AI Choreographer多模式內(nèi)容創(chuàng)建模型訓(xùn)練基礎(chǔ)設(shè)施

conda create -n mint python=3.7
conda activate mint
conda install protobuf numpy
pip install tensorflow absl-py tensorflow-datasets librosa

sudo apt-get install libopenexr-dev
pip install --upgrade OpenEXR
pip install tensorflow-graphics tensorflow-graphics-gpu

git clone https://github.com/arogozhnikov/einops /tmp/einops
cd /tmp/einops/ && pip install . -U

git clone https://github.com/google/aistplusplus_api /tmp/aistplusplus_api
cd /tmp/aistplusplus_api && pip install -r requirements.txt && pip install . -U

注意如果遇到 numpy 的環(huán)境沖突，可以試試 pip install numpy==1.20

獲取數(shù)據(jù)

數(shù)據(jù)在該網(wǎng)站。

運行代碼

• 編譯協(xié)議

protoc ./mint/protos/*.proto

• 將數(shù)據(jù)集預(yù)處理為 tfrecord 

python tools/preprocessing.py \
    --anno_dir="/mnt/data/aist_plusplus_final/" \
    --audio_dir="/mnt/data/AIST/music/" \
    --split=train
python tools/preprocessing.py \
    --anno_dir="/mnt/data/aist_plusplus_final/" \
    --audio_dir="/mnt/data/AIST/music/" \
    --split=testval

• 訓(xùn)練

python trainer.py --config_path ./configs/fact_v5_deeper_t10_cm12.config --model_dir ./checkpoints

• 運行測試和評估 

# caching the generated motions (seed included) to `./outputs`
python evaluator.py --config_path ./configs/fact_v5_deeper_t10_cm12.config --model_dir ./checkpoints
# calculate FIDs
python tools/calculate_scores.py