure maximal prompt + generation length <= 16 seconds (due to our limited compute, we had to drop utterances longer than 16s in training data). Even stronger models forthcomming, stay tuned!
@@ -30,7 +33,7 @@ If you want to do model development such as training/finetuning, I recommend fol
 - [x] Inference demo for speech editing and TTS
 - [x] Training guidance
 - [x] RealEdit dataset and training manifest
-- [x] Model weights (giga330M.pth, giga830M.pth, and gigaHalfLibri330M_TTSEnhanced_max16s.pth)
+- [x] Model weights
 - [x] Better guidance on training/finetuning
 - [x] Colab notebooks
 - [x] HuggingFace Spaces demo
@@ -210,7 +213,7 @@ We thank Feiteng for his [VALL-E reproduction](https://github.com/lifeiteng/vall
 ## Citation
 ```
 @article{peng2024voicecraft,
-  author    = {Peng, Puyuan and Huang, Po-Yao and Li, Daniel and Mohamed, Abdelrahman and Harwath, David},
+  author    = {Peng, Puyuan and Huang, Po-Yao and Mohamed, Abdelrahman and Harwath, David},
   title     = {VoiceCraft: Zero-Shot Speech Editing and Text-to-Speech in the Wild},
   journal   = {arXiv},
   year      = {2024},

cog.yaml

@@ -0,0 +1,24 @@
+# Configuration for Cog ⚙️
+# Reference: https://github.com/replicate/cog/blob/main/docs/yaml.md
+
+build:
+  gpu: true
+  system_packages:
+    - libgl1-mesa-glx
+    - libglib2.0-0
+    - ffmpeg
+    - espeak-ng
+  python_version: "3.11"
+  python_packages:
+    - torch==2.1.0
+    - torchaudio==2.1.0 
+    - xformers
+    - phonemizer==3.2.1
+    - whisperx==3.1.1
+    - openai-whisper>=20231117
+  run:
+    - git clone https://github.com/facebookresearch/audiocraft && pip install -e ./audiocraft
+    - pip install "pydantic<2.0.0"
+    - curl -o /usr/local/bin/pget -L "https://github.com/replicate/pget/releases/download/v0.6.0/pget_linux_x86_64" && chmod +x /usr/local/bin/pget
+    - mkdir -p /root/.cache/torch/hub/checkpoints/ && wget --output-document "/root/.cache/torch/hub/checkpoints/wav2vec2_fairseq_base_ls960_asr_ls960.pth" "https://download.pytorch.org/torchaudio/models/wav2vec2_fairseq_base_ls960_asr_ls960.pth"
+predict: "predict.py:Predictor"

gradio_app.py

@@ -1,4 +1,6 @@
 import os
+import re
+from num2words import num2words
 import gradio as gr
 import torch
 import torchaudio
@@ -83,7 +85,7 @@ def load_models(whisper_backend_name, whisper_model_name, alignment_model_name,
     elif voicecraft_model_name == "830M":
         voicecraft_model_name = "giga830M"
     elif voicecraft_model_name == "330M_TTSEnhanced":
-        voicecraft_model_name = "gigaHalfLibri330M_TTSEnhanced_max16s"
+        voicecraft_model_name = "330M_TTSEnhanced"
     elif voicecraft_model_name == "830M_TTSEnhanced":
         voicecraft_model_name = "830M_TTSEnhanced"
 
@@ -201,6 +203,15 @@ def get_output_audio(audio_tensors, codec_audio_sr):
     buffer.seek(0)
     return buffer.read()
 
+def replace_numbers_with_words(sentence):
+    sentence = re.sub(r'(\d+)', r' \1 ', sentence) # add spaces around numbers
+    def replace_with_words(match):
+        num = match.group(0)
+        try:
+            return num2words(num) # Convert numbers to words
+        except:
+            return num # In case num2words fails (unlikely with digits but just to be safe)
+    return re.sub(r'\b\d+\b', replace_with_words, sentence) # Regular expression that matches numbers
 
 def run(seed, left_margin, right_margin, codec_audio_sr, codec_sr, top_k, top_p, temperature,
         stop_repetition, sample_batch_size, kvcache, silence_tokens,
@@ -213,6 +224,8 @@ def run(seed, left_margin, right_margin, codec_audio_sr, codec_sr, top_k, top_p,
         raise gr.Error("Can't use smart transcript: whisper transcript not found")
 
     seed_everything(seed)
+    transcript = replace_numbers_with_words(transcript).replace("  ", " ").replace("  ", " ") # replace numbers with words, so that the phonemizer can do a better job
+
     if mode == "Long TTS":
         if split_text == "Newline":
             sentences = transcript.split('\n')

gradio_requirements.txt

@@ -4,3 +4,4 @@ openai-whisper>=20231117
 aeneas>=1.7.3.0
 whisperx>=3.1.1
 huggingface_hub==0.22.2
+num2words==0.5.13

inference_tts.ipynb

@@ -71,7 +71,7 @@
