Open Junglesl opened 2 years ago
kkoutini
commented
2 years ago Hi, can you report the output of this command:
pip list | grep sacredmaybe running :
pip install -e 'git+https://github.com/kkoutini/sacred@v0.0.1#egg=sacred' should install the correct sacred version.
Junglesl
commented
2 years ago Hi, can you report the output of this command:
pip list | grep sacredmaybe running :
pip install -e 'git+https://github.com/kkoutini/sacred@v0.0.1#egg=sacred'should install the correct sacred version.
I reinstalled sacred as you suggested and the output of pip list | grep sacred is as follows:
This version should be correct. I have another question:is the audio input is normalized to [-1,1] before they enter your model? I want to verify this because this is important. Thank you very much!
kkoutini
commented
2 years ago I think you need to uninstall sacred first:
pip uninstall -y sacred && pip install -e 'git+https://github.com/kkoutini/sacred@v0.0.1#egg=sacred'you can check wether you have the correct version like this:
python -c "from sacred.config_helpers import CMD"yes the model expects the range [-1,1]
Junglesl
commented
2 years ago I think you need to uninstall sacred first:
pip uninstall -y sacred && pip install -e 'git+https://github.com/kkoutini/sacred@v0.0.1#egg=sacred'you can check wether you have the correct version like this:
python -c "from sacred.config_helpers import CMD"yes the model expects the range [-1,1]
Thank you very much! I will try your suggestion. I have two other questiones and I need your suggestions.
and I run the following recommanded code in passt_hear21:
it runs successfully.But when I change the audio input to audio = torch.ones((1, 32000 seconds))0.5,there is a segmentation fault. when I change the input audio to audio = torch.ones((5, 32000 seconds))0.5, the following error occurred:
/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/torch/functional.py:581: UserWarning: stft will soon require the return_complex parameter be given for real inputs, and will further require that return_complex=True in a future PyTorch release. (Triggered internally at /pytorch/aten/src/ATen/native/SpectralOps.cpp:639.)
normalized, onesided, return_complex)
x torch.Size([5, 1, 128, 16])
/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/hear21passt/models/passt.py:260: UserWarning: Input image size (12816) doesn't match model (128998).
warnings.warn(f"Input image size ({H}{W}) doesn't match model ({self.img_size[0]}{self.img_size[1]}).")
self.norm(x) torch.Size([5, 768, 12, 1])
patch_embed : torch.Size([5, 768, 12, 1])
self.time_new_pos_embed.shape torch.Size([1, 768, 1, 99])
CUT time_new_pos_embed.shape torch.Size([1, 768, 1, 1])
self.freq_new_pos_embed.shape torch.Size([1, 768, 12, 1])
X flattened torch.Size([5, 12, 768])
self.new_pos_embed.shape torch.Size([1, 2, 768])
self.cls_tokens.shape torch.Size([5, 1, 768])
self.dist_token.shape torch.Size([5, 1, 768])
final sequence x torch.Size([5, 14, 768])
Traceback (most recent call last):
File "default.py", line 19, in cublasSgemm( handle, opa, opb, m, n, k, &alpha, a, lda, b, ldb, &beta, c, ldc)
I want to know whether the number 3 represents batch size? And why can't it changed to other numbers?
2.The input of the recommanded code is a 15s audio, I want to know what is the longest audio that can be input to the model?
Thank you very much!
Junglesl
commented
2 years ago I think you need to uninstall sacred first:
pip uninstall -y sacred && pip install -e 'git+https://github.com/kkoutini/sacred@v0.0.1#egg=sacred'you can check wether you have the correct version like this:
python -c "from sacred.config_helpers import CMD"yes the model expects the range [-1,1]
Thank you very much! I will try your suggestion. I have three other questiones and I need your suggestions.
