Spaces:
Runtime error
Runtime error
File size: 1,605 Bytes
357c94c 5aef126 a1c4331 357c94c a1c4331 bd69de9 a1c4331 bd69de9 5aef126 bd69de9 5aef126 a1c4331 bd69de9 357c94c a1c4331 bd69de9 a1c4331 357c94c a1c4331 357c94c bd69de9 357c94c a1c4331 bd69de9 a1c4331 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 |
#!/bin/bash
set -e # Exit immediately on error
set -o pipefail # Catch errors in pipelines
JOBS_DIR=$(dirname "$(dirname "$0")")
export PYTHONPATH=./
MODEL_BASE=./weights
MODEL_DIR="${MODEL_BASE}/ckpts/hunyuan-video-t2v-720p"
CHECKPOINT_FILE="${MODEL_DIR}/transformers/mp_rank_00_model_states.pt"
# Step 1: Ensure model is downloaded
echo "📦 Checking model checkpoint..."
if [ -f "$CHECKPOINT_FILE" ]; then
echo "✅ Model checkpoint already exists. Skipping download."
else
echo "⬇️ Model not found. Downloading from Hugging Face..."
python3 - <<EOF
from huggingface_hub import snapshot_download
import os
target_dir = "${MODEL_DIR}"
os.makedirs(target_dir, exist_ok=True)
snapshot_download(
repo_id="tencent/HunyuanVideo-Avatar",
local_dir=target_dir,
local_dir_use_symlinks=False
)
EOF
# Confirm model was downloaded
if [ ! -f "$CHECKPOINT_FILE" ]; then
echo "❌ Failed to download model checkpoint. Exiting."
exit 1
fi
echo "✅ Model downloaded successfully."
fi
# Step 2: Launch the Flask audio server
echo "🚀 Starting flask_audio.py..."
torchrun --nnodes=1 --nproc_per_node=8 --master_port=29605 hymm_gradio/flask_audio.py \
--input 'assets/test.csv' \
--ckpt "${CHECKPOINT_FILE}" \
--sample-n-frames 129 \
--seed 128 \
--image-size 704 \
--cfg-scale 7.5 \
--infer-steps 50 \
--use-deepcache 1 \
--flow-shift-eval-video 5.0 &
# Step 3: Run Gradio UI
echo "🟢 Starting gradio_audio.py UI..."
# Add your command to run gradio_audio.py here
# Example (uncomment and adjust as needed):
|