     "# load model, encodec, and phn2num\n",
     "# # load model, tokenizer, and other necessary files\n",
     "device = \"cuda\" if torch.cuda.is_available() else \"cpu\"\n",
-    "voicecraft_name=\"830M_TTSEnhanced.pth\" # or giga330M.pth, gigaHalfLibri330M_TTSEnhanced_max16s.pth, giga830M.pth\n",
+    "voicecraft_name=\"830M_TTSEnhanced.pth\" # or giga330M.pth, 330M_TTSEnhanced.pth, giga830M.pth\n",
     "\n",
     "# the new way of loading the model, with huggingface, recommended\n",
     "from models import voicecraft\n",

models/voicecraft.py

@@ -711,7 +711,7 @@ class VoiceCraft(
         ##################### silence repetition handling #####################
         # prepare the cache placeholder
         # n_layers, 2, bsz, num_heads, src_len, head_dim
-        past = torch.ones([self.args.num_decoder_layers, 2, x.shape[0]], device=x.device, dtype=torch.float32) if kvcache else None
+        past = torch.ones([self.args.num_decoder_layers, 2, x.shape[0]], device=x.device, dtype=torch.float16) if kvcache else None
         # handle multi-span kv-cache
         new_masked_span = False
         
@@ -1011,7 +1011,7 @@ class VoiceCraft(
 
         # prepare the cache placeholder
         # n_layers, 2, bsz, num_heads, src_len, head_dim
-        past = torch.ones([self.args.num_decoder_layers, 2, x.shape[0]], device=x.device, dtype=torch.float32) if kvcache else None
+        past = torch.ones([self.args.num_decoder_layers, 2, x.shape[0]], device=x.device, dtype=torch.float16) if kvcache else None
         # logging.info(f"number of decoder layers: {self.args.num_decoder_layers}")
         # logging.info(f"number of decoder layers: {self.args.num_decoder_layers}")
         # logging.info(f"number of decoder layers: {self.args.num_decoder_layers}")
@@ -1261,7 +1261,7 @@ class VoiceCraft(
 
         # prepare the cache placeholder
         # n_layers, 2, bsz, num_heads, src_len, head_dim
-        past = torch.ones([self.args.num_decoder_layers, 2, x.shape[0]], device=x.device, dtype=torch.float32) if kvcache else None
+        past = torch.ones([self.args.num_decoder_layers, 2, x.shape[0]], device=x.device, dtype=torch.float16) if kvcache else None
         # logging.info(f"number of decoder layers: {self.args.num_decoder_layers}")
         # logging.info(f"number of decoder layers: {self.args.num_decoder_layers}")
         # logging.info(f"number of decoder layers: {self.args.num_decoder_layers}")

predict.py

@@ -0,0 +1,389 @@
+# Prediction interface for Cog ⚙️
+# https://github.com/replicate/cog/blob/main/docs/python.md
+
+import os
+import time
+import random
+import getpass
+import shutil
+import subprocess
+import torch
+import numpy as np
+import torchaudio
+from cog import BasePredictor, Input, Path, BaseModel
+
+os.environ["USER"] = getpass.getuser()
+
+from data.tokenizer import (
+    AudioTokenizer,
+    TextTokenizer,
+)
+from models import voicecraft
+from inference_tts_scale import inference_one_sample
+from edit_utils import get_span
+from inference_speech_editing_scale import (
+    inference_one_sample as inference_one_sample_editing,
+)
+
+
+MODEL_URL = "https://weights.replicate.delivery/default/pyp1/VoiceCraft-models.tar"  # all the models are cached and uploaded to replicate.delivery for faster booting
+MODEL_CACHE = "model_cache"
+
+
+class ModelOutput(BaseModel):
+    whisper_transcript_orig_audio: str
+    generated_audio: Path
+
+
+class WhisperxAlignModel:
+    def __init__(self):
+        from whisperx import load_align_model
+
+        self.model, self.metadata = load_align_model(
+            language_code="en", device="cuda:0"
+        )
+
+    def align(self, segments, audio_path):
+        from whisperx import align, load_audio
+
+        audio = load_audio(audio_path)
+        return align(
+            segments,
+            self.model,
+            self.metadata,
+            audio,
+            device="cuda:0",
+            return_char_alignments=False,
+        )["segments"]
+
+
+class WhisperxModel:
+    def __init__(self, model_name, align_model: WhisperxAlignModel, device="cuda"):
+        from whisperx import load_model
+
