Toolmaker. Software creator, optimizer and harmonizer.
Current domains: LLM/Scalability/NLP/Machine Learning.
-
Stasosphere Online Inc.
- Nanaimo, BC, Canada
- https://stasosphere.com/machine-learning/
- @StasBekman
Block or Report
Block or report stas00
Report abuse
Contact GitHub support about this user’s behavior. Learn more about reporting abuse.
Report abusePinned
-
ipyexperiments Public
jupyter/ipython experiment containers for GPU and general RAM re-use
-
-
1,928 contributions in the last year
Activity overview
Contributed to
bigscience-workshop/bigscience,
huggingface/transformers,
bigscience-workshop/Megatron-DeepSpeed
and 39 other
repositories
Contribution activity
May 2022
Created 10 commits in 4 repositories
Created a pull request in microsoft/DeepSpeed that received 4 comments
[pipe] prevent deadlock with multiple evals sequence
This PR solves a race condition that leads to deadlocks at random times at eval@pipe. Here is the diagnostics with tracebacks: https://github.com/b…
+4
−0
•
4
comments
Opened 4 other pull requests in 2 repositories
huggingface/transformers
2
merged
1
open
microsoft/DeepSpeed
1
merged
Reviewed 8 pull requests in 2 repositories
huggingface/transformers
7 pull requests
- Bigscience176b
- Add Mistral GPT-2 Stability Tweaks
- Update self-push workflow
- Extend Transformers Trainer Class to Enable PyTorch SGD/Adagrad Optimizers for Training
- Extend Transformers Trainer Class to Enable CPU AMP and Integrate Intel Extension for PyTorch
- Fix self-push CI report path in cat
- Move test model folders
bigscience-workshop/Megatron-DeepSpeed
1 pull request
Created an issue in pytorch/pytorch that received 3 comments
[distributed] c10d crashing on assert
3
comments