- I pip the model passt_hear21 as suggests and I use the model weight passt-s-f128-p16-s10-ap.476-swa.pt,
and I run the following recommanded code in passt_hear21:
it runs successfully.But when I change the audio input to audio = torch.ones((1, 32000 seconds))0.5,there is a segmentation fault. when I change the input audio to audio = torch.ones((5, 32000 seconds))0.5, the following error occurred: /home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/torch/functional.py:581: UserWarning: stft will soon require the return_complex parameter be given for real inputs, and will further require that return_complex=True in a future PyTorch release. (Triggered internally at /pytorch/aten/src/ATen/native/SpectralOps.cpp:639.) normalized, onesided, return_complex) x torch.Size([5, 1, 128, 16]) /home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/hear21passt/models/passt.py:260: UserWarning: Input image size (128_16) doesn't match model (128998). warnings.warn(f"Input image size ({H}{W}) doesn't match model ({self.imgsize[0]}{self.img_size[1]}).") self.norm(x) torch.Size([5, 768, 12, 1]) patch_embed : torch.Size([5, 768, 12, 1]) self.time_new_pos_embed.shape torch.Size([1, 768, 1, 99]) CUT time_new_pos_embed.shape torch.Size([1, 768, 1, 1]) self.freq_new_pos_embed.shape torch.Size([1, 768, 12, 1]) X flattened torch.Size([5, 12, 768]) self.new_pos_embed.shape torch.Size([1, 2, 768]) self.cls_tokens.shape torch.Size([5, 1, 768]) self.dist_token.shape torch.Size([5, 1, 768]) final sequence x torch.Size([5, 14, 768]) Traceback (most recent call last): File "default.py", line 19, in embed, time_stamps = get_timestamp_embeddings(audio, model) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/hear21passt/base.py", line 35, in get_timestamp_embeddings return get_basic_timestamp_embeddings(audio, model) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/hear21passt/base.py", line 48, in get_basic_timestamp_embeddings return model.get_timestamp_embeddings(audio) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/hear21passt/wrapper.py", line 91, in get_timestamp_embeddings embeddings.append(self.forward(segment.to(self.device())).cpu()) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/hear21passt/wrapper.py", line 38, in forward x, features = self.net(specs) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, kwargs) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/hear21passt/models/passt.py", line 503, in forward x = self.forward_features(x) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/hear21passt/models/passt.py", line 491, in forward_features x = self.blocks(x) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, *kwargs) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/torch/nn/modules/container.py", line 119, in forward input = module(input) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(input, kwargs) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/hear21passt/models/passt.py", line 312, in forward x = x + self.drop_path(self.mlp(self.norm2(x))) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, *kwargs) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/hear21passt/models/passt.py", line 226, in forward x = self.fc1(x) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(input, **kwargs) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/torch/nn/modules/linear.py", line 94, in forward return F.linear(input, self.weight, self.bias) File "/home/jovyan/miniconda3/envs/ba3l/lib/python3.7/site-packages/torch/nn/functional.py", line 1753, in linear return torch._C._nn.linear(input, weight, bias) RuntimeError: CUDA error: CUBLAS_STATUS_EXECUTION_FAILED when calling
cublasSgemm( handle, opa, opb, m, n, k, &alpha, a, lda, b, ldb, &beta, c, ldc)
I want to know whether the number 3 represents batch size? And why can't it changed to other numbers?
kkoutini
commented
2 years ago The model is trained on 10-seconds audio clips and has time positional encodings for 10 seconds only. If you want to do inference on longer clips, one possibility is using a sliding windows like the one here .
Timestamp embbedings will cut the audio into small snippets and pass these sequentially to the model (see here ). A simple optimization is to stack the snippets and pass them as a batch to the GPU.
The basic model can only deal with 10-seconds audio clips. This release contains models passt-s-f128-30sec-p16-s10-ap.473-swa.pt passt-s-f128-20sec-p16-s10-ap.474-swa.pt that can do inference on 20 and 30 seconds audio clips. These 2 model has time positional encodings 20 to 30 seconds. Although they are trained on 10-seconds audio clips. In order to make this possible, I sample a sub-sequence corresponding to 10 seconds of time positional encoding during training.
Junglesl
commented
2 years ago
- The model is trained on 10-seconds audio clips and has time positional encodings for 10 seconds only. If you want to do inference on longer clips, one possibility is using a sliding windows like the one here .
- Timestamp embbedings will cut the audio into small snippets and pass these sequentially to the model (see here ). A simple optimization is to stack the snippets and pass them as a batch to the GPU.
- The basic model can only deal with 10-seconds audio clips. This release contains models passt-s-f128-30sec-p16-s10-ap.473-swa.pt passt-s-f128-20sec-p16-s10-ap.474-swa.pt that can do inference on 20 and 30 seconds audio clips. These 2 model has time positional encodings 20 to 30 seconds. Although they are trained on 10-seconds audio clips. In order to make this possible, I sample a sub-sequence corresponding to 10 seconds of time positional encoding during training.
