Merge devel into master#2380
Merged
Merged
Conversation
Co-authored-by: AnuragKr <anuragkrsingh02@outlook.com>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Chun Cai <amoycaic@gmail.com> Signed-off-by: Chun Cai <amoycaic@gmail.com>
## Beep boop. Your images are optimized! Your image file size has been reduced by **23%** 🎉 <details> <summary> Details </summary> | File | Before | After | Percent reduction | |:--|:--|:--|:--| | /doc/nvnmd/figure_6.png | 24.15kb | 17.38kb | 28.03% | | /doc/nvnmd/figure_2.png | 27.60kb | 20.08kb | 27.23% | | /doc/nvnmd/figure_4.png | 33.31kb | 24.89kb | 25.27% | | /doc/nvnmd/figure_5.png | 36.86kb | 27.78kb | 24.65% | | /doc/nvnmd/figure_7.png | 42.60kb | 32.28kb | 24.23% | | /doc/nvnmd/figure_1.png | 21.24kb | 16.77kb | 21.05% | | /doc/nvnmd/figure_3.png | 34.43kb | 29.69kb | 13.76% | | | | | | | **Total :** | **220.19kb** | **168.87kb** | **23.31%** | </details> --- [📝 docs](https://imgbot.net/docs) | [:octocat: repo](https://github.com/imgbot/ImgBot) | [🙋🏾 issues](https://github.com/imgbot/ImgBot/issues) | [🏪 marketplace](https://github.com/marketplace/imgbot) <i>~Imgbot - Part of [Optimole](https://optimole.com/) family</i> Signed-off-by: ImgBotApp <ImgBotHelp@gmail.com> Co-authored-by: ImgBotApp <ImgBotHelp@gmail.com>
Fix deepmodeling#2176. See also tensorflow/tensorflow#58867. Note that CUDA Toolkit 12.0 requires CUDA driver 525.60.13. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Add `DipoleChargeModifier` to C API. In addition, this patch also contains the following changes: * fix bugs in `deepmd::DeepTensor::DeepTensor` and `deepmd::DipoleChargeModifier::DipoleChargeModifier` * add comments to `deepmd::DipoleChargeModifier` * support `gpu_rank` and other parameters for `deepmd::hpp::DeepPot` and `deepmd::hpp::DeepTensor` Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
* merge the implementation of `deepmd::DeepPot::print_summary` and `deepmd::DeepTensor::print_summary` to `deepmd::print_summary` * add the implementation of `deepmd::DipoleChargeModifier::print_summary` * add `DP_PrintSummary`, `deepmd::hpp::DeepPot::print_summary`, `deepmd::hpp::DeepTensor::print_summary`, and `deepmd::hpp::DipoleChargeModifier::print_summary` * add tests for above Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Before, `sys_probs` only specifies the probability of each system. This patch allows `numb_copy.npy` (`int`) to specify each frame is copied by the "numb_copy" (int) times. The default is 1 for all frames, which keeps the same behavior as before. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
In this way, one can install CUDA Toolkit and cuDNN via pip if one does not want to install them manually. These packages are provided officially by NVIDIA and can be dynamically opened before importing TensorFlow. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
) Fix deepmodeling#2201. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Add `--unit-model` for `dp freeze` in multitask mode; Add init_frz_model api for multi model
Co-authored-by: AnuragKr <anuragkrsingh02@outlook.com>
Add `layer_name` parameter to share network parameters among different fitting network layers. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Support using a text file(E.G. `list.txt`) to specify the list of test sets list.txt ``` data_dir/sys.000 data_dir/sys.001 data_dir/sys.002 ... ``` `/data_dir/sys.000` should be a directory with "type.raw". Then one can use `-l list.txt` to replace `-s data_dir`.
…2239) When `atom_ener` is set, there will be variables named `layer_xx_suffix_1/MatMul` (the automatic name when `layer_xx_suffix/MatMul` has been used).
