upload seamless files

Dockerfile

@@ -0,0 +1,56 @@
+FROM nvidia/cuda:11.7.1-cudnn8-devel-ubuntu22.04
+ENV DEBIAN_FRONTEND=noninteractive
+RUN apt-get update && \
+    apt-get upgrade -y && \
+    apt-get install -y --no-install-recommends \
+    git \
+    git-lfs \
+    wget \
+    curl \
+    # python build dependencies \
+    build-essential \
+    libssl-dev \
+    zlib1g-dev \
+    libbz2-dev \
+    libreadline-dev \
+    libsqlite3-dev \
+    libncursesw5-dev \
+    xz-utils \
+    tk-dev \
+    libxml2-dev \
+    libxmlsec1-dev \
+    libffi-dev \
+    liblzma-dev \
+    # gradio dependencies \
+    ffmpeg \
+    # fairseq2 dependencies \
+    libsndfile-dev && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/*
+
+RUN useradd -m -u 1000 user
+USER user
+ENV HOME=/home/user \
+    PATH=/home/user/.local/bin:${PATH}
+WORKDIR ${HOME}/app
+
+RUN curl https://pyenv.run | bash
+ENV PATH=${HOME}/.pyenv/shims:${HOME}/.pyenv/bin:${PATH}
+ARG PYTHON_VERSION=3.10.12
+RUN pyenv install ${PYTHON_VERSION} && \
+    pyenv global ${PYTHON_VERSION} && \
+    pyenv rehash && \
+    pip install --no-cache-dir -U pip setuptools wheel
+
+COPY --chown=1000 ./requirements.txt /tmp/requirements.txt
+RUN pip install --no-cache-dir --upgrade -r /tmp/requirements.txt
+
+COPY --chown=1000 . ${HOME}/app
+ENV PYTHONPATH=${HOME}/app \
+    PYTHONUNBUFFERED=1 \
+    GRADIO_ALLOW_FLAGGING=never \
+    GRADIO_NUM_PORTS=1 \
+    GRADIO_SERVER_NAME=0.0.0.0 \
+    GRADIO_THEME=huggingface \
+    SYSTEM=spaces
+CMD ["python", "app.py"]

app.py

@@ -0,0 +1,140 @@
+from __future__ import annotations
+
+import os
+
+import gradio as gr
+import numpy as np
+import torch
+import torchaudio
+from seamless_communication.models.inference import Translator
+
+from lang_list import LANGUAGE_NAME_TO_CODE, S2TT_TARGET_LANGUAGE_NAMES 
+
+
+DESCRIPTION = """# Speech to Text Translation 
+
+[SeamlessM4T](https://github.com/facebookresearch/seamless_communication) is designed to provide high-quality
+translation, allowing people from different linguistic communities to communicate effortlessly through speech and text.
+
+This unified model enables multiple tasks like Speech-to-Speech (S2ST), Speech-to-Text (S2TT), Text-to-Speech (T2ST)
+translation and more, without relying on multiple separate models. Here the task is to do the speech to text translation. 
+"""
+
+CACHE_EXAMPLES = os.getenv("CACHE_EXAMPLES") == "1"
+
+
+AUDIO_SAMPLE_RATE  = 44100
+#MAX_INPUT_AUDIO_LENGTH =  1800 # in seconds
+DEFAULT_TARGET_LANGUAGE = "French"
+
+device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
+translator = Translator(
+    model_name_or_card="seamlessM4T_medium",
+    vocoder_name_or_card="vocoder_36langs",
+    device=device,
+)
+
+task_name = "S2TT (Speech to Text translation)"
+task_name = task_name.split()[0]
+
+def predict(
+    audio_source: str,
+    input_audio_mic: str | None,
+    input_audio_file: str | None,
+    target_language: str,
+) -> str:
+    target_language_code = LANGUAGE_NAME_TO_CODE[target_language]
+
+    if audio_source == "microphone": 
+        input_data = input_audio_mic
+    else:
+        input_data = input_audio_file
+
+    arr, org_sr = torchaudio.load(input_data)
+    new_arr = torchaudio.functional.resample(arr, orig_freq=org_sr, new_freq=AUDIO_SAMPLE_RATE)
+    #max_length = int(MAX_INPUT_AUDIO_LENGTH * AUDIO_SAMPLE_RATE)
+    # if new_arr.shape[1] > max_length:
+    #     new_arr = new_arr[:, :max_length]
+    #     gr.Warning(f"Input audio is too long. Only the first {MAX_INPUT_AUDIO_LENGTH} seconds is used.")
+    torchaudio.save(input_data, new_arr, sample_rate=int(AUDIO_SAMPLE_RATE))
+
+    text_out, wav, sr = translator.predict(
+        input = input_data,
+        task_str = task_name,
+        tgt_lang=target_language_code,
+        ngram_filtering=True,
+    )
+    
+    return text_out
+
+
+
+def update_audio_ui(audio_source: str) -> tuple[dict, dict]:
+    mic = audio_source == "microphone"
+    return (
+        gr.update(visible=mic, value=None),  # input_audio_mic
+        gr.update(visible=not mic, value=None),  # input_audio_file
+    )
+
+
+# ... (previous parts of the code remain unchanged)
+
+with gr.Blocks(css="style.css") as demo:
+
+    gr.Markdown(DESCRIPTION)
+    
+    with gr.Group():
+        with gr.Row():
+            target_language = gr.Dropdown(
+                label="Target language",
+                choices=S2TT_TARGET_LANGUAGE_NAMES,
+            )
+            
+        with gr.Row() as audio_box:
+        
+            audio_source = gr.Radio(
+                label="Audio source",
+                choices=["file", "microphone"],
+                value="file",
+            )
+            input_audio_mic = gr.Audio(
+                label="Input speech",
+                type="filepath",
+                source="microphone",
+                visible=False,
+            )
+            input_audio_file = gr.Audio(
+                label="Input speech",
+                type="filepath",
+                source="upload",
+                visible=True,
+            )
+                
+        btn = gr.Button("Translate") 
+        with gr.Column():
+            output_text = gr.Textbox(label="Translated text")
+
+        audio_source.change(
+            fn=update_audio_ui,
+            inputs=audio_source,
+            outputs=[
+                input_audio_mic,
+                input_audio_file,
+            ],
+            queue=False,
+            api_name=False,
+        )
+    
+        btn.click(
+            fn=predict,
+            inputs=[
+                audio_source,
+                input_audio_mic,
+                input_audio_file,
+                target_language,
+            ],
+            outputs=[output_text],
+            api_name="run",
+        )
+
+demo.launch()

lang_list.py

@@ -0,0 +1,254 @@
+# Language dict
+language_code_to_name = {
+    "afr": "Afrikaans",
+    "amh": "Amharic",
+    "arb": "Modern Standard Arabic",
+    "ary": "Moroccan Arabic",
+    "arz": "Egyptian Arabic",
+    "asm": "Assamese",
+    "ast": "Asturian",
+    "azj": "North Azerbaijani",
+    "bel": "Belarusian",
+    "ben": "Bengali",
+    "bos": "Bosnian",
+    "bul": "Bulgarian",
+    "cat": "Catalan",
+    "ceb": "Cebuano",
+    "ces": "Czech",
+    "ckb": "Central Kurdish",
+    "cmn": "Mandarin Chinese",
+    "cym": "Welsh",
+    "dan": "Danish",
+    "deu": "German",
+    "ell": "Greek",
+    "eng": "English",
+    "est": "Estonian",
+    "eus": "Basque",
+    "fin": "Finnish",
+    "fra": "French",
+    "gaz": "West Central Oromo",
+    "gle": "Irish",
+    "glg": "Galician",
+    "guj": "Gujarati",
+    "heb": "Hebrew",
+    "hin": "Hindi",
+    "hrv": "Croatian",
+    "hun": "Hungarian",
+    "hye": "Armenian",
+    "ibo": "Igbo",
+    "ind": "Indonesian",
+    "isl": "Icelandic",
+    "ita": "Italian",
+    "jav": "Javanese",
+    "jpn": "Japanese",
+    "kam": "Kamba",
+    "kan": "Kannada",
+    "kat": "Georgian",
+    "kaz": "Kazakh",
+    "kea": "Kabuverdianu",
+    "khk": "Halh Mongolian",
+    "khm": "Khmer",
+    "kir": "Kyrgyz",
+    "kor": "Korean",
+    "lao": "Lao",
+    "lit": "Lithuanian",
+    "ltz": "Luxembourgish",
+    "lug": "Ganda",
+    "luo": "Luo",
+    "lvs": "Standard Latvian",
+    "mai": "Maithili",
+    "mal": "Malayalam",
+    "mar": "Marathi",
+    "mkd": "Macedonian",
+    "mlt": "Maltese",
+    "mni": "Meitei",
+    "mya": "Burmese",
+    "nld": "Dutch",
+    "nno": "Norwegian Nynorsk",
+    "nob": "Norwegian Bokm\u00e5l",
+    "npi": "Nepali",
+    "nya": "Nyanja",
+    "oci": "Occitan",
+    "ory": "Odia",
+    "pan": "Punjabi",
+    "pbt": "Southern Pashto",
+    "pes": "Western Persian",
+    "pol": "Polish",
+    "por": "Portuguese",
+    "ron": "Romanian",
+    "rus": "Russian",
+    "slk": "Slovak",
+    "slv": "Slovenian",
+    "sna": "Shona",
+    "snd": "Sindhi",
+    "som": "Somali",
+    "spa": "Spanish",
+    "srp": "Serbian",
+    "swe": "Swedish",
+    "swh": "Swahili",
+    "tam": "Tamil",
+    "tel": "Telugu",
+    "tgk": "Tajik",
+    "tgl": "Tagalog",
+    "tha": "Thai",
+    "tur": "Turkish",
+    "ukr": "Ukrainian",
+    "urd": "Urdu",
+    "uzn": "Northern Uzbek",
+    "vie": "Vietnamese",
+    "xho": "Xhosa",
+    "yor": "Yoruba",
+    "yue": "Cantonese",
+    "zlm": "Colloquial Malay",
+    "zsm": "Standard Malay",
+    "zul": "Zulu",
+}
+LANGUAGE_NAME_TO_CODE = {v: k for k, v in language_code_to_name.items()}