+        # the model weights are cached from Systran/faster-whisper-base.en etc
+        self.model = load_model(
+            model_name,
+            device,
+            asr_options={
+                "suppress_numerals": True,
+                "max_new_tokens": None,
+                "clip_timestamps": None,
+                "hallucination_silence_threshold": None,
+            },
+        )
+        self.align_model = align_model
+
+    def transcribe(self, audio_path):
+        segments = self.model.transcribe(audio_path, language="en", batch_size=8)[
+            "segments"
+        ]
+        return self.align_model.align(segments, audio_path)
+
+
+def download_weights(url, dest):
+    start = time.time()
+    print("downloading url: ", url)
+    print("downloading to: ", dest)
+    subprocess.check_call(["pget", "-x", url, dest], close_fds=False)
+    print("downloading took: ", time.time() - start)
+
+
+class Predictor(BasePredictor):
+    def setup(self):
+        """Load the model into memory to make running multiple predictions efficient"""
+        self.device = "cuda"
+
+        if not os.path.exists(MODEL_CACHE):
+            download_weights(MODEL_URL, MODEL_CACHE)
+
+        encodec_fn = f"{MODEL_CACHE}/encodec_4cb2048_giga.th"
+        self.models, self.ckpt, self.phn2num = {}, {}, {}
+        for voicecraft_name in [
+            "giga830M.pth",
+            "giga330M.pth",
+            "gigaHalfLibri330M_TTSEnhanced_max16s.pth",
+        ]:
+            ckpt_fn = f"{MODEL_CACHE}/{voicecraft_name}"
+
+            self.ckpt[voicecraft_name] = torch.load(ckpt_fn, map_location="cpu")
+            self.models[voicecraft_name] = voicecraft.VoiceCraft(
+                self.ckpt[voicecraft_name]["config"]
+            )
+            self.models[voicecraft_name].load_state_dict(
+                self.ckpt[voicecraft_name]["model"]
+            )
+            self.models[voicecraft_name].to(self.device)
+            self.models[voicecraft_name].eval()
+
+            self.phn2num[voicecraft_name] = self.ckpt[voicecraft_name]["phn2num"]
+
+        self.text_tokenizer = TextTokenizer(backend="espeak")
+        self.audio_tokenizer = AudioTokenizer(signature=encodec_fn, device=self.device)
+
+        align_model = WhisperxAlignModel()
+        self.transcribe_models = {
+            k: WhisperxModel(f"{MODEL_CACHE}/whisperx_{k.split('.')[0]}", align_model)
+            for k in ["base.en", "small.en", "medium.en"]
+        }
+
+    def predict(
+        self,
+        task: str = Input(
+            description="Choose a task",
+            choices=[
+                "speech_editing-substitution",
+                "speech_editing-insertion",
+                "speech_editing-deletion",
+                "zero-shot text-to-speech",
+            ],
+            default="zero-shot text-to-speech",
+        ),
+        voicecraft_model: str = Input(
+            description="Choose a model",
+            choices=["giga830M.pth", "giga330M.pth", "giga330M_TTSEnhanced.pth"],
+            default="giga330M_TTSEnhanced.pth",
+        ),
+        orig_audio: Path = Input(description="Original audio file"),
+        orig_transcript: str = Input(
+            description="Optionally provide the transcript of the input audio. Leave it blank to use the WhisperX model below to generate the transcript. Inaccurate transcription may lead to error TTS or speech editing",
+            default="",
+        ),
+        whisperx_model: str = Input(
+            description="If orig_transcript is not provided above, choose a WhisperX model for generating the transcript. Inaccurate transcription may lead to error TTS or speech editing. You can modify the generated transcript and provide it directly to orig_transcript above",
+            choices=[
+                "base.en",
+                "small.en",
+                "medium.en",
+            ],
+            default="base.en",
+        ),
+        target_transcript: str = Input(
+            description="Transcript of the target audio file",
+        ),
+        cut_off_sec: float = Input(
