diff --git a/README.md b/README.md index 43129b5..3207859 100644 --- a/README.md +++ b/README.md @@ -240,6 +240,15 @@ If you find our work helpful, feel free to give us a cite. ```bibtex +@misc{Mingflash2025, + title = {Ming-Flash-Omni: A Sparse, Unified Architecture for Multimodal Perception and Generation}, + author = {Inclusion AI}, + year = {2025}, + eprint = {2510.24821}, + archivePrefix = {arXiv}, + url = {https://arxiv.org/abs/2510.24821} +} + @misc{Mingomni2025, title = {Ming-Omni: A Unified Multimodal Model for Perception and Generation}, author = {Inclusion AI}, diff --git a/docker/docker-py310-cu121/Dockerfile b/docker/docker-py310-cu121/Dockerfile deleted file mode 100644 index 9d0534c..0000000 --- a/docker/docker-py310-cu121/Dockerfile +++ /dev/null @@ -1,28 +0,0 @@ -FROM pytorch/pytorch:2.1.2-cuda12.1-cudnn8-devel - -WORKDIR /workspace - -# Define environments -ENV DEBIAN_FRONTEND=noninteractive -ENV MAX_JOBS=16 - -# Install base packages like wget, git, etc. -RUN apt update && apt install -y wget git build-essential libxml2 ninja-build \ - && apt clean && rm -rf /var/lib/apt/lists/* - -# Install python packages including torch and flash_attn -RUN pip install --no-cache-dir torch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2 --index-url https://download.pytorch.org/whl/cu121 \ - && pip install packaging \ - && pip install flash-attn==2.3.6 --no-build-isolation - -# Install dependence of Ming -COPY matcha_tts-0.0.5.1-cp310-cp310-linux_x86_64.whl . -RUN pip install --no-cache-dir funasr==1.1.14 flash-attn==2.3.6 peft==0.13.2 diffusers tokenizers==0.20.3 transformers==4.45.0 decord==0.6.0 hyperpyyaml modelscope onnxruntime inflect conformer diffusers lightning gdown openai-whisper==20240930 \ - && pip install --no-cache-dir matcha_tts-0.0.5.1-cp310-cp310-linux_x86_64.whl \ - && pip install --no-cache-dir diffusers==0.33.0 \ - && pip install --no-cache-dir nvidia-cublas-cu12==12.4.5.8 \ - && rm matcha_tts-0.0.5.1-cp310-cp310-linux_x86_64.whl - -COPY entrypoint.sh . -RUN chmod +x ./entrypoint.sh -ENTRYPOINT ["./entrypoint.sh"] diff --git a/docker/docker-py310-cu121/entrypoint.sh b/docker/docker-py310-cu121/entrypoint.sh deleted file mode 100644 index be5d467..0000000 --- a/docker/docker-py310-cu121/entrypoint.sh +++ /dev/null @@ -1,3 +0,0 @@ -#!/bin/bash -pip install grouped_gemm==0.1.4 -exec "$@" \ No newline at end of file diff --git a/docker/docker-py310-cu124/Dockerfile b/docker/docker-py310-cu124/Dockerfile new file mode 100644 index 0000000..699a487 --- /dev/null +++ b/docker/docker-py310-cu124/Dockerfile @@ -0,0 +1,50 @@ +FROM pytorch/pytorch:2.4.1-cuda12.4-cudnn9-devel + +WORKDIR /workspace + +# Define environments +ENV DEBIAN_FRONTEND=noninteractive +ENV MAX_JOBS=16 +ENV PIP_INDEX_URL=https://mirrors.aliyun.com/pypi/simple/ +ENV CUDNN_PATH=/opt/conda/lib/python3.11/site-packages/nvidia/cudnn +ENV LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$CUDNN_PATH/lib + +# # Substitude sources with aliyun +# RUN sed -i 's/archive.ubuntu.com/mirrors.aliyun.com/g' /etc/apt/sources.list && \ +# sed -i 's/security.ubuntu.com/mirrors.aliyun.com/g' /etc/apt/sources.list + +# Install base packages like wget, git, etc. +RUN apt update && apt install -y wget git build-essential libxml2 ninja-build pkg-config && \ + apt clean && rm -rf /var/lib/apt/lists/* + +# Install python packages including torch and flash_attn +RUN pip install packaging +RUN pip install flash-attn==2.6.3 --no-build-isolation +# RUN pip install flash-attn==2.7.2.post1 --no-build-isolation + +# Install dependence of Ming +COPY matcha_tts-0.0.5.1-cp310-cp310-linux_x86_64.whl . +RUN pip install --no-cache-dir matcha_tts-0.0.5.1-cp310-cp310-linux_x86_64.whl \ +&& rm matcha_tts-0.0.5.1-cp310-cp310-linux_x86_64.whl +# RUN pip install --no-cache-dir transformer_engine[pytorch]==1.11.0 +RUN pip install --no-cache-dir transformers==4.51.3 accelerate==0.33.0 peft==0.15.0 diffusers==0.33.0 +RUN pip install --no-cache-dir hyperpyyaml