Thank you for your reply very much! I have another question, could the model be loaded on multi-gpus? It costs about 2.86 seconds on a 10 seconds' audio to get its audio embedding,it's too slow since we have more than 400,000 audios. And I want to know if I cut the audio to 10 seconds audio segments,how can I integrate them together to get the final embedding? For example, if I have a 46 seconds audio,I cut them to 5 segments(Whether should I pad zero for this 46 seconds audio to 50 seconds?), for each segment,I obtain an 1295 dimension embedding,how can I obtain the final 1295 embedding for this 1295 dimension embedding? And I still have question on why the batch size must be 3,and why can't the batch size be 1? Looking forward to your reply.Thank you very much!
kkoutini
commented
2 years ago Hi, yes the model can be loaded on multiple-gpus and be used with DistributedDataParallel similar to other torch models.
And I want to know if I cut the audio to 10 seconds audio segments,how can I integrate them together to get the final embedding? For example, if I have a 46 seconds audio,I cut them to 5 segments(Whether should I pad zero for this 46 seconds audio to 50 seconds?), for each segment,I obtain an 1295 dimension embedding,how can I obtain the final 1295 embedding for this 1295 dimension embedding?
I didn't look much into the best way to aggregate local embedding. We tried the simple method of taking the global mean in HEAR, it worked well.
And I still have question on why the batch size must be 3,and why can't the batch size be 1?
There are no constraints on the batch size.
Junglesl
commented
2 years ago Hi, yes the model can be loaded on multiple-gpus and be used with
DistributedDataParallelsimilar to other torch models.And I want to know if I cut the audio to 10 seconds audio segments,how can I integrate them together to get the final embedding? For example, if I have a 46 seconds audio,I cut them to 5 segments(Whether should I pad zero for this 46 seconds audio to 50 seconds?), for each segment,I obtain an 1295 dimension embedding,how can I obtain the final 1295 embedding for this 1295 dimension embedding?
I didn't look much into the best way to aggregate local embedding. We tried the simple method of taking the global mean in HEAR, it worked well.
And I still have question on why the batch size must be 3,and why can't the batch size be 1?
There are no constraints on the batch size.
Thank you for your reply very much! There is another question,when I want to obtain the embeddings from the function get_timestamp_embeddings, Segmentation fault often happen. I don't know what happened and why. This is disturbed very much. Do you know the reason and how to solve it? I will be apperiate it very much!
kkoutini
commented
2 years ago Unfortunately I've never got a Segmentation fault. It can be caused by any of the dependencies or packages with native code, in the environment.
Junglesl
commented
2 years ago Unfortunately I've never got a Segmentation fault. It can be caused by any of the dependencies or packages with native code, in the environment.
I want to know if I only want to do inferfence use the pretained model,If I only need to install the packages in the green line? Do I need to install the packages in the red lines? I only install the passt_hear21-0.0.8 but I force the following bug:
I don't konw why?Could you give some suggestions? Thank you very much!
Junglesl
commented
2 years ago Unfortunately I've never got a Segmentation fault. It can be caused by any of the dependencies or packages with native code, in the environment.
And when I download the hear21passt manually and run:python setup.py install,I enforce the following bug:
Why torch-11 is installed? I run the command following
should't torch=1.8 is installed? It's really disturbing since I have expended much time on the environment settings and I don't know where the problem is.Hope you suggestions.Thank you very much!
Junglesl
commented
2 years ago Unfortunately I've never got a Segmentation fault. It can be caused by any of the dependencies or packages with native code, in the environment.
I find in environment.yml that conda-forge channel is above the defaults.
Which is different from the common environment. If I don't put the conda-forge above the defaults, some bugs may arise? And the cuda and torchaudio can't use? Thank you!
