wip containerization

animation/StableAnimator/command_basic_infer.sh

@@ -1,18 +1,35 @@
-CUDA_VISIBLE_DEVICES=0 python inference_basic.py \
- --pretrained_model_name_or_path="path/checkpoints/SVD/stable-video-diffusion-img2vid-xt" \
- --output_dir="path/basic_infer" \
- --validation_control_folder="path/inference/case-1/poses" \
- --validation_image="path/inference/case-1/reference.png" \
+#!/bin/bash
+
+OUTPUT_DIR="inference/output_1/final_output"
+
+POSE_FRAME_DIR="inference/output_1/poses"
+STARTING_IMAGE="inference/output_1/frames/frame_0.png"
+
+mkdir -p OUTPUT_DIR
+
+#CUDA_VISIBLE_DEVICES=0 
+python inference_basic.py \
+ --pretrained_model_name_or_path="checkpoints/SVD/stable-video-diffusion-img2vid-xt" \
+ --output_dir="${OUTPUT_DIR}" \
+ --validation_control_folder="${POSE_FRAME_DIR}" \
+ --validation_image="${STARTING_IMAGE}" \
  --width=576 \
  --height=1024 \
  --guidance_scale=3.0 \
  --num_inference_steps=25 \
- --posenet_model_name_or_path="path/checkpoints/Animation/pose_net.pth" \
- --face_encoder_model_name_or_path="path/checkpoints/Animation/face_encoder.pth" \
- --unet_model_name_or_path="path/checkpoints/Animation/unet.pth" \
+ --posenet_model_name_or_path="checkpoints/Animation/pose_net.pth" \
+ --face_encoder_model_name_or_path="checkpoints/Animation/face_encoder.pth" \
+ --unet_model_name_or_path="checkpoints/Animation/unet.pth" \
  --tile_size=16 \
  --overlap=4 \
  --noise_aug_strength=0.02 \
  --frames_overlap=4 \
  --decode_chunk_size=4 \
- --gradient_checkpointing
+ --gradient_checkpointing
+
+
+ffmpeg -framerate 30 \
+  -start_number 0 \
+  -i "${OUTPUT_DIR}/animated_images/frame_%d.png" \
+  -c:v libx264 -pix_fmt yuv420p "${OUTPUT_DIR}/out.mp4" -y
+