Hi, I've installed the latest version of Boltz from the sources. I'm experiencing an error related to the triton package:

Exception: Failed to import Triton-based component: triangle_multiplicative_update:

Not Supported Please make sure to install triton==3.3.0. Other versions may not work! Predicting DataLoader 0: 0%| | 0/1 [00:01<?, ?it/s]

But inspecting the version of the triton package I receive the following:

(boltz2-env) xxxx@xxxx-Predator-PHN16-71:~/boltz_2_affinity_example$ pip3 list | grep triton triton 3.3.0

Any suggestion?

Also for the warning:

"You are using a CUDA device ('NVIDIA GeForce RTX 4070 Laptop GPU') that has Tensor Cores. To properly utilize them, you should set torch.set_float32_matmul_precision('medium' | 'high') which will trade-off precision for performance."

How it is possible to make the suggested settings when calling boltz?

Thanks.

Saverio

PS: The following is the list of the used commands:

(pyenv) xxxx@xxxx-Predator-PHN16-71:~$ source boltz2-env/bin/activate (boltz2-env) xxxx@xxxx-Predator-PHN16-71:~$ boltz --help Usage: boltz [OPTIONS] COMMAND [ARGS]...

Boltz.

Options: --help Show this message and exit.

Commands: predict Run predictions with Boltz.

(boltz2-env) xxxx@xxxx-Predator-PHN16-71:~$ cd boltz_2_affinity_example/ (boltz2-env) xxxx@xxxx-Predator-PHN16-71:~/boltz_2_affinity_example$ boltz predict /home/xxxx/sources/boltz-2.1.1/examples/affinity.yaml --use_msa_server Checking input data. All inputs are already processed. Processing 0 inputs with 0 threads. 0it [00:00, ?it/s] Using bfloat16 Automatic Mixed Precision (AMP) GPU available: True (cuda), used: True TPU available: False, using: 0 TPU cores HPU available: False, using: 0 HPUs /home/xxxx/boltz2-env/lib/python3.12/site-packages/pytorch_lightning/trainer/connectors/logger_connector/logger_connector.py:76: Starting from v1.9.0, tensorboardX has been removed as a dependency of the pytorch_lightning package, due to potential conflicts with other packages in the ML ecosystem. For this reason, logger=True will use CSVLogger as the default logger, unless the tensorboard or tensorboardX packages are found. Please pip install lightning[extra] or one of them to enable TensorBoard support by default Running structure prediction for 1 input. /home/xxxx/boltz2-env/lib/python3.12/site-packages/pytorch_lightning/utilities/migration/utils.py:56: The loaded checkpoint was produced with Lightning v2.5.0.post0, which is newer than your current Lightning version: v2.5.0 You are using a CUDA device ('NVIDIA GeForce RTX 4070 Laptop GPU') that has Tensor Cores. To properly utilize them, you should set torch.set_float32_matmul_precision('medium' | 'high') which will trade-off precision for performance. For more details, read https://pytorch.org/docs/stable/generated/torch.set_float32_matmul_precision.html#torch.set_float32_matmul_precision LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0] Predicting DataLoader 0: 0%| | 0/1 [00:00<?, ?it/s]Traceback (most recent call last): File "/home/xxxx/boltz2-env/bin/boltz", line 8, in sys.exit(cli()) ^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/click/core.py", line 1157, in call return self.main(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/click/core.py", line 1078, in main rv = self.invoke(ctx) ^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/click/core.py", line 1688, in invoke return _process_result(sub_ctx.command.invoke(sub_ctx)) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/click/core.py", line 1434, in invoke return ctx.invoke(self.callback, **ctx.params) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/click/core.py", line 783, in invoke return __callback(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/boltz/main.py", line 1190, in predict trainer.predict( File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/pytorch_lightning/trainer/trainer.py", line 859, in predict return call._call_and_handle_interrupt( ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/pytorch_lightning/trainer/call.py", line 47, in _call_and_handle_interrupt return trainer_fn(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/pytorch_lightning/trainer/trainer.py", line 898, in _predict_impl results = self._run(model, ckpt_path=ckpt_path) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/pytorch_lightning/trainer/trainer.py", line 982, in _run results = self._run_stage() ^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/pytorch_lightning/trainer/trainer.py", line 1021, in _run_stage return self.predict_loop.run() ^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/pytorch_lightning/loops/utilities.py", line 179, in _decorator return loop_run(self, *args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/pytorch_lightning/loops/prediction_loop.py", line 125, in run self._predict_step(batch, batch_idx, dataloader_idx, dataloader_iter) File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/pytorch_lightning/loops/prediction_loop.py", line 255, in _predict_step predictions = call._call_strategy_hook(trainer, "predict_step", *step_args) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/pytorch_lightning/trainer/call.py", line 323, in _call_strategy_hook output = fn(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/pytorch_lightning/strategies/strategy.py", line 438, in predict_step return self.lightning_module.predict_step(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/boltz/model/models/boltz2.py", line 1055, in predict_step out = self( ^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/boltz/model/models/boltz2.py", line 464, in forward z = z + template_module( ^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/boltz/model/modules/trunkv2.py", line 498, in forward v = v + self.pairformer(v, pair_mask, use_kernels=use_kernels) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/boltz/model/layers/pairformer.py", line 331, in forward z = layer( ^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/boltz/model/layers/pairformer.py", line 245, in forward z = z + dropout * self.tri_mul_out( ^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/boltz/model/layers/triangular_mult.py", line 92, in forward return kernel_triangular_mult( ^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/torch/_dynamo/eval_frame.py", line 838, in _fn return fn(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/boltz/model/layers/triangular_mult.py", line 23, in kernel_triangular_mult return triangle_multiplicative_update( ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/cuequivariance_torch/primitives/triangle.py", line 223, in triangle_multiplicative_update return f( ^^ File "/home/xxxx/boltz2-env/lib/python3.12/site-packages/cuequivariance_ops_torch/init.py", line 68, in triangle_multiplicative_update raise Exception(f"Failed to import Triton-based component: triangle_multiplicative_update:\n{IMPORT_EXCEPTION}\n" Exception: Failed to import Triton-based component: triangle_multiplicative_update: Not Supported Please make sure to install triton==3.3.0. Other versions may not work! Predicting DataLoader 0: 0%| | 0/1 [00:01<?, ?it/s]
(boltz2-env) xxxx@xxxx-Predator-PHN16-71:~/boltz_2_affinity_example$ pip3 list | grep triton triton 3.3.0

Jul 02 '25 09:07 xavgit

The multiplication precision is currently hard-coded to highest, but there's a PR (#413) to make it a command-line option.

Jul 02 '25 21:07 xavierholt

Hi, thanks. For the error:

Exception: Failed to import Triton-based component: triangle_multiplicative_update:

Not Supported. Please make sure to install triton==3.3.0. Other versions may not work!

