Update README.md

README.md

@@ -1,5 +1,10 @@
 ---
 license: mit
+tags:
+- smpl
+- human-pose-and-shape-estimation
+- human-mesh-recovery
+- inverse-kinematics
 ---
 # Learnable SMPLify: A Neural Solution for Optimization-Free Human Pose Inverse Kinematics
 
@@ -8,36 +13,6 @@ license: mit
 ``TL;DR`` Given X_{t-s} and X_{t} 3D keypoints, 
 calculate residual SMPL parameters from t-s to t.
 
-## Preparation
-Refer to [PREPARATION.md](doc/PREPARATION.md) for installation and data preparation details.
-
-## Checkpoints
-The pretrained model checkpoint is available at [Google Drive](https://drive.google.com/drive/folders/1oyG2gbB3EMcc6NgTIT1p1uJ_Em0dJwXz?usp=sharing).
-
-## Usage
-### Training
-cd to `src` folder and run the following command.
-
-```
-torchrun --nproc-per-node <NUM_GPUS> main.py --config configs/net.yaml (--extra_tag <EXTRA_TAG> --batch_size <BATCH_SIZE> --epochs <EPOCHS>)
-```
-
-You can get logs, tensorboard and checkpoints in the corresponding `logs/<MODEL_NAME>_net_<EXTRA_TAG>` folder.
-
-### Evaluation
-To evaluate the model, run the following command:
-
-```
-torchrun --nproc-per-node <NUM_GPUS> main.py --config configs/net.yaml --eval --checkpoint <PATH_TO_CHECKPOINT>
-```
-
-### Sequential Inference
-To run sequential inference, you can use the following command:
-
-```
-python inference.py <PATH_TO_CHECKPOINT> (<DATASET_NAME> <SAMPLE_RATIO>)
-```
-
 ## Citation
 If you find this work useful in your research, please consider citing: