Skip to content

Pull requests: AI-Hypercomputer/maxtext

New pull request New
134 Open 1,006 Closed
134 Open 1,006 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

bq draft
#1250 opened Feb 7, 2025 by Obliviour Loading…
4 tasks
1
JStS for GPUs Preview Integration
#1249 opened Feb 7, 2025 by parambole Loading…
4 tasks done
Use causal mask type for training or prefill with fixed seqlen when
#1248 opened Feb 7, 2025 by wenchenvincent • Draft
3 of 4 tasks
Ahead of time compilation with auto layouts
#1247 opened Feb 7, 2025 by s2bk Loading…
4 tasks done
@s2bk
Enable mlperf to use mixtral-v1 tokenized dataset to avoid dropping in total_weights when splitting data across multiple host
#1246 opened Feb 7, 2025 by aireenmei Loading…
4 tasks done
Get partition spec only when needed
#1245 opened Feb 7, 2025 by gobbleturk Loading…
4 tasks done
@anfals
Upload HLOs
#1244 opened Feb 6, 2025 by Obliviour • Draft
4 tasks
1
POC for GPU e2e feature testing
#1243 opened Feb 6, 2025 by sfvaroglu • Draft
Onboard DeepSeek MoE with shared experts
#1242 opened Feb 6, 2025 by RissyRan Loading…
4 tasks done
@gagika @gobbleturk
3
[wip]chunked prefill
#1241 opened Feb 6, 2025 by mailvijayasingh Loading…
4 tasks
Gpu inference
#1240 opened Feb 5, 2025 by jwyang-google Loading…
4 tasks
Integrate MaxText with Goodput v5
#1238 opened Feb 4, 2025 by dipannita08 Loading…
4 tasks done
[wip-prototype]chunked prefill on 1k prompt len
#1237 opened Feb 4, 2025 by mailvijayasingh Loading…
4 tasks
Add a basic sampling test
#1235 opened Feb 4, 2025 by liurupeng Loading…
4 tasks done
2
Make float32_qk_product and float32_logits apply during inference
#1225 opened Feb 1, 2025 by philip-essential Loading…
4 tasks done
7
MaxText Trillium Benchmarking
#1224 opened Jan 31, 2025 by anfals • Draft
4 tasks
[NVIDIA] Support FP8 quantization for MOE layers
#1221 opened Jan 31, 2025 by kaixih Loading…
7
Add Pathways Benchmarking Recipes for Scale Testing (and fix bugs)
#1220 opened Jan 31, 2025 by SujeethJinesh • Draft
4 tasks
[DRAFT] Initial commit Maxtext unit tests with Pathways backend.
#1211 opened Jan 28, 2025 by RoshaniN Loading…
4 tasks
Mattdavidow ago rebased [DO NOT SUBMIT]
#1208 opened Jan 28, 2025 by gobbleturk • Draft
4 tasks
Mattdavidow ago refactor
#1207 opened Jan 28, 2025 by gobbleturk • Draft
4 tasks
Paged Attention
#1206 opened Jan 28, 2025 by patemotter Loading…
4 tasks done
2
Run Maxtext unit test on candidate JStS Image
#1204 opened Jan 27, 2025 by parambole Loading…
4 tasks done
add GCP platform check for workload observability
#1203 opened Jan 27, 2025 by jcyang43 • Draft
4 tasks done