What can be done? I'm using Ubuntu 24.04.

Thanks.

Saverio

Jul 03 '25 00:07 xavgit

I have the same issue with the newest versions, as a temporary workaround I'm now using boltz==2.0.3 and that is working for me.

Jul 03 '25 15:07 coreyhowe999

Ok. Thanks. I hope that this problem will be fixed.

Saverio

Jul 04 '25 08:07 xavgit

I also was also not able to solve the triton issue with other versions of packages. Since boltz 2.1.0, the prediction fails every time. I have to use 2.0.3.

Jul 08 '25 11:07 skacelhonza

Exactly the same problem here. I hope the Boltz team will publish an updated dependency list soon.

Jul 10 '25 15:07 Shredderroy

I address this issue please try open /home/user/anaconda3/envs/boltz2environmen/lib/python3.12/site-packages/cuequivariance_ops/triton/cache_manager.py and change gpu_core_count = 10240（your cuda core count）

Jul 12 '25 09:07 wyb63136

Hi, I've installed boltz2 using python3 -m venv boltz2-env and then compiling from the sources. I've modified cache_manager.py in the directory /home/xxxx/boltz2-env/lib/python3.12/site-packages/cuequivariance_ops/triton from gpu_core_count = pynvml.nvmlDeviceGetNumGpuCores(handle) to gpu_core_count = 5888 # rtx 4070 mobile cuda core count but the error remains. Any suggestions?

Thanks.

Saverio

Jul 12 '25 12:07 xavgit

maybe you should reinstall in a new conda env Use conda create -n boltz2new python=3.12; then use pip install boltz -U , when all done you can change that /home/user/anaconda3/envs/boltz2environmen/lib/python3.12/site-packages/cuequivariance_ops/triton/cache_manager.py replace gpu_core_count = pynvml.nvmlDeviceGetNumGpuCores(handle) with gpu_core_count = 5888 . I think that will work . If not please paste the report. In my opinion I suggest you can try on WSL or linux and change a GPU (memory>16GB)

Jul 12 '25 12:07 wyb63136

Thanks GPT let me tell you why

❗ Error Explanation

The error occurs because:

Inside cuequivariance_ops_torch, the code tries to retrieve GPU information using pynvml, specifically calling a method that is not supported by your current GPU or driver:
pynvml.nvmlDeviceGetNumGpuCores(handle)

Which results in the error:

pynvml.NVMLError_NotSupported

🔍 Root Cause Analysis

The root of the issue:

The function nvmlDeviceGetNumGpuCores is not part of the official NVIDIA NVML API, and it is not a standard function in pynvml.

You can verify this with:

import pynvml
print(dir(pynvml))  # You won't find nvmlDeviceGetNumGpuCores here

This usually means:

The installed pynvml.py file has been custom-modified by some library (possibly cuequivariance), or
The package is calling a private or unofficial helper function that’s not part of the public API.

✅ Solution

Modify the source code to bypass the invalid function call

Locate the error source:

/home/wyb/.pixi/envs/default/lib/python3.12/site-packages/cuequivariance_ops/triton/cache_manager.py

Find this line:

gpu_core_count = pynvml.nvmlDeviceGetNumGpuCores(handle)

And manually comment it out or replace it with something like:

gpu_core_count = 128  # Or set it based on your GPU's actual CUDA core count

If you're using an RTX 4080 SUPER, you could replace it with:

gpu_core_count = 10240

Or, to make it more robust:

try:
    gpu_core_count = pynvml.nvmlDeviceGetNumGpuCores(handle)
except pynvml.NVMLError_NotSupported:
    gpu_core_count = 10240  # Fallback value for RTX 4080 SUPER

Jul 12 '25 12:07 wyb63136

Hi, I've used your suggestions ( before Santo (( italian )) GTP but I receive the same error. I've modified gpu_core_count = 5888

The command and its output: (boltz2) xxxx@xxxx-Predator-PHN16-71:~$ cd boltz_2_affinity_example/ (boltz2) xxxx@xxxx-Predator-PHN16-71:~/boltz_2_affinity_example$ boltz predict /home/xxxx/sources/boltz-2.1.1/examples/affinity.yaml --use_msa_server Checking input data. All inputs are already processed. Processing 0 inputs with 0 threads. 0it [00:00, ?it/s] Using bfloat16 Automatic Mixed Precision (AMP) GPU available: True (cuda), used: True TPU available: False, using: 0 TPU cores HPU available: False, using: 0 HPUs /home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/pytorch_lightning/trainer/connectors/logger_connector/logger_connector.py:76: Starting from v1.9.0, tensorboardX has been removed as a dependency of the pytorch_lightning package, due to potential conflicts with other packages in the ML ecosystem. For this reason, logger=True will use CSVLogger as the default logger, unless the tensorboard or tensorboardX packages are found. Please pip install lightning[extra] or one of them to enable TensorBoard support by default Running structure prediction for 1 input. /home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/pytorch_lightning/utilities/migration/utils.py:56: The loaded checkpoint was produced with Lightning v2.5.0.post0, which is newer than your current Lightning version: v2.5.0 You are using a CUDA device ('NVIDIA GeForce RTX 4070 Laptop GPU') that has Tensor Cores. To properly utilize them, you should set torch.set_float32_matmul_precision('medium' | 'high') which will trade-off precision for performance. For more details, read https://pytorch.org/docs/stable/generated/torch.set_float32_matmul_precision.html#torch.set_float32_matmul_precision LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0] Predicting DataLoader 0: 0%| | 0/1 [00:00<?, ?it/s]Traceback (most recent call last): File "/home/xxxx/anaconda3/envs/boltz2/bin/boltz", line 8, in sys.exit(cli()) ^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/click/core.py", line 1157, in call return self.main(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/click/core.py", line 1078, in main rv = self.invoke(ctx) ^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/click/core.py", line 1688, in invoke return _process_result(sub_ctx.command.invoke(sub_ctx)) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/click/core.py", line 1434, in invoke return ctx.invoke(self.callback, **ctx.params) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/click/core.py", line 783, in invoke return __callback(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/boltz/main.py", line 1190, in predict trainer.predict( File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/pytorch_lightning/trainer/trainer.py", line 859, in predict return call._call_and_handle_interrupt( ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/pytorch_lightning/trainer/call.py", line 47, in _call_and_handle_interrupt return trainer_fn(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/pytorch_lightning/trainer/trainer.py", line 898, in _predict_impl results = self._run(model, ckpt_path=ckpt_path) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/pytorch_lightning/trainer/trainer.py", line 982, in _run results = self._run_stage() ^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/pytorch_lightning/trainer/trainer.py", line 1021, in _run_stage return self.predict_loop.run() ^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/pytorch_lightning/loops/utilities.py", line 179, in _decorator return loop_run(self, *args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/pytorch_lightning/loops/prediction_loop.py", line 125, in run self._predict_step(batch, batch_idx, dataloader_idx, dataloader_iter) File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/pytorch_lightning/loops/prediction_loop.py", line 255, in _predict_step predictions = call._call_strategy_hook(trainer, "predict_step", *step_args) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/pytorch_lightning/trainer/call.py", line 323, in _call_strategy_hook output = fn(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/pytorch_lightning/strategies/strategy.py", line 438, in predict_step return self.lightning_module.predict_step(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/boltz/model/models/boltz2.py", line 1055, in predict_step out = self( ^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/boltz/model/models/boltz2.py", line 464, in forward z = z + template_module( ^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/boltz/model/modules/trunkv2.py", line 498, in forward v = v + self.pairformer(v, pair_mask, use_kernels=use_kernels) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/boltz/model/layers/pairformer.py", line 331, in forward z = layer( ^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/boltz/model/layers/pairformer.py", line 245, in forward z = z + dropout * self.tri_mul_out( ^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/boltz/model/layers/triangular_mult.py", line 92, in forward return kernel_triangular_mult( ^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/torch/_dynamo/eval_frame.py", line 838, in _fn return fn(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/boltz/model/layers/triangular_mult.py", line 23, in kernel_triangular_mult return triangle_multiplicative_update( ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/cuequivariance_torch/primitives/triangle.py", line 223, in triangle_multiplicative_update return f( ^^ File "/home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages/cuequivariance_ops_torch/init.py", line 68, in triangle_multiplicative_update raise Exception(f"Failed to import Triton-based component: triangle_multiplicative_update:\n{IMPORT_EXCEPTION}\n" Exception: Failed to import Triton-based component: triangle_multiplicative_update: Not Supported Please make sure to install triton==3.3.0. Other versions may not work! Predicting DataLoader 0: 0%| | 0/1 [00:01<?, ?it/s]

(boltz2) xxxx@xxxx-Predator-PHN16-71:~/boltz_2_affinity_example$ pip3 list | grep triton triton 3.3.1 (boltz2) xxxx@xxxx-Predator-PHN16-71

Maybe this?

Thanks.

Saverio

Used commands for the installation of boltz 2 with anaconda:

(boltz2-env) xxxx@xxxx-Predator-PHN16-71:~$ deactivate xxxx@xxxx-Predator-PHN16-71:~$ rm -rf boltz2-env/ xxxx@xxxx-Predator-PHN16-71:~$ which python xxxx@xxxx-Predator-PHN16-71:~$ conda create -n boltz2 python=3.12 Retrieving notices: ...working... done Channels:

conda-forge
nodefaults Platform: linux-64 Collecting package metadata (repodata.json): done Solving environment: done

Package Plan

environment location: /home/xxxx/anaconda3/envs/boltz2

added / updated specs: - python=3.12

The following packages will be downloaded:

package                    |            build
---------------------------|-----------------
ca-certificates-2025.7.9   |       hbd8a1cb_0         149 KB  conda-forge
icu-75.1                   |       he02047a_0        11.6 MB  conda-forge
ld_impl_linux-64-2.44      |       h1423503_1         660 KB  conda-forge
libexpat-2.7.0             |       h5888daf_0          73 KB  conda-forge
libgcc-15.1.0              |       h767d61c_3         806 KB  conda-forge
libgcc-ng-15.1.0           |       h69a702a_3          28 KB  conda-forge
libgomp-15.1.0             |       h767d61c_3         437 KB  conda-forge
liblzma-5.8.1              |       hb9d3cd8_2         110 KB  conda-forge
libnsl-2.0.1               |       hb9d3cd8_1          33 KB  conda-forge
libsqlite-3.50.2           |       hee844dc_2         914 KB  conda-forge
libstdcxx-15.1.0           |       h8f9b012_3         3.7 MB  conda-forge
libstdcxx-ng-15.1.0        |       h4852527_3          28 KB  conda-forge
openssl-3.5.1              |       h7b32b05_0         3.0 MB  conda-forge
pip-25.1.1                 |     pyh8b19718_0         1.2 MB  conda-forge
python-3.12.11             |h9e4cc4f_0_cpython        30.0 MB  conda-forge
setuptools-80.9.0          |     pyhff2d567_0         731 KB  conda-forge
tk-8.6.13                  |noxft_hd72426e_102         3.1 MB  conda-forge
------------------------------------------------------------
                                       Total:        56.5 MB

The following NEW packages will be INSTALLED:

_libgcc_mutex conda-forge/linux-64::_libgcc_mutex-0.1-conda_forge _openmp_mutex conda-forge/linux-64::_openmp_mutex-4.5-2_gnu bzip2 conda-forge/linux-64::bzip2-1.0.8-h4bc722e_7 ca-certificates conda-forge/noarch::ca-certificates-2025.7.9-hbd8a1cb_0 icu conda-forge/linux-64::icu-75.1-he02047a_0 ld_impl_linux-64 conda-forge/linux-64::ld_impl_linux-64-2.44-h1423503_1 libexpat conda-forge/linux-64::libexpat-2.7.0-h5888daf_0 libffi conda-forge/linux-64::libffi-3.4.6-h2dba641_1 libgcc conda-forge/linux-64::libgcc-15.1.0-h767d61c_3 libgcc-ng conda-forge/linux-64::libgcc-ng-15.1.0-h69a702a_3 libgomp conda-forge/linux-64::libgomp-15.1.0-h767d61c_3 liblzma conda-forge/linux-64::liblzma-5.8.1-hb9d3cd8_2 libnsl conda-forge/linux-64::libnsl-2.0.1-hb9d3cd8_1 libsqlite conda-forge/linux-64::libsqlite-3.50.2-hee844dc_2 libstdcxx conda-forge/linux-64::libstdcxx-15.1.0-h8f9b012_3 libstdcxx-ng conda-forge/linux-64::libstdcxx-ng-15.1.0-h4852527_3 libuuid conda-forge/linux-64::libuuid-2.38.1-h0b41bf4_0 libxcrypt conda-forge/linux-64::libxcrypt-4.4.36-hd590300_1 libzlib conda-forge/linux-64::libzlib-1.3.1-hb9d3cd8_2 ncurses conda-forge/linux-64::ncurses-6.5-h2d0b736_3 openssl conda-forge/linux-64::openssl-3.5.1-h7b32b05_0 pip conda-forge/noarch::pip-25.1.1-pyh8b19718_0 python conda-forge/linux-64::python-3.12.11-h9e4cc4f_0_cpython readline conda-forge/linux-64::readline-8.2-h8c095d6_2 setuptools conda-forge/noarch::setuptools-80.9.0-pyhff2d567_0 tk conda-forge/linux-64::tk-8.6.13-noxft_hd72426e_102 tzdata conda-forge/noarch::tzdata-2025b-h78e105d_0 wheel conda-forge/noarch::wheel-0.45.1-pyhd8ed1ab_1

Proceed ([y]/n)? y

Downloading and Extracting Packages:

Preparing transaction: done
Verifying transaction: done
Executing transaction: done

To activate this environment, use

$ conda activate boltz2

To deactivate an active environment, use

$ conda deactivate

xxxx@xxxx-Predator-PHN16-71:~$ conda activate boltz2
(boltz2) xxxx@xxxx-Predator-PHN16-71:~$ pip install boltz -U
Collecting boltz
Using cached boltz-2.1.1-py3-none-any.whl.metadata (7.1 kB)
Collecting torch>=2.2 (from boltz) Using cached torch-2.7.1-cp312-cp312-manylinux_2_28_x86_64.whl.metadata (29 kB) Collecting numpy<2.0,>=1.26 (from boltz) Using cached numpy-1.26.4-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (61 kB) Collecting hydra-core==1.3.2 (from boltz) Using cached hydra_core-1.3.2-py3-none-any.whl.metadata (5.5 kB) Collecting pytorch-lightning==2.5.0 (from boltz) Using cached pytorch_lightning-2.5.0-py3-none-any.whl.metadata (21 kB) Collecting rdkit>=2024.3.2 (from boltz) Using cached rdkit-2025.3.3-cp312-cp312-manylinux_2_28_x86_64.whl.metadata (4.0 kB) Collecting dm-tree==0.1.8 (from boltz) Using cached dm_tree-0.1.8-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (1.9 kB) Collecting requests==2.32.3 (from boltz) Using cached requests-2.32.3-py3-none-any.whl.metadata (4.6 kB) Collecting pandas>=2.2.2 (from boltz) Using cached pandas-2.3.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (91 kB) Collecting types-requests (from boltz) Using cached types_requests-2.32.4.20250611-py3-none-any.whl.metadata (2.1 kB) Collecting einops==0.8.0 (from boltz) Using cached einops-0.8.0-py3-none-any.whl.metadata (12 kB) Collecting einx==0.3.0 (from boltz) Using cached einx-0.3.0-py3-none-any.whl.metadata (6.9 kB) Collecting fairscale==0.4.13 (from boltz) Using cached fairscale-0.4.13-py3-none-any.whl Collecting mashumaro==3.14 (from boltz) Using cached mashumaro-3.14-py3-none-any.whl.metadata (114 kB) Collecting modelcif==1.2 (from boltz) Using cached modelcif-1.2-py3-none-any.whl Collecting wandb==0.18.7 (from boltz) Using cached wandb-0.18.7-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (9.7 kB) Collecting click==8.1.7 (from boltz) Using cached click-8.1.7-py3-none-any.whl.metadata (3.0 kB) Collecting pyyaml==6.0.2 (from boltz) Using cached PyYAML-6.0.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (2.1 kB) Collecting biopython==1.84 (from boltz) Using cached biopython-1.84-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (12 kB) Collecting scipy==1.13.1 (from boltz) Using cached scipy-1.13.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (60 kB) Collecting numba==0.61.0 (from boltz) Using cached numba-0.61.0-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (2.8 kB) Collecting gemmi==0.6.5 (from boltz) Using cached gemmi-0.6.5-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (2.3 kB) Collecting scikit-learn==1.6.1 (from boltz) Using cached scikit_learn-1.6.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (18 kB) Collecting chembl_structure_pipeline==1.2.2 (from boltz) Using cached chembl_structure_pipeline-1.2.2-py3-none-any.whl.metadata (3.9 kB) Collecting cuequivariance_ops_cu12>=0.5.0 (from boltz) Using cached cuequivariance_ops_cu12-0.5.1-py3-none-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl.metadata (20 kB) Collecting cuequivariance_ops_torch_cu12>=0.5.0 (from boltz) Using cached cuequivariance_ops_torch_cu12-0.5.1-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.metadata (92 kB) Collecting cuequivariance_torch>=0.5.0 (from boltz) Using cached cuequivariance_torch-0.5.1-py3-none-any.whl.metadata (15 kB) Requirement already satisfied: setuptools>=46.4.0 in ./anaconda3/envs/boltz2/lib/python3.12/site-packages (from chembl_structure_pipeline==1.2.2->boltz) (80.9.0) Collecting sympy (from einx==0.3.0->boltz) Using cached sympy-1.14.0-py3-none-any.whl.metadata (12 kB) Collecting frozendict (from einx==0.3.0->boltz) Using cached frozendict-2.4.6-py312-none-any.whl.metadata (23 kB) Collecting omegaconf<2.4,>=2.2 (from hydra-core==1.3.2->boltz) Using cached omegaconf-2.3.0-py3-none-any.whl.metadata (3.9 kB) Collecting antlr4-python3-runtime==4.9.* (from hydra-core==1.3.2->boltz) Using cached antlr4_python3_runtime-4.9.3-py3-none-any.whl Collecting packaging (from hydra-core==1.3.2->boltz) Using cached packaging-25.0-py3-none-any.whl.metadata (3.3 kB) Collecting typing-extensions>=4.1.0 (from mashumaro==3.14->boltz) Downloading typing_extensions-4.14.1-py3-none-any.whl.metadata (3.0 kB) Collecting ihm>=1.7 (from modelcif==1.2->boltz) Downloading ihm-2.7.tar.gz (392 kB) Preparing metadata (setup.py) ... done Collecting llvmlite<0.45,>=0.44.0dev0 (from numba==0.61.0->boltz) Using cached llvmlite-0.44.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (5.0 kB) Collecting tqdm>=4.57.0 (from pytorch-lightning==2.5.0->boltz) Using cached tqdm-4.67.1-py3-none-any.whl.metadata (57 kB) Collecting fsspec>=2022.5.0 (from fsspec[http]>=2022.5.0->pytorch-lightning==2.5.0->boltz) Using cached fsspec-2025.5.1-py3-none-any.whl.metadata (11 kB) Collecting torchmetrics>=0.7.0 (from pytorch-lightning==2.5.0->boltz) Downloading torchmetrics-1.7.4-py3-none-any.whl.metadata (21 kB) Collecting lightning-utilities>=0.10.0 (from pytorch-lightning==2.5.0->boltz) Using cached lightning_utilities-0.14.3-py3-none-any.whl.metadata (5.6 kB) Collecting charset-normalizer<4,>=2 (from requests==2.32.3->boltz) Using cached charset_normalizer-3.4.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (35 kB) Collecting idna<4,>=2.5 (from requests==2.32.3->boltz) Using cached idna-3.10-py3-none-any.whl.metadata (10 kB) Collecting urllib3<3,>=1.21.1 (from requests==2.32.3->boltz) Using cached urllib3-2.5.0-py3-none-any.whl.metadata (6.5 kB) Collecting certifi>=2017.4.17 (from requests==2.32.3->boltz) Downloading certifi-2025.7.9-py3-none-any.whl.metadata (2.4 kB) Collecting joblib>=1.2.0 (from scikit-learn==1.6.1->boltz) Using cached joblib-1.5.1-py3-none-any.whl.metadata (5.6 kB) Collecting threadpoolctl>=3.1.0 (from scikit-learn==1.6.1->boltz) Using cached threadpoolctl-3.6.0-py3-none-any.whl.metadata (13 kB) Collecting docker-pycreds>=0.4.0 (from wandb==0.18.7->boltz) Using cached docker_pycreds-0.4.0-py2.py3-none-any.whl.metadata (1.8 kB) Collecting gitpython!=3.1.29,>=1.0.0 (from wandb==0.18.7->boltz) Using cached GitPython-3.1.44-py3-none-any.whl.metadata (13 kB) Collecting platformdirs (from wandb==0.18.7->boltz) Using cached platformdirs-4.3.8-py3-none-any.whl.metadata (12 kB) Collecting protobuf!=4.21.0,!=5.28.0,<6,>=3.19.0 (from wandb==0.18.7->boltz) Using cached protobuf-5.29.5-cp38-abi3-manylinux2014_x86_64.whl.metadata (592 bytes) Collecting psutil>=5.0.0 (from wandb==0.18.7->boltz) Using cached psutil-7.0.0-cp36-abi3-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (22 kB) Collecting sentry-sdk>=2.0.0 (from wandb==0.18.7->boltz) Using cached sentry_sdk-2.32.0-py2.py3-none-any.whl.metadata (10 kB) Collecting setproctitle (from wandb==0.18.7->boltz) Using cached setproctitle-1.3.6-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (10 kB) Collecting nvidia-cublas-cu12>=12.5.0 (from cuequivariance_ops_cu12>=0.5.0->boltz) Using cached nvidia_cublas_cu12-12.9.1.4-py3-none-manylinux_2_27_x86_64.whl.metadata (1.7 kB) Collecting pynvml (from cuequivariance_ops_cu12>=0.5.0->boltz) Using cached pynvml-12.0.0-py3-none-any.whl.metadata (5.4 kB) Collecting cuequivariance (from cuequivariance_torch>=0.5.0->boltz) Using cached cuequivariance-0.5.1-py3-none-any.whl.metadata (15 kB) Collecting six>=1.4.0 (from docker-pycreds>=0.4.0->wandb==0.18.7->boltz) Using cached six-1.17.0-py2.py3-none-any.whl.metadata (1.7 kB) Collecting aiohttp!=4.0.0a0,!=4.0.0a1 (from fsspec[http]>=2022.5.0->pytorch-lightning==2.5.0->boltz) Downloading aiohttp-3.12.14-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (7.6 kB) Collecting aiohappyeyeballs>=2.5.0 (from aiohttp!=4.0.0a0,!=4.0.0a1->fsspec[http]>=2022.5.0->pytorch-lightning==2.5.0->boltz) Using cached aiohappyeyeballs-2.6.1-py3-none-any.whl.metadata (5.9 kB) Collecting aiosignal>=1.4.0 (from aiohttp!=4.0.0a0,!=4.0.0a1->fsspec[http]>=2022.5.0->pytorch-lightning==2.5.0->boltz) Downloading aiosignal-1.4.0-py3-none-any.whl.metadata (3.7 kB) Collecting attrs>=17.3.0 (from aiohttp!=4.0.0a0,!=4.0.0a1->fsspec[http]>=2022.5.0->pytorch-lightning==2.5.0->boltz) Using cached attrs-25.3.0-py3-none-any.whl.metadata (10 kB) Collecting frozenlist>=1.1.1 (from aiohttp!=4.0.0a0,!=4.0.0a1->fsspec[http]>=2022.5.0->pytorch-lightning==2.5.0->boltz) Using cached frozenlist-1.7.0-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (18 kB) Collecting multidict<7.0,>=4.5 (from aiohttp!=4.0.0a0,!=4.0.0a1->fsspec[http]>=2022.5.0->pytorch-lightning==2.5.0->boltz) Using cached multidict-6.6.3-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl.metadata (5.3 kB) Collecting propcache>=0.2.0 (from aiohttp!=4.0.0a0,!=4.0.0a1->fsspec[http]>=2022.5.0->pytorch-lightning==2.5.0->boltz) Using cached propcache-0.3.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (12 kB) Collecting yarl<2.0,>=1.17.0 (from aiohttp!=4.0.0a0,!=4.0.0a1->fsspec[http]>=2022.5.0->pytorch-lightning==2.5.0->boltz) Using cached yarl-1.20.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (73 kB) Collecting gitdb<5,>=4.0.1 (from gitpython!=3.1.29,>=1.0.0->wandb==0.18.7->boltz) Using cached gitdb-4.0.12-py3-none-any.whl.metadata (1.2 kB) Collecting smmap<6,>=3.0.1 (from gitdb<5,>=4.0.1->gitpython!=3.1.29,>=1.0.0->wandb==0.18.7->boltz) Using cached smmap-5.0.2-py3-none-any.whl.metadata (4.3 kB) Collecting msgpack (from ihm>=1.7->modelcif==1.2->boltz) Using cached msgpack-1.1.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (8.4 kB) Collecting python-dateutil>=2.8.2 (from pandas>=2.2.2->boltz) Using cached python_dateutil-2.9.0.post0-py2.py3-none-any.whl.metadata (8.4 kB) Collecting pytz>=2020.1 (from pandas>=2.2.2->boltz) Using cached pytz-2025.2-py2.py3-none-any.whl.metadata (22 kB) Collecting tzdata>=2022.7 (from pandas>=2.2.2->boltz) Using cached tzdata-2025.2-py2.py3-none-any.whl.metadata (1.4 kB) Collecting Pillow (from rdkit>=2024.3.2->boltz) Using cached pillow-11.3.0-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.metadata (9.0 kB) Collecting filelock (from torch>=2.2->boltz) Using cached filelock-3.18.0-py3-none-any.whl.metadata (2.9 kB) Collecting networkx (from torch>=2.2->boltz) Using cached networkx-3.5-py3-none-any.whl.metadata (6.3 kB) Collecting jinja2 (from torch>=2.2->boltz) Using cached jinja2-3.1.6-py3-none-any.whl.metadata (2.9 kB) Collecting nvidia-cuda-nvrtc-cu12==12.6.77 (from torch>=2.2->boltz) Using cached nvidia_cuda_nvrtc_cu12-12.6.77-py3-none-manylinux2014_x86_64.whl.metadata (1.5 kB) Collecting nvidia-cuda-runtime-cu12==12.6.77 (from torch>=2.2->boltz) Using cached nvidia_cuda_runtime_cu12-12.6.77-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.5 kB) Collecting nvidia-cuda-cupti-cu12==12.6.80 (from torch>=2.2->boltz) Using cached nvidia_cuda_cupti_cu12-12.6.80-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.6 kB) Collecting nvidia-cudnn-cu12==9.5.1.17 (from torch>=2.2->boltz) Using cached nvidia_cudnn_cu12-9.5.1.17-py3-none-manylinux_2_28_x86_64.whl.metadata (1.6 kB) Collecting nvidia-cublas-cu12>=12.5.0 (from cuequivariance_ops_cu12>=0.5.0->boltz) Using cached nvidia_cublas_cu12-12.6.4.1-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.5 kB) Collecting nvidia-cufft-cu12==11.3.0.4 (from torch>=2.2->boltz) Using cached nvidia_cufft_cu12-11.3.0.4-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.5 kB) Collecting nvidia-curand-cu12==10.3.7.77 (from torch>=2.2->boltz) Using cached nvidia_curand_cu12-10.3.7.77-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.5 kB) Collecting nvidia-cusolver-cu12==11.7.1.2 (from torch>=2.2->boltz) Using cached nvidia_cusolver_cu12-11.7.1.2-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.6 kB) Collecting nvidia-cusparse-cu12==12.5.4.2 (from torch>=2.2->boltz) Using cached nvidia_cusparse_cu12-12.5.4.2-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.6 kB) Collecting nvidia-cusparselt-cu12==0.6.3 (from torch>=2.2->boltz) Using cached nvidia_cusparselt_cu12-0.6.3-py3-none-manylinux2014_x86_64.whl.metadata (6.8 kB) Collecting nvidia-nccl-cu12==2.26.2 (from torch>=2.2->boltz) Using cached nvidia_nccl_cu12-2.26.2-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (2.0 kB) Collecting nvidia-nvtx-cu12==12.6.77 (from torch>=2.2->boltz) Using cached nvidia_nvtx_cu12-12.6.77-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.6 kB) Collecting nvidia-nvjitlink-cu12==12.6.85 (from torch>=2.2->boltz) Using cached nvidia_nvjitlink_cu12-12.6.85-py3-none-manylinux2010_x86_64.manylinux_2_12_x86_64.whl.metadata (1.5 kB) Collecting nvidia-cufile-cu12==1.11.1.6 (from torch>=2.2->boltz) Using cached nvidia_cufile_cu12-1.11.1.6-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.5 kB) Collecting triton==3.3.1 (from torch>=2.2->boltz) Using cached triton-3.3.1-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.metadata (1.5 kB) Collecting mpmath<1.4,>=1.1.0 (from sympy->einx==0.3.0->boltz) Using cached mpmath-1.3.0-py3-none-any.whl.metadata (8.6 kB) Collecting opt-einsum (from cuequivariance->cuequivariance_torch>=0.5.0->boltz) Using cached opt_einsum-3.4.0-py3-none-any.whl.metadata (6.3 kB) Collecting MarkupSafe>=2.0 (from jinja2->torch>=2.2->boltz) Using cached MarkupSafe-3.0.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (4.0 kB) Collecting nvidia-ml-py<13.0.0a0,>=12.0.0 (from pynvml->cuequivariance_ops_cu12>=0.5.0->boltz) Using cached nvidia_ml_py-12.575.51-py3-none-any.whl.metadata (9.3 kB) Using cached boltz-2.1.1-py3-none-any.whl (262 kB) Using cached biopython-1.84-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (3.2 MB) Using cached chembl_structure_pipeline-1.2.2-py3-none-any.whl (17 kB) Using cached click-8.1.7-py3-none-any.whl (97 kB) Using cached dm_tree-0.1.8-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (152 kB) Using cached einops-0.8.0-py3-none-any.whl (43 kB) Using cached einx-0.3.0-py3-none-any.whl (102 kB) Using cached gemmi-0.6.5-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (3.2 MB) Using cached hydra_core-1.3.2-py3-none-any.whl (154 kB) Using cached mashumaro-3.14-py3-none-any.whl (92 kB) Using cached numba-0.61.0-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (3.9 MB) Using cached pytorch_lightning-2.5.0-py3-none-any.whl (819 kB) Using cached PyYAML-6.0.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (767 kB) Using cached requests-2.32.3-py3-none-any.whl (64 kB) Using cached scikit_learn-1.6.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (13.1 MB) Using cached scipy-1.13.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (38.2 MB) Using cached wandb-0.18.7-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (16.1 MB) Using cached charset_normalizer-3.4.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (148 kB) Using cached idna-3.10-py3-none-any.whl (70 kB) Using cached llvmlite-0.44.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (42.4 MB) Using cached numpy-1.26.4-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (18.0 MB) Using cached omegaconf-2.3.0-py3-none-any.whl (79 kB) Using cached protobuf-5.29.5-cp38-abi3-manylinux2014_x86_64.whl (319 kB) Using cached urllib3-2.5.0-py3-none-any.whl (129 kB) Downloading certifi-2025.7.9-py3-none-any.whl (159 kB) Using cached cuequivariance_ops_cu12-0.5.1-py3-none-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl (31.1 MB) Using cached cuequivariance_ops_torch_cu12-0.5.1-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (196 kB) Using cached cuequivariance_torch-0.5.1-py3-none-any.whl (56 kB) Using cached docker_pycreds-0.4.0-py2.py3-none-any.whl (9.0 kB) Using cached fsspec-2025.5.1-py3-none-any.whl (199 kB) Downloading aiohttp-3.12.14-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.7 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.7/1.7 MB 6.9 MB/s eta 0:00:00 Using cached multidict-6.6.3-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl (256 kB) Using cached yarl-1.20.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (355 kB) Using cached aiohappyeyeballs-2.6.1-py3-none-any.whl (15 kB) Downloading aiosignal-1.4.0-py3-none-any.whl (7.5 kB) Using cached attrs-25.3.0-py3-none-any.whl (63 kB) Using cached frozenlist-1.7.0-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (241 kB) Using cached GitPython-3.1.44-py3-none-any.whl (207 kB) Using cached gitdb-4.0.12-py3-none-any.whl (62 kB) Using cached smmap-5.0.2-py3-none-any.whl (24 kB) Using cached joblib-1.5.1-py3-none-any.whl (307 kB) Using cached lightning_utilities-0.14.3-py3-none-any.whl (28 kB) Using cached packaging-25.0-py3-none-any.whl (66 kB) Using cached pandas-2.3.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (12.0 MB) Using cached propcache-0.3.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (224 kB) Using cached psutil-7.0.0-cp36-abi3-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (277 kB) Using cached python_dateutil-2.9.0.post0-py2.py3-none-any.whl (229 kB) Using cached pytz-2025.2-py2.py3-none-any.whl (509 kB) Using cached rdkit-2025.3.3-cp312-cp312-manylinux_2_28_x86_64.whl (34.8 MB) Using cached sentry_sdk-2.32.0-py2.py3-none-any.whl (356 kB) Using cached six-1.17.0-py2.py3-none-any.whl (11 kB) Using cached threadpoolctl-3.6.0-py3-none-any.whl (18 kB) Using cached torch-2.7.1-cp312-cp312-manylinux_2_28_x86_64.whl (821.0 MB) Using cached nvidia_cublas_cu12-12.6.4.1-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (393.1 MB) Using cached nvidia_cuda_cupti_cu12-12.6.80-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (8.9 MB) Using cached nvidia_cuda_nvrtc_cu12-12.6.77-py3-none-manylinux2014_x86_64.whl (23.7 MB) Using cached nvidia_cuda_runtime_cu12-12.6.77-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (897 kB) Using cached nvidia_cudnn_cu12-9.5.1.17-py3-none-manylinux_2_28_x86_64.whl (571.0 MB) Using cached nvidia_cufft_cu12-11.3.0.4-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (200.2 MB) Using cached nvidia_cufile_cu12-1.11.1.6-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (1.1 MB) Using cached nvidia_curand_cu12-10.3.7.77-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (56.3 MB) Using cached