kkoutini
commented
2 years ago Hi, the environment.yml is a snapshot of the environment I'm using. It was exported using:
conda env export --no-builds | grep -v "prefix" > environment.ymlI'm not sure why you have torch 1.11 installed and if it matters. I think the segment fault can happen in case of a problem in the complied dependencies. Maybe you can try to uninstall it first, then install the correct version:
pip3 install torch==1.8.1+cu111 torchaudio==0.8.1 -f https://download.pytorch.org/whl/lts/1.8/torch_lts.htmlIf it helps this the full environment I have, on another machine:
(ba3l) ➜ ~ conda list
# packages in environment at /system/user/koutini/miniconda3/envs/ba3l:
#
# Name Version Build Channel
_libgcc_mutex 0.1 conda_forge conda-forge
_openmp_mutex 4.5 1_llvm conda-forge
_pytorch_select 0.1 cpu_0
absl-py 0.12.0 pypi_0 pypi
aiohttp 3.7.4.post0 pypi_0 pypi
appdirs 1.4.4 pyh9f0ad1d_0 conda-forge
astunparse 1.6.3 pyhd8ed1ab_0 conda-forge
async-timeout 3.0.1 pypi_0 pypi
attrs 20.3.0 pypi_0 pypi
audioread 2.1.9 py37h89c1867_0 conda-forge
av 8.0.3 pypi_0 pypi
ba3l 0.0.0.1 dev_0 <develop>
blas 1.0 openblas conda-forge
brotlipy 0.7.0 py37h5e8e339_1001 conda-forge
bzip2 1.0.8 h7f98852_4 conda-forge
c-ares 1.17.1 h7f98852_1 conda-forge
ca-certificates 2021.5.30 ha878542_0 conda-forge
cached-property 1.5.2 hd8ed1ab_1 conda-forge
cached_property 1.5.2 pyha770c72_1 conda-forge
cachetools 4.2.1 pypi_0 pypi
certifi 2021.5.30 py37h89c1867_0 conda-forge
cffi 1.14.5 py37hc58025e_0 conda-forge
chardet 4.0.0 py37h89c1867_1 conda-forge
cmake 3.20.3 h8897547_0 conda-forge
colorama 0.4.4 pyh9f0ad1d_0 conda-forge
cryptography 3.4.6 py37h5d9358c_0 conda-forge
cudatoolkit 11.2.0 h73cb219_8 conda-forge
cycler 0.10.0 py_2 conda-forge
dataclasses 0.8 pyhc8e2a94_1 conda-forge
decorator 4.4.2 py_0 conda-forge
docopt 0.6.2 py_1 conda-forge
einops 0.3.0 pypi_0 pypi
expat 2.4.1 h9c3ff4c_0 conda-forge
ffmpeg 4.3.1 hca11adc_2 conda-forge
freetype 2.10.4 h0708190_1 conda-forge
fsspec 0.8.7 pypi_0 pypi
future 0.18.2 py37h89c1867_3 conda-forge
gettext 0.19.8.1 h0b5b191_1005 conda-forge
gitdb 4.0.5 pyhd8ed1ab_1 conda-forge
gitpython 3.1.14 pyhd8ed1ab_0 conda-forge
gmp 6.2.1 h58526e2_0 conda-forge
gnutls 3.6.13 h85f3911_1 conda-forge
google-auth 1.28.0 pypi_0 pypi
google-auth-oauthlib 0.4.3 pypi_0 pypi
gpuinfo 1.0.0a7 pypi_0 pypi
grpcio 1.36.1 pypi_0 pypi
h5py 3.1.0 nompi_py37h1e651dc_100 conda-forge
hdf5 1.10.6 nompi_h6a2412b_1114 conda-forge
idna 2.10 pyh9f0ad1d_0 conda-forge
imageio 2.9.0 pypi_0 pypi
importlib-metadata 3.7.3 py37h89c1867_0 conda-forge
importlib_metadata 3.7.3 hd8ed1ab_0 conda-forge
intel-openmp 2020.2 254
jbig 2.1 h7f98852_2003 conda-forge
joblib 1.0.1 pyhd8ed1ab_0 conda-forge
jpeg 9d h36c2ea0_0 conda-forge
jsonpickle 1.4.1 pyh9f0ad1d_0 conda-forge
kiwisolver 1.3.1 py37h2527ec5_1 conda-forge
krb5 1.19.1 hcc1bbae_0 conda-forge
