Principal Research Scientist working on Deep Learning, Time Series Forecasting, Reinforcement Learning and HPC.
- Berlin, Germany
- @krasul
Highlights
- Pro
Block or Report
Block or report kashif
Report abuse
Contact GitHub support about this user’s behavior. Learn more about reporting abuse.
Report abusePopular repositories
-
ICLR2022-OpenReviewData Public
Forked from evanzd/ICLR2021-OpenReviewData
Crawl & visualize ICLR papers and reviews.
-
-
pytorch-transformer-ts Public
Repository of Transformer based PyTorch Time Series Models
601 contributions in the last year
Less
More
Activity overview
Contributed to
kashif/pytorch-transformer-ts,
huggingface/diffusers,
awslabs/gluonts
and 55 other
repositories
Contribution activity
February 2023
Created 10 commits in 6 repositories
Created 2 repositories
- kashif/peft Python
- kashif/accelerate Python
Created a pull request in huggingface/accelerate that received 5 comments
do not scale gradient in bf16 mode
GradScaler only makes sense in float16 mode and not for bf16 on the GPU. Removed block of code which did this for bf16.
+0
−5
•
5
comments
Opened 5 other pull requests in 4 repositories
awslabs/gluonts
2
merged
huggingface/peft
1
merged
huggingface/transformers
1
merged
huggingface/diffusers
1
merged
Reviewed 3 pull requests in 3 repositories
awslabs/gluonts
1 pull request
huggingface/transformers
1 pull request
huggingface/diffusers
1 pull request
5
contributions
in private repositories
Feb 1 – Feb 10