nvidia_cusolver_cu12-11.7.1.2-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (158.2 MB) Using cached nvidia_cusparse_cu12-12.5.4.2-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (216.6 MB) Using cached nvidia_cusparselt_cu12-0.6.3-py3-none-manylinux2014_x86_64.whl (156.8 MB) Using cached nvidia_nccl_cu12-2.26.2-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (201.3 MB) Using cached nvidia_nvjitlink_cu12-12.6.85-py3-none-manylinux2010_x86_64.manylinux_2_12_x86_64.whl (19.7 MB) Using cached nvidia_nvtx_cu12-12.6.77-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (89 kB) Using cached triton-3.3.1-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (155.7 MB) Using cached sympy-1.14.0-py3-none-any.whl (6.3 MB) Using cached mpmath-1.3.0-py3-none-any.whl (536 kB) Downloading torchmetrics-1.7.4-py3-none-any.whl (963 kB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 963.5/963.5 kB 7.6 MB/s eta 0:00:00 Using cached tqdm-4.67.1-py3-none-any.whl (78 kB) Downloading typing_extensions-4.14.1-py3-none-any.whl (43 kB) Using cached tzdata-2025.2-py2.py3-none-any.whl (347 kB) Using cached cuequivariance-0.5.1-py3-none-any.whl (126 kB) Using cached filelock-3.18.0-py3-none-any.whl (16 kB) Using cached frozendict-2.4.6-py312-none-any.whl (16 kB) Using cached jinja2-3.1.6-py3-none-any.whl (134 kB) Using cached MarkupSafe-3.0.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (23 kB) Using cached msgpack-1.1.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (426 kB) Using cached networkx-3.5-py3-none-any.whl (2.0 MB) Using cached opt_einsum-3.4.0-py3-none-any.whl (71 kB) Using cached pillow-11.3.0-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (6.6 MB) Using cached platformdirs-4.3.8-py3-none-any.whl (18 kB) Using cached pynvml-12.0.0-py3-none-any.whl (26 kB) Using cached nvidia_ml_py-12.575.51-py3-none-any.whl (47 kB) Using cached setproctitle-1.3.6-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (31 kB) Using cached types_requests-2.32.4.20250611-py3-none-any.whl (20 kB) Building wheels for collected packages: ihm DEPRECATION: Building 'ihm' using the legacy setup.py bdist_wheel mechanism, which will be removed in a future version. pip 25.3 will enforce this behaviour change. A possible replacement is to use the standardized build interface by setting the --use-pep517 option, (possibly combined with --no-build-isolation), or adding a pyproject.toml file to the source tree of 'ihm'. Discussion can be found at https://github.com/pypa/pip/issues/6334 Building wheel for ihm (setup.py) ... done Created wheel for ihm: filename=ihm-2.7-cp312-cp312-linux_x86_64.whl size=235613 sha256=276eef366b760387e335bbf33221e2931f286e1e7176e88c1353d15d762631e0 Stored in directory: /home/xxxx/.cache/pip/wheels/b0/21/5f/1358a14f7c48c79a4060ecce6eb1d2bb78dc72f5e5050edd28 Successfully built ihm Installing collected packages: pytz, nvidia-ml-py, nvidia-cusparselt-cu12, mpmath, dm-tree, antlr4-python3-runtime, urllib3, tzdata, typing-extensions, triton, tqdm, threadpoolctl, sympy, smmap, six, setproctitle, pyyaml, pynvml, psutil, protobuf, propcache, platformdirs, Pillow, packaging, opt-einsum, nvidia-nvtx-cu12, nvidia-nvjitlink-cu12, nvidia-nccl-cu12, nvidia-curand-cu12, nvidia-cufile-cu12, nvidia-cuda-runtime-cu12, nvidia-cuda-nvrtc-cu12, nvidia-cuda-cupti-cu12, nvidia-cublas-cu12, numpy, networkx, multidict, msgpack, MarkupSafe, llvmlite, joblib, idna, gemmi, fsspec, frozenlist, frozendict, filelock, einops, click, charset-normalizer, certifi, attrs, aiohappyeyeballs, yarl, types-requests, sentry-sdk, scipy, requests, rdkit, python-dateutil, omegaconf, nvidia-cusparse-cu12, nvidia-cufft-cu12, nvidia-cudnn-cu12, numba, mashumaro, lightning-utilities, jinja2, ihm, gitdb, einx, docker-pycreds, cuequivariance_ops_cu12, biopython, aiosignal, scikit-learn, pandas, nvidia-cusolver-cu12, modelcif, hydra-core, gitpython, cuequivariance_ops_torch_cu12, cuequivariance, chembl_structure_pipeline, aiohttp, wandb, torch, cuequivariance_torch, torchmetrics, fairscale, pytorch-lightning, boltz ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. packmol-memgen 2024.3.27 requires matplotlib, which is not installed. edgembar 3.0 requires matplotlib, which is not installed. ndfes 3.0 requires matplotlib, which is not installed. Successfully installed MarkupSafe-3.0.2 Pillow-11.3.0 aiohappyeyeballs-2.6.1 aiohttp-3.12.14 aiosignal-1.4.0 antlr4-python3-runtime-4.9.3 attrs-25.3.0 biopython-1.84 boltz-2.1.1 certifi-2025.7.9 charset-normalizer-3.4.2 chembl_structure_pipeline-1.2.2 click-8.1.7 cuequivariance-0.5.1 cuequivariance_ops_cu12-0.5.1 cuequivariance_ops_torch_cu12-0.5.1 cuequivariance_torch-0.5.1 dm-tree-0.1.8 docker-pycreds-0.4.0 einops-0.8.0 einx-0.3.0 fairscale-0.4.13 filelock-3.18.0 frozendict-2.4.6 frozenlist-1.7.0 fsspec-2025.5.1 gemmi-0.6.5 gitdb-4.0.12 gitpython-3.1.44 hydra-core-1.3.2 idna-3.10 ihm-2.7 jinja2-3.1.6 joblib-1.5.1 lightning-utilities-0.14.3 llvmlite-0.44.0 mashumaro-3.14 modelcif-1.2 mpmath-1.3.0 msgpack-1.1.1 multidict-6.6.3 networkx-3.5 numba-0.61.0 numpy-1.26.4 nvidia-cublas-cu12-12.6.4.1 nvidia-cuda-cupti-cu12-12.6.80 nvidia-cuda-nvrtc-cu12-12.6.77 nvidia-cuda-runtime-cu12-12.6.77 nvidia-cudnn-cu12-9.5.1.17 nvidia-cufft-cu12-11.3.0.4 nvidia-cufile-cu12-1.11.1.6 nvidia-curand-cu12-10.3.7.77 nvidia-cusolver-cu12-11.7.1.2 nvidia-cusparse-cu12-12.5.4.2 nvidia-cusparselt-cu12-0.6.3 nvidia-ml-py-12.575.51 nvidia-nccl-cu12-2.26.2 nvidia-nvjitlink-cu12-12.6.85 nvidia-nvtx-cu12-12.6.77 omegaconf-2.3.0 opt-einsum-3.4.0 packaging-25.0 pandas-2.3.1 platformdirs-4.3.8 propcache-0.3.2 protobuf-5.29.5 psutil-7.0.0 pynvml-12.0.0 python-dateutil-2.9.0.post0 pytorch-lightning-2.5.0 pytz-2025.2 pyyaml-6.0.2 rdkit-2025.3.3 requests-2.32.3 scikit-learn-1.6.1 scipy-1.13.1 sentry-sdk-2.32.0 setproctitle-1.3.6 six-1.17.0 smmap-5.0.2 sympy-1.14.0 threadpoolctl-3.6.0 torch-2.7.1 torchmetrics-1.7.4 tqdm-4.67.1 triton-3.3.1 types-requests-2.32.4.20250611 typing-extensions-4.14.1 tzdata-2025.2 urllib3-2.5.0 wandb-0.18.7 yarl-1.20.1 (boltz2) xxxx@xxxx-Predator-PHN16-71:~$ boltz Usage: boltz [OPTIONS] COMMAND [ARGS]...