lame 3.100 h7f98852_1001 conda-forge
lcms2 2.12 hddcbb42_0 conda-forge
ld_impl_linux-64 2.35.1 hea4e1c9_2 conda-forge
lerc 2.2.1 h9c3ff4c_0 conda-forge
libblas 3.9.0 8_openblas conda-forge
libcblas 3.9.0 8_openblas conda-forge
libcurl 7.77.0 h2574ce0_0 conda-forge
libdeflate 1.7 h7f98852_5 conda-forge
libedit 3.1.20191231 he28a2e2_2 conda-forge
libev 4.33 h516909a_1 conda-forge
libffi 3.3 h58526e2_2 conda-forge
libflac 1.3.3 h9c3ff4c_1 conda-forge
libgcc-ng 9.3.0 h2828fa1_19 conda-forge
libgfortran-ng 9.3.0 hff62375_19 conda-forge
libgfortran5 9.3.0 hff62375_19 conda-forge
libgomp 9.3.0 h2828fa1_19 conda-forge
liblapack 3.9.0 8_openblas conda-forge
libllvm10 10.0.1 he513fc3_3 conda-forge
libnghttp2 1.43.0 h812cca2_0 conda-forge
libogg 1.3.4 h7f98852_1 conda-forge
libopenblas 0.3.12 pthreads_h4812303_1 conda-forge
libopus 1.3.1 h7f98852_1 conda-forge
libpng 1.6.37 h21135ba_2 conda-forge
librosa 0.8.0 pyh9f0ad1d_0 conda-forge
libsndfile 1.0.31 h9c3ff4c_1 conda-forge
libssh2 1.9.0 ha56f1ee_6 conda-forge
libstdcxx-ng 9.3.0 h6de172a_19 conda-forge
libtiff 4.3.0 hf544144_1 conda-forge
libuv 1.41.0 h7f98852_0 conda-forge
libvorbis 1.3.7 h9c3ff4c_0 conda-forge
libwebp-base 1.2.0 h7f98852_2 conda-forge
llvm-openmp 11.1.0 h4bd325d_1 conda-forge
llvmlite 0.36.0 py37h9d7f4d0_0 conda-forge
lz4-c 1.9.3 h9c3ff4c_0 conda-forge
magma-cuda112 2.5.2 1 pytorch
markdown 3.3.4 pypi_0 pypi
matplotlib-base 3.3.4 py37h0c9df89_0 conda-forge
mkl 2020.2 ha770c72_256 conda-forge
mkl-include 2021.2.0 h726a3e6_389 conda-forge
mkl-service 2.3.0 py37h8f50634_2 conda-forge
multidict 5.1.0 pypi_0 pypi
munch 2.5.0 py_0 conda-forge
ncurses 6.2 h58526e2_4 conda-forge
nettle 3.6 he412f7d_0 conda-forge
ninja 1.10.2 h4bd325d_0 conda-forge
numba 0.53.0 py37h7dd73a4_1 conda-forge
numpy 1.20.1 py37haa41c4c_0 conda-forge
oauthlib 3.1.0 pypi_0 pypi
olefile 0.46 pyh9f0ad1d_1 conda-forge
openblas 0.3.12 pthreads_h04b7a96_1 conda-forge
openh264 2.1.1 h780b84a_0 conda-forge
openjpeg 2.4.0 hb52868f_1 conda-forge
openssl 1.1.1k h7f98852_0 conda-forge
packaging 20.9 pyh44b312d_0 conda-forge
pandas 1.2.3 py37hdc94413_0 conda-forge
pillow 8.1.2 py37h4600e1f_1 conda-forge
pip 21.0.1 pyhd8ed1ab_0 conda-forge
pooch 1.3.0 pyhd8ed1ab_0 conda-forge
protobuf 3.15.6 pypi_0 pypi
py-cpuinfo 7.0.0 pyh9f0ad1d_0 conda-forge
pyasn1 0.4.8 pypi_0 pypi
pyasn1-modules 0.2.8 pypi_0 pypi
pycparser 2.20 pyh9f0ad1d_2 conda-forge
pymongo 3.11.3 pypi_0 pypi
pyopenssl 20.0.1 pyhd8ed1ab_0 conda-forge
pyparsing 2.4.7 pyh9f0ad1d_0 conda-forge
pysocks 1.7.1 py37h89c1867_3 conda-forge
pysoundfile 0.10.3.post1 pyhd3deb0d_0 conda-forge
python 3.7.10 hffdb5ce_100_cpython conda-forge
python-dateutil 2.8.1 py_0 conda-forge
python_abi 3.7 1_cp37m conda-forge
pytorch-lightning 1.3.0.dev0 dev_0 <develop>
pytz 2021.1 pyhd8ed1ab_0 conda-forge
pyyaml 5.3.1 pypi_0 pypi
readline 8.0 he28a2e2_2 conda-forge
requests 2.25.1 pyhd3deb0d_0 conda-forge
