LI TIANCHENG

### Describe the bug I run abacus on 4 V100 GPU using PW basis with dav_subspace solver for a 216 atoms system with the command `OMP_NUM_THREADS=12 mpirun --allow-run-as-root -np 4...

feat: do not output spin info for lmp format

4

fix issue #885 in dpdata and issue #312 in dpgen2 However, adding an additional parameter is not the best method. Please provide another way to fix this issue, thanks. ##...

### Bug summary When STRU contains below info: ### dpdata Version latest ### Input Files, Running Commands, Error Log, etc. Error: ### Steps to Reproduce By using commands below to...

feat: add streaming output for dp train

3

For trainning step, there is no output in log files. We can only see the output after the trainning is completed. However, the trainning normally lasts for days, so the...

when I use configurations in abacus/stru format in dpgen2, the lammps will raise error: ```text read_data conf.lmp Reading data file ... triclinic box = (0 0 0) to (13.160973 13.160973...

For DP_ZBL model, additional txt file for ZBL table is required both in training and freezing. But there is not option for exploring stage to copy this file. For trainning,...

[BUG] Cannot recognise CUDA device when installed along with tensorflow

1

### Description I created a new environment using mamba, and installed uv in it, and then I want to install both jax and tensorflow using `uv pip install -U tensorflow...

[Bug] Compilation warnings with recent TensorFlow due to deprecated API 'tsl::errors::InvalidArgument'

2

### Bug summary When compiling the C++ part of DeePMD-kit from source against a recent version of TensorFlow, a large number of deprecation warnings (`-Wdeprecated-declarations`) are generated. The warnings indicate...

[Feature Request] Accelerate DPA-3 descriptor with TileLang backend

1

### Summary This is a feature request to explore and potentially implement a [TileLang](https://github.com/tile-ai/tilelang) backend for the DPA-3 descriptor in DeepMD-kit, with the primary goal of significantly improving performance while...

PyTorch Backend: Missing autograd wrapper for se_t descriptor's second-order derivatives

2

### Summary The se_t descriptor (TabulateFusionSeTOp) does not seem to support second-order derivative calculations (required for virial/stress), even though the underlying C++/CUDA kernel (deepmd::tabulate_fusion_se_t_grad_grad_gpu) and its wrapper function (TabulateFusionSeTGradGradForward) are...