modelscope onnxruntime inflect conformer lightning triton matplotlib gdown rich ipdb +RUN pip install --no-cache-dir nvidia-cublas-cu12==12.4.5.8 + +RUN pip install --no-cache-dir decord==0.6.0 iopath webdataset==0.2.100 omegaconf timm sortedcontainers +RUN pip install --no-cache-dir jsonlines fairscale typing_extensions==4.12.2 six regex ftfy +RUN pip install --no-cache-dir bitsandbytes funasr==1.1.14 openai-whisper==20240930 grpcio==1.59.0 "pynvml>=11.4.1" dlrover[torch] +RUN pip install --no-cache-dir xformers==0.0.22.post7 --no-deps +RUN pip install --no-cache-dir opencv-python-headless lmdb imageio +RUN pip install --no-cache-dir pytorch-lightning==1.8.6 +# RUN pip install --no-cache-dir pytorch-lightning==1.6.5 +RUN pip install --no-cache-dir "tensorboard>=2.11.0" torchscale jieba av==10.0.0 +RUN pip install --no-cache-dir -U mdatasets +RUN pip install --no-cache-dir opencc-python-reimplemented==0.1.7 +RUN pip install --no-cache-dir pypinyin addict +RUN pip install --no-cache-dir mmengine==0.10.7 --no-deps +RUN pip install --no-cache-dir pycryptodome==3.21.0 cryptography==43.0.0 x_transformers==2.4.12 + +COPY entrypoint.sh . +RUN chmod +x ./entrypoint.sh +ENTRYPOINT ["./entrypoint.sh"] \ No newline at end of file diff --git a/docker/docker-py310-cu124/entrypoint.sh b/docker/docker-py310-cu124/entrypoint.sh new file mode 100644 index 0000000..a17740b --- /dev/null +++ b/docker/docker-py310-cu124/entrypoint.sh @@ -0,0 +1,4 @@ +#!/bin/bash +pip install grouped_gemm==0.1.4 +pip install --no-cache-dir transformer_engine[pytorch]==1.11.0 +exec "$@" \ No newline at end of file diff --git a/docker/docker-py310-cu121/matcha_tts-0.0.5.1-cp310-cp310-linux_x86_64.whl b/docker/docker-py310-cu124/matcha_tts-0.0.5.1-cp310-cp310-linux_x86_64.whl similarity index 100% rename from docker/docker-py310-cu121/matcha_tts-0.0.5.1-cp310-cp310-linux_x86_64.whl rename to docker/docker-py310-cu124/matcha_tts-0.0.5.1-cp310-cp310-linux_x86_64.whl diff --git a/docker/docker-py38-cu122/Dockerfile b/docker/docker-py38-cu122/Dockerfile deleted file mode 100644 index 91278ef..0000000 --- a/docker/docker-py38-cu122/Dockerfile +++ /dev/null @@ -1,39 +0,0 @@ -FROM python:3.8 - -WORKDIR /workspace - -# Define environments -ENV DEBIAN_FRONTEND=noninteractive -ENV MAX_JOBS=16 -ENV CUDA_HOME=/usr/local/cuda -ENV PATH=/usr/local/cuda/bin:/usr/local/cuda-12.2/bin${PATH:+:${PATH}} -ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/cuda-12.2/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}} - -# Install base packages like wget, git, etc. -RUN apt update && apt install -y wget git build-essential libxml2 ninja-build \ - && apt clean && rm -rf /var/lib/apt/lists/* - -# Install python packages including torch and flash_attn -RUN wget https://developer.download.nvidia.com/compute/cuda/12.2.2/local_installers/cuda_12.2.2_535.104.05_linux.run \ - && sh cuda_12.2.2_535.104.05_linux.run --toolkit --silent --override \ - && echo '/usr/local/cuda/lib64' >> /etc/ld.so.conf \ - && echo '/usr/local/cuda-12.2/lib64' >> /etc/ld.so.conf \ - && ldconfig \ - && rm cuda_12.2.2_535.104.05_linux.run - -RUN pip install --no-cache-dir torch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2 \ - && pip install packaging \ - && pip install flash-attn==2.3.6 --no-build-isolation -# --index-url https://download.pytorch.org/whl/cu122 - -# Install dependence of Ming -RUN git clone --depth 1 https://github.com/inclusionAI/Ming.git \ - && cd Ming \ - && pip install --no-cache-dir funasr==1.1.14 flash-attn==2.3.6 peft==0.13.2 diffusers tokenizers==0.20.3 transformers==4.45.0 decord==0.6.0 hyperpyyaml modelscope onnxruntime inflect conformer diffusers lightning gdown openai-whisper==20240930 \ - && pip install --no-cache-dir data/matcha_tts-0.0.5.1-cp38-cp38-linux_x86_64.whl \ - && pip install --no-cache-dir diffusers==0.33.0 \ - && pip install --no-cache-dir nvidia-cublas-cu12==12.4.5.8 - -COPY entrypoint.sh . -RUN chmod +x ./entrypoint.sh -ENTRYPOINT ["./entrypoint.sh"]