+            description="Only used for for zero-shot text-to-speech task. The first seconds of the original audio that are used for zero-shot text-to-speech. 3 sec of reference is generally enough for high quality voice cloning, but longer is generally better, try e.g. 3~6 sec",
+            default=3.01,
+        ),
+        kvcache: int = Input(
+            description="Set to 0 to use less VRAM, but with slower inference",
+            choices=[0, 1],
+            default=1,
+        ),
+        left_margin: float = Input(
+            description="Margin to the left of the editing segment",
+            default=0.08,
+        ),
+        right_margin: float = Input(
+            description="Margin to the right of the editing segment",
+            default=0.08,
+        ),
+        temperature: float = Input(
+            description="Adjusts randomness of outputs, greater than 1 is random and 0 is deterministic. Do not recommend to change",
+            default=1,
+        ),
+        top_p: float = Input(
+            description="Default value for TTS is 0.9, and 0.8 for speech editing",
+            default=0.9,
+        ),
+        stop_repetition: int = Input(
+            default=3,
+            description="Default value for TTS is 3, and -1 for speech editing. -1 means do not adjust prob of silence tokens. if there are long silence or unnaturally stretched words, increase sample_batch_size to 2, 3 or even 4",
+        ),
+        sample_batch_size: int = Input(
+            description="Default value for TTS is 4, and 1 for speech editing. The higher the number, the faster the output will be. Under the hood, the model will generate this many samples and choose the shortest one",
+            default=4,
+        ),
+        seed: int = Input(
+            description="Random seed. Leave blank to randomize the seed", default=None
+        ),
+    ) -> ModelOutput:
+        """Run a single prediction on the model"""
+
+        if seed is None:
+            seed = int.from_bytes(os.urandom(2), "big")
+        print(f"Using seed: {seed}")
+
+        seed_everything(seed)
+
+        segments = self.transcribe_models[whisperx_model].transcribe(
+            str(orig_audio)
+        )
+
+        state = get_transcribe_state(segments)
+
+        whisper_transcript = state["transcript"].strip()
+
+        if len(orig_transcript.strip()) == 0:
+            orig_transcript = whisper_transcript
+
+        print(f"The transcript from the Whisper model: {whisper_transcript}")
+
+        temp_folder = "exp_dir"
+        if os.path.exists(temp_folder):
+            shutil.rmtree(temp_folder)
+
+        os.makedirs(temp_folder)
+
+        filename = "orig_audio"
+        audio_fn = str(orig_audio)
+
+        info = torchaudio.info(audio_fn)
+        audio_dur = info.num_frames / info.sample_rate
+
+        # hyperparameters for inference
+        codec_audio_sr = 16000
+        codec_sr = 50
+        top_k = 0
+        silence_tokens = [1388, 1898, 131]
+
+        if voicecraft_model == "giga330M_TTSEnhanced.pth":
+            voicecraft_model = "gigaHalfLibri330M_TTSEnhanced_max16s.pth"
+
+        if task == "zero-shot text-to-speech":
+            assert (
+                cut_off_sec < audio_dur
+            ), f"cut_off_sec {cut_off_sec} is larger than the audio duration {audio_dur}"
+            prompt_end_frame = int(cut_off_sec * info.sample_rate)
+
+            idx = find_closest_cut_off_word(state["word_bounds"], cut_off_sec)
+            orig_transcript_until_cutoff_time = " ".join(
+                [word_bound["word"] for word_bound in state["word_bounds"][: idx + 1]]
+            )
+        else:
+            edit_type = task.split("-")[-1]
+            orig_span, new_span = get_span(
+                orig_transcript, target_transcript, edit_type
+            )
+            if orig_span[0] > orig_span[1]:
+                RuntimeError(f"example {audio_fn} failed")
+            if orig_span[0] == orig_span[1]:
+                orig_span_save = [orig_span[0]]
+            else:
+                orig_span_save = orig_span
+            if new_span[0] == new_span[1]:
+                new_span_save = [new_span[0]]
+            else:
+                new_span_save = new_span
+            orig_span_save = ",".join([str(item) for item in orig_span_save])
+            new_span_save = ",".join([str(item) for item in new_span_save])
+
+            start, end = get_mask_interval_from_word_bounds(
+                state["word_bounds"], orig_span_save, edit_type
+            )
+
+            # span in codec frames
+            morphed_span = (
+                max(start - left_margin, 1 / codec_sr),
+                min(end + right_margin, audio_dur),
+            )  # in seconds
+            mask_interval = [
+                [round(morphed_span[0] * codec_sr), round(morphed_span[1] * codec_sr)]
+            ]
+            mask_interval = torch.LongTensor(mask_interval)  # [M,2], M==1 for now
+
+        decode_config = {
+            "top_k": top_k,
+            "top_p": top_p,
+            "temperature": temperature,
+            "stop_repetition": stop_repetition,
+            "kvcache": kvcache,
+            "codec_audio_sr": codec_audio_sr,
+            "codec_sr": codec_sr,
+            "silence_tokens": silence_tokens,
+        }
+
+        if task == "zero-shot text-to-speech":
+            decode_config["sample_batch_size"] = sample_batch_size
+            _, gen_audio = inference_one_sample(
+                self.models[voicecraft_model],
+                self.ckpt[voicecraft_model]["config"],
+                self.phn2num[voicecraft_model],
+                self.text_tokenizer,
+                self.audio_tokenizer,
+                audio_fn,
+                orig_transcript_until_cutoff_time.strip()
+                + " "
+                + target_transcript.strip(),
+                self.device,
+                decode_config,
+                prompt_end_frame,
+            )
+        else:
+            _, gen_audio = inference_one_sample_editing(
+                self.models[voicecraft_model],
+                self.ckpt[voicecraft_model]["config"],
+                self.phn2num[voicecraft_model],
+                self.text_tokenizer,
+                self.audio_tokenizer,
+                audio_fn,
+                target_transcript,
+                mask_interval,
+                self.device,
+                decode_config,
+            )
+
+        # save segments for comparison
+        gen_audio = gen_audio[0].cpu()
+
+        out = "/tmp/out.wav"
+        torchaudio.save(out, gen_audio, codec_audio_sr)
+        return ModelOutput(
+            generated_audio=Path(out), whisper_transcript_orig_audio=whisper_transcript
+        )
+
+
+def seed_everything(seed):
+    os.environ["PYTHONHASHSEED"] = str(seed)
+    random.seed(seed)
+    np.random.seed(seed)
+    torch.manual_seed(seed)
+    torch.cuda.manual_seed(seed)
+    torch.backends.cudnn.benchmark = False
+    torch.backends.cudnn.deterministic = True
+
+
+def get_transcribe_state(segments):
+    words_info = [word_info for segment in segments for word_info in segment["words"]]
+    return {
+        "transcript": " ".join([segment["text"].strip() for segment in segments]),
+        "word_bounds": [
+            {"word": word["word"], "start": word["start"], "end": word["end"]}
+            for word in words_info
+        ],
+    }
+
+
+def find_closest_cut_off_word(word_bounds, cut_off_sec):
+    min_distance = float("inf")
+
+    for i, word_bound in enumerate(word_bounds):
+        distance = abs(word_bound["start"] - cut_off_sec)
+
+        if distance < min_distance:
+            min_distance = distance
+
+        if word_bound["end"] > cut_off_sec:
+            break
+
+    return i
+
+
+def get_mask_interval_from_word_bounds(word_bounds, word_span_ind, editType):
+    tmp = word_span_ind.split(",")
+    s, e = int(tmp[0]), int(tmp[-1])
+    start = None
+    for j, item in enumerate(word_bounds):
+        if j == s:
+            if editType == "insertion":
+                start = float(item["end"])
+            else:
+                start = float(item["start"])
+        if j == e:
+            if editType == "insertion":
+                end = float(item["start"])
+            else:
+                end = float(item["end"])
+            assert start is not None
+            break
+    return (start, end)