Good evening,
I would like to try this library for research purposes, but tracking terminate with the following error:
AttributeError Traceback (most recent call last)
Cell In[11], line 5
3 from torch.optim import AdamW
4 from experiment_impact_tracker.compute_tracker import ImpactTracker
----> 5 tracker = ImpactTracker("eit_log")
7 # Accelerate is a library that enables the same PyTorch code to be run across any distributed configuration
8 accelerator = Accelerator()
File ~/miniconda3/lib/python3.9/site-packages/experiment_impact_tracker/compute_tracker.py:173, in ImpactTracker.init(self, logdir)
171 self._setup_logging()
172 self.logger.warn("Gathering system info for reproducibility...")
--> 173 self.initial_info = gather_initial_info(logdir)
174 self.logger.warn("Done initial setup and information gathering...")
File ~/miniconda3/lib/python3.9/site-packages/experiment_impact_tracker/compute_tracker.py:150, in gather_initial_info(logdir)
148 compatabilities = info["compatability"]
149 if _validatecompatabilities(compatabilities):
--> 150 data[key] = info["routing"]["function"]()
152 with open(info_path, 'wb') as info_file:
153 pickle.dump(data, info_file)
File ~/miniconda3/lib/python3.9/site-packages/experiment_impact_tracker/gpu/nvidia.py:47, in get_gpu_info(*args, **kwargs)
44 driver_version = xml.findall('driver_version')[0].text
45 cuda_version = xml.findall('cuda_version')[0].text
---> 47 for gpu_id, gpu in enumerate(xml.getiterator('gpu')):
48 gpu_data = {}
49 name = [x for x in gpu.getiterator('product_name')][0].text
AttributeError: 'xml.etree.ElementTree.Element' object has no attribute 'getiterator'
My environment has:
python 3.9
experiment-impact-tracker 0.1.8
I tried to install xlrd, as suggested in: stackoverflow_answer (both latest and 1.2.0 version) but it still gives me the same error.
Good evening, I would like to try this library for research purposes, but tracking terminate with the following error:
My environment has:
I tried to install xlrd, as suggested in: stackoverflow_answer (both latest and 1.2.0 version) but it still gives me the same error.