-
Notifications
You must be signed in to change notification settings - Fork 226
Insights: google/maxtext
Overview
-
- 8 Merged pull requests
- 2 Open pull requests
- 0 Closed issues
- 1 New issue
Could not load contribution data
Please try again later
8 Pull requests merged by 7 people
-
Update flops calculation to active experts in moe
#678 merged
May 31, 2024 -
Pass metrics client config through to Jetstream
#665 merged
May 31, 2024 -
fix OUTPUT_PATH in v5e/128b.sh
#676 merged
May 31, 2024 -
Add MoE matmul implementation
#659 merged
May 31, 2024 -
integrate nsys profiler into MaxText
#663 merged
May 29, 2024 -
FP8 Clipping Fix
#672 merged
May 29, 2024 -
Add llama2-13b tests
#565 merged
May 28, 2024 -
fix OOM issue running inference microbenchmark with llama13b on v5e4
#670 merged
May 26, 2024
2 Pull requests opened by 2 people
-
Add MoE e2e test on GPU
#673 opened
May 28, 2024 -
Pipeline parallelism
#677 opened
May 31, 2024
1 Issue opened by 1 person
-
llama_or_mistral_ckpt.py file requiring checkpoints in local file system
#674 opened
May 30, 2024
3 Unresolved conversations
Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.
-
Sweep across KV cache layouts
#662 commented on
May 31, 2024 • 14 new comments -
Llama3-8b 🦙
#653 commented on
May 28, 2024 • 5 new comments -
Add KV Cache Layout Control and Attention Tests
#667 commented on
May 31, 2024 • 0 new comments