This fixes an error: ``` OSError: /home/jz748/anaconda3/envs/tf/lib/python3.10/site-packages/nvidia/cublas/lib/libcublas.so.11: symbol cublasLtHSHMatmulAlgoInit, version libcublasLt.so.11 not defined in file libcublasLt.so.11 with link time reference ``` Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
0.2.8 does not pass the tests. See https://github.com/njzjz/deepmd-kit/actions/runs/3890065119/jobs/6638845358. See also deepmodeling/dargs#8. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
See the docstring for detials. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Fix deepmodeling#2238. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
A loop using `np.append` is $O(N^2)$. See numpy/numpy#17090. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Here are my results (2080 Super, TF 2.9, CUDA 11.6): (unit: ms/step) | no JIT | JIT -- | -- | -- water tanh fp32 | 8.45 | 7.78 water tanh fp64 | 33.37 | 34.09 water gelu fp32 | 19.69 | 14.23 water gelu fp64 | 48.39 | 49.19 Conclusion: 1. TF JIT only works with FP32 and slows down FP64 (or my card only has good FP32 performance?). 2. JIT has a great performance on TF's GeLU. 3. JIT only works for static shapes, so training with different natoms (or different atom types) needs a lot of compilation time (not sure if it's worth it). 4. JIT only works for a GPU. Usage ```sh export DP_JIT=1 ``` (It sets `TF_XLA_FLAGS=--tf_xla_auto_jit=2` but `DP_JIT=1` is easier to remember) Co-authored-by: Han Wang <92130845+wanghan-iapcm@users.noreply.github.com>
Fix deepmodeling#2231. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
…ing#2264) Co-authored-by: Han Wang <wang_han@iapcm.ac.cn>
Fix deepmodeling#2237. Use the LAMMPS `pair_coeff` command to set `type_map`, like ReaxFF, EAM, etc. Add and improve the documentation about `type_map`. From now on, setting `type_map` explicitly is encouraged (but not necessary). Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
…2269) Fix to deepmodeling#1387. Co-authored-by: Sigbjoern Loeland Bore <sigbjobo@eduroam-193-157-248-42.wlan.uio.no>
I prefer manually check the coverage. --------- Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Fix deepmodeling#2346. --------- Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
…ng#2357) Follow deepmodeling#830 to fix deepmodeling#2303. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
It looks like the `yum install` is broken. Switch to the local run file. --------- Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Fix `error: ‘strlen’ was not declared in this scope` --------- Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Han Wang <92130845+wanghan-iapcm@users.noreply.github.com>
Support virtual atom types (`-1`) for `se_atten`. Thus, mixed numbers of atoms can be input for inference with mixed_type. --------- Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu> Co-authored-by: Han Wang <92130845+wanghan-iapcm@users.noreply.github.com> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
…#2351) Fix C++ inference of multiple frames with fparam/aparam. The input fparam/aparam can be in a single or multiple frames. Add Python and C++ tests for fparam/aparam. --------- Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Also, add tests. --------- Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
The URL links to https://deepmd.rtfd.io/parallelism/. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
This PR does the same thing as deepmodeling#2139. --------- Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Make this status pass:  Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
…eepmodeling#2363) Since the Red Hat Developer Toolset does not support the custom `_GLIBCXX_USE_CXX11_ABI` flag, this PR improves the flag and the message. - if both `_GLIBCXX_USE_CXX11_ABI=0` and `_GLIBCXX_USE_CXX11_ABI=1` work, throw a warning and set `_GLIBCXX_USE_CXX11_ABI=1` (this flag should not take actual effect). - if both flags do not work, throw an error message about the compiler issue. Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
This PR creates a test for `se_a_ebd` descriptor to increase test coverage. --------- Signed-off-by: Han Wang <92130845+wanghan-iapcm@users.noreply.github.com> Co-authored-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Han Wang <92130845+wanghan-iapcm@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
13 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.