Boltz.

Options: --help Show this message and exit.

Commands: predict Run predictions with Boltz.

Jul 12 '25 13:07 xavgit

Hi, with (boltz2) xxxx@xxxx-Predator-PHN16-71:~/boltz_2_affinity_example$ pip3 install triton==3.3.0 Collecting triton==3.3.0 Using cached triton-3.3.0-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.metadata (1.5 kB) Requirement already satisfied: setuptools>=40.8.0 in /home/xxxx/anaconda3/envs/boltz2/lib/python3.12/site-packages (from triton==3.3.0) (80.9.0) Using cached triton-3.3.0-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (156.5 MB) Installing collected packages: triton Attempting uninstall: triton Found existing installation: triton 3.3.1 Uninstalling triton-3.3.1: Successfully uninstalled triton-3.3.1 ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. torch 2.7.1 requires triton==3.3.1; platform_system == "Linux" and platform_machine == "x86_64", but you have triton 3.3.0 which is incompatible. Successfully installed triton-3.3.0

The same error.

Thanks.

Saverio

Jul 12 '25 13:07 xavgit

very strange issue I've try my best，If you want use boltz2 maybe you should try online service on https://rowansci.com/blog/how-to-run-boltz-2 or follow this page using pixi instead of conda（maybe you should install pixi ）which i‘m using just fine

Good luck！ yb W

Jul 12 '25 13:07 wyb63136

Hi, with pixi exactly the same error.

Saverio

Jul 12 '25 14:07 xavgit

Hi @wyb63136 @xavgit @coreyhowe999 @Shredderroy @xavierholt,

The solution here should be to add the flag --no_kernels to the prediction. The issue occurs in older NVIDIA devices that do not support the latest kernels.

Let me know if you still have issues.

Jul 15 '25 00:07 gcorso

Hi gcorso, thanks for your suggestion. For which version of boltz it should work? I've installed the latest version using a fresh python env. I'm receiving a very long and repeating error calling both boltz predict /home/xxxx/sources/boltz_2.2.0/examples/affinity.yaml --use_msa_server --no_kernels and boltz predict /home/xxxx/sources/boltz_2.2.0/examples/affinity.yaml --use_msa_server

The gpu is: (boltz2-env) xxxx@xxxx-Predator-PHN16-71:~/boltz_2_affinity_example$ lspci -v | grep VGA 0000:00:02.0 VGA compatible controller: Intel Corporation Raptor Lake-S UHD Graphics (rev 04) (prog-if 00 [VGA controller]) 0000:01:00.0 VGA compatible controller: NVIDIA Corporation AD106M [GeForce RTX 4070 Max-Q / Mobile] (rev a1) (prog-if 00 [VGA controller])

I'm wrong somewhere?

Thanks.

Saverio

boltz_cmd_--no_kernels.log

boltz_cmd.log

Jul 15 '25 08:07 xavgit

Hi gcorso, if this can be useful , with boltz 2.1.1 there are no errors.

Thanks for the suggestion.

Saverio

Jul 17 '25 15:07 xavgit

Same problem. Use 2.1.1 and add no_kernels param works for me.

pip install 'boltz[cuda]==2.1.1' -U

boltz predict  examples/prot_custom_msa.yaml  --no_kernels

Aug 02 '25 19:08 charlesxu90

This solution worked well for me https://github.com/jwohlwend/boltz/issues/355. I previously had the same issue, and now it seems to be running (and also faster, as expected).

Oct 17 '25 12:10 RohanEapen

Problem with the triton package when calling boltz predict

Thanks GPT let me tell you why

❗ Error Explanation

🔍 Root Cause Analysis

The root of the issue:

✅ Solution

Modify the source code to bypass the invalid function call

Package Plan

To activate this environment, use

$ conda activate boltz2

To deactivate an active environment, use

$ conda deactivate