requests-oauthlib 1.3.0 pypi_0 pypi
resampy 0.2.2 py_0 conda-forge
rhash 1.4.1 h7f98852_0 conda-forge
rsa 4.7.2 pypi_0 pypi
sacred 0.8.1 dev_0 <develop>
scikit-learn 0.24.1 py37h69acf81_0 conda-forge
scipy 1.6.1 py37h14a347d_0 conda-forge
setuptools 49.6.0 py37h89c1867_3 conda-forge
six 1.15.0 pyh9f0ad1d_0 conda-forge
smmap 3.0.5 pyh44b312d_0 conda-forge
sqlite 3.34.0 h74cdb3f_0 conda-forge
tensorboard 2.4.1 pypi_0 pypi
tensorboard-plugin-wit 1.8.0 pypi_0 pypi
test-tube 0.7.5 pypi_0 pypi
threadpoolctl 2.1.0 pyh5ca1d4c_0 conda-forge
timm 0.4.12 pypi_0 pypi
tk 8.6.10 h21135ba_1 conda-forge
torch 1.10.0a0+gitc51abf8 pypi_0 pypi
torchaudio 0.8.2 pypi_0 pypi
torchmetrics 0.2.0 pypi_0 pypi
torchvision 0.9.2+cu111 pypi_0 pypi
tornado 6.1 py37h5e8e339_1 conda-forge
tqdm 4.59.0 pypi_0 pypi
typing_extensions 3.7.4.3 py_0 conda-forge
urllib3 1.26.4 pyhd8ed1ab_0 conda-forge
werkzeug 1.0.1 pypi_0 pypi
wheel 0.36.2 pypi_0 pypi
wrapt 1.12.1 py37h5e8e339_3 conda-forge
x264 1!161.3030 h7f98852_0 conda-forge
xz 5.2.5 h516909a_1 conda-forge
yaml 0.2.5 h516909a_0 conda-forge
yarl 1.6.3 pypi_0 pypi
zipp 3.4.1 pyhd8ed1ab_0 conda-forge
zlib 1.2.11 h516909a_1010 conda-forge
zstd 1.5.0 ha95c52a_0 conda-forge
Hi, the environment.yml is a snapshot of the environment I'm using. It was exported using:
conda env export --no-builds | grep -v "prefix" > environment.ymlI'm not sure why you have torch 1.11 installed and if it matters. I think the segment fault can happen in case of a problem in the complied dependencies. Maybe you can try to uninstall it first, then install the correct version:
pip3 install torch==1.8.1+cu111 torchaudio==0.8.1 -f https://download.pytorch.org/whl/lts/1.8/torch_lts.htmlIf it helps this the full environment I have, on another machine:
(ba3l) ➜ ~ conda list # packages in environment at /system/user/koutini/miniconda3/envs/ba3l: # # Name Version Build Channel _libgcc_mutex 0.1 conda_forge conda-forge _openmp_mutex 4.5 1_llvm conda-forge _pytorch_select 0.1 cpu_0 absl-py 0.12.0 pypi_0 pypi aiohttp 3.7.4.post0 pypi_0 pypi appdirs 1.4.4 pyh9f0ad1d_0 conda-forge astunparse 1.6.3 pyhd8ed1ab_0 conda-forge async-timeout 3.0.1 pypi_0 pypi attrs 20.3.0 pypi_0 pypi audioread 2.1.9 py37h89c1867_0 conda-forge av 8.0.3 pypi_0 pypi ba3l 0.0.0.1 dev_0 <develop> blas 1.0 openblas conda-forge brotlipy 0.7.0 py37h5e8e339_1001 conda-forge bzip2 1.0.8 h7f98852_4 conda-forge c-ares 1.17.1 h7f98852_1 conda-forge ca-certificates 2021.5.30 ha878542_0 conda-forge cached-property 1.5.2 hd8ed1ab_1 conda-forge cached_property 1.5.2 pyha770c72_1 conda-forge cachetools 4.2.1 pypi_0 pypi certifi 2021.5.30 py37h89c1867_0 conda-forge cffi 1.14.5 py37hc58025e_0 conda-forge chardet 4.0.0 py37h89c1867_1 conda-forge cmake 3.20.3 h8897547_0 conda-forge colorama 0.4.4 pyh9f0ad1d_0 conda-forge cryptography 3.4.6 py37h5d9358c_0 conda-forge cudatoolkit 11.2.0 h73cb219_8 conda-forge cycler 0.10.0 py_2 conda-forge dataclasses 0.8 pyhc8e2a94_1 conda-forge decorator 4.4.2 py_0 conda-forge docopt 0.6.2 py_1 conda-forge einops 0.3.0 pypi_0 pypi expat 2.4.1 h9c3ff4c_0 conda-forge ffmpeg 4.3.1 hca11adc_2 conda-forge freetype 2.10.4 h0708190_1 conda-forge fsspec 0.8.7 pypi_0 pypi future 0.18.2 py37h89c1867_3 conda-forge gettext 0.19.8.1 h0b5b191_1005 conda-forge gitdb 4.0.5 pyhd8ed1ab_1 conda-forge gitpython 3.1.14 pyhd8ed1ab_0 conda-forge gmp 6.2.1 h58526e2_0 conda-forge gnutls 3.6.13 h85f3911_1 conda-forge google-auth 1.28.0 pypi_0 pypi google-auth-oauthlib 0.4.3 pypi_0 pypi gpuinfo 1.0.0a7 pypi_0 pypi grpcio 1.36.1 pypi_0 pypi h5py 3.1.0 nompi_py37h1e651dc_100 conda-forge hdf5 1.10.6 nompi_h6a2412b_1114 conda-forge idna 2.10 pyh9f0ad1d_0 conda-forge imageio 2.9.0 pypi_0 pypi importlib-metadata 3.7.3 py37h89c1867_0 conda-forge importlib_metadata 3.7.3 hd8ed1ab_0 conda-forge intel-openmp 2020.2 254 jbig 2.1 h7f98852_2003 conda-forge joblib 1.0.1 pyhd8ed1ab_0 conda-forge jpeg 9d h36c2ea0_0 conda-forge jsonpickle 1.4.1 pyh9f0ad1d_0 conda-forge kiwisolver 1.3.1 py37h2527ec5_1 conda-forge krb5 1.19.1 hcc1bbae_0 conda-forge lame 3.100 h7f98852_1001 conda-forge lcms2 2.12 hddcbb42_0 conda-forge ld_impl_linux-64 2.35.1 hea4e1c9_2 conda-forge lerc 2.2.1 h9c3ff4c_0 conda-forge libblas 3.9.0 8_openblas conda-forge libcblas 3.9.0 8_openblas conda-forge libcurl 7.77.0 h2574ce0_0 conda-forge libdeflate 1.7 h7f98852_5 conda-forge libedit 3.1.20191231 he28a2e2_2 conda-forge libev 4.33 h516909a_1 conda-forge libffi 3.3 h58526e2_2 conda-forge libflac 1.3.3 h9c3ff4c_1 conda-forge libgcc-ng 9.3.0 h2828fa1_19 conda-forge libgfortran-ng 9.3.0 hff62375_19 conda-forge libgfortran5 9.3.0 hff62375_19 conda-forge libgomp 9.3.0 h2828fa1_19 conda-forge liblapack 3.9.0 8_openblas conda-forge libllvm10 10.0.1 he513fc3_3 conda-forge libnghttp2 1.43.0 h812cca2_0 conda-forge libogg 1.3.4 h7f98852_1 conda-forge libopenblas 0.3.12 pthreads_h4812303_1 conda-forge libopus 1.3.1 h7f98852_1 conda-forge libpng 1.6.37 h21135ba_2 conda-forge librosa 0.8.0 pyh9f0ad1d_0 conda-forge libsndfile 1.0.31 h9c3ff4c_1 conda-forge libssh2 1.9.0 ha56f1ee_6 conda-forge libstdcxx-ng 9.3.0 h6de172a_19 conda-forge libtiff 4.3.0 hf544144_1 conda-forge libuv 1.41.0 h7f98852_0 conda-forge libvorbis 1.3.7 h9c3ff4c_0 conda-forge libwebp-base 1.2.0 h7f98852_2 conda-forge llvm-openmp 11.1.0 h4bd325d_1 conda-forge llvmlite 0.36.0 py37h9d7f4d0_0 conda-forge lz4-c 1.9.3 h9c3ff4c_0 conda-forge magma-cuda112 2.5.2 1 pytorch markdown 3.3.4 pypi_0 pypi matplotlib-base 3.3.4 py37h0c9df89_0 conda-forge mkl 2020.2 ha770c72_256 conda-forge mkl-include 2021.2.0 h726a3e6_389 conda-forge mkl-service 2.3.0 py37h8f50634_2 conda-forge multidict 5.1.0 pypi_0 pypi munch 2.5.0 py_0 conda-forge ncurses 6.2 h58526e2_4 conda-forge nettle 3.6 he412f7d_0 conda-forge ninja 1.10.2 