+
+# Source langs: S2ST / S2TT / ASR don't need source lang
+# T2TT / T2ST use this
+text_source_language_codes = [
+    "afr",
+    "amh",
+    "arb",
+    "ary",
+    "arz",
+    "asm",
+    "azj",
+    "bel",
+    "ben",
+    "bos",
+    "bul",
+    "cat",
+    "ceb",
+    "ces",
+    "ckb",
+    "cmn",
+    "cym",
+    "dan",
+    "deu",
+    "ell",
+    "eng",
+    "est",
+    "eus",
+    "fin",
+    "fra",
+    "gaz",
+    "gle",
+    "glg",
+    "guj",
+    "heb",
+    "hin",
+    "hrv",
+    "hun",
+    "hye",
+    "ibo",
+    "ind",
+    "isl",
+    "ita",
+    "jav",
+    "jpn",
+    "kan",
+    "kat",
+    "kaz",
+    "khk",
+    "khm",
+    "kir",
+    "kor",
+    "lao",
+    "lit",
+    "lug",
+    "luo",
+    "lvs",
+    "mai",
+    "mal",
+    "mar",
+    "mkd",
+    "mlt",
+    "mni",
+    "mya",
+    "nld",
+    "nno",
+    "nob",
+    "npi",
+    "nya",
+    "ory",
+    "pan",
+    "pbt",
+    "pes",
+    "pol",
+    "por",
+    "ron",
+    "rus",
+    "slk",
+    "slv",
+    "sna",
+    "snd",
+    "som",
+    "spa",
+    "srp",
+    "swe",
+    "swh",
+    "tam",
+    "tel",
+    "tgk",
+    "tgl",
+    "tha",
+    "tur",
+    "ukr",
+    "urd",
+    "uzn",
+    "vie",
+    "yor",
+    "yue",
+    "zsm",
+    "zul",
+]
+TEXT_SOURCE_LANGUAGE_NAMES = sorted([language_code_to_name[code] for code in text_source_language_codes])
+
+# Target langs:
+# S2ST / T2ST
+s2st_target_language_codes = [
+    "eng",
+    "arb",
+    "ben",
+    "cat",
+    "ces",
+    "cmn",
+    "cym",
+    "dan",
+    "deu",
+    "est",
+    "fin",
+    "fra",
+    "hin",
+    "ind",
+    "ita",
+    "jpn",
+    "kor",
+    "mlt",
+    "nld",
+    "pes",
+    "pol",
+    "por",
+    "ron",
+    "rus",
+    "slk",
+    "spa",
+    "swe",
+    "swh",
+    "tel",
+    "tgl",
+    "tha",
+    "tur",
+    "ukr",
+    "urd",
+    "uzn",
+    "vie",
+]
+S2ST_TARGET_LANGUAGE_NAMES = sorted([language_code_to_name[code] for code in s2st_target_language_codes])
+
+# S2TT / ASR
+S2TT_TARGET_LANGUAGE_NAMES = TEXT_SOURCE_LANGUAGE_NAMES
+# T2TT
+T2TT_TARGET_LANGUAGE_NAMES = TEXT_SOURCE_LANGUAGE_NAMES

requirements.txt

@@ -0,0 +1,6 @@
+fairseq2==0.1.0
+git+https://github.com/facebookresearch/seamless_communication
+gradio==3.40.1
+huggingface_hub==0.16.4
+torch==2.0.1
+torchaudio==2.0.2

style.css

@@ -0,0 +1,17 @@
+h1 {
+    text-align: center;
+  }
+  
+  #duplicate-button {
+    margin: auto;
+    color: #fff;
+    background: #1565c0;
+    border-radius: 100vh;
+  }
+  
+  #component-0 {
+    max-width: 730px;
+    margin: auto;
+    padding-top: 1.5rem;
+  }
+