h4bd325d_0 conda-forge numba 0.53.0 py37h7dd73a4_1 conda-forge numpy 1.20.1 py37haa41c4c_0 conda-forge oauthlib 3.1.0 pypi_0 pypi olefile 0.46 pyh9f0ad1d_1 conda-forge openblas 0.3.12 pthreads_h04b7a96_1 conda-forge openh264 2.1.1 h780b84a_0 conda-forge openjpeg 2.4.0 hb52868f_1 conda-forge openssl 1.1.1k h7f98852_0 conda-forge packaging 20.9 pyh44b312d_0 conda-forge pandas 1.2.3 py37hdc94413_0 conda-forge pillow 8.1.2 py37h4600e1f_1 conda-forge pip 21.0.1 pyhd8ed1ab_0 conda-forge pooch 1.3.0 pyhd8ed1ab_0 conda-forge protobuf 3.15.6 pypi_0 pypi py-cpuinfo 7.0.0 pyh9f0ad1d_0 conda-forge pyasn1 0.4.8 pypi_0 pypi pyasn1-modules 0.2.8 pypi_0 pypi pycparser 2.20 pyh9f0ad1d_2 conda-forge pymongo 3.11.3 pypi_0 pypi pyopenssl 20.0.1 pyhd8ed1ab_0 conda-forge pyparsing 2.4.7 pyh9f0ad1d_0 conda-forge pysocks 1.7.1 py37h89c1867_3 conda-forge pysoundfile 0.10.3.post1 pyhd3deb0d_0 conda-forge python 3.7.10 hffdb5ce_100_cpython conda-forge python-dateutil 2.8.1 py_0 conda-forge python_abi 3.7 1_cp37m conda-forge pytorch-lightning 1.3.0.dev0 dev_0 <develop> pytz 2021.1 pyhd8ed1ab_0 conda-forge pyyaml 5.3.1 pypi_0 pypi readline 8.0 he28a2e2_2 conda-forge requests 2.25.1 pyhd3deb0d_0 conda-forge requests-oauthlib 1.3.0 pypi_0 pypi resampy 0.2.2 py_0 conda-forge rhash 1.4.1 h7f98852_0 conda-forge rsa 4.7.2 pypi_0 pypi sacred 0.8.1 dev_0 <develop> scikit-learn 0.24.1 py37h69acf81_0 conda-forge scipy 1.6.1 py37h14a347d_0 conda-forge setuptools 49.6.0 py37h89c1867_3 conda-forge six 1.15.0 pyh9f0ad1d_0 conda-forge smmap 3.0.5 pyh44b312d_0 conda-forge sqlite 3.34.0 h74cdb3f_0 conda-forge tensorboard 2.4.1 pypi_0 pypi tensorboard-plugin-wit 1.8.0 pypi_0 pypi test-tube 0.7.5 pypi_0 pypi threadpoolctl 2.1.0 pyh5ca1d4c_0 conda-forge timm 0.4.12 pypi_0 pypi tk 8.6.10 h21135ba_1 conda-forge torch 1.10.0a0+gitc51abf8 pypi_0 pypi torchaudio 0.8.2 pypi_0 pypi torchmetrics 0.2.0 pypi_0 pypi torchvision 0.9.2+cu111 pypi_0 pypi tornado 6.1 py37h5e8e339_1 conda-forge tqdm 4.59.0 pypi_0 pypi typing_extensions 3.7.4.3 py_0 conda-forge urllib3 1.26.4 pyhd8ed1ab_0 conda-forge werkzeug 1.0.1 pypi_0 pypi wheel 0.36.2 pypi_0 pypi wrapt 1.12.1 py37h5e8e339_3 conda-forge x264 1!161.3030 h7f98852_0 conda-forge xz 5.2.5 h516909a_1 conda-forge yaml 0.2.5 h516909a_0 conda-forge yarl 1.6.3 pypi_0 pypi zipp 3.4.1 pyhd8ed1ab_0 conda-forge zlib 1.2.11 h516909a_1010 conda-forge zstd 1.5.0 ha95c52a_0 conda-forge
Thank you for your reply. I uninstall all the environment before and install them as you suggest,now I can run the code successfully. Thank you very much!
Hello,when I after running the code following,
and I run the code
but I encontered the issue:
Is this the wrong version of pytorch-lighting and sarced?When I upgrad the pytorch-lighting to the latest version,the issue is solved but the issue with sacred has not been solved. Could you please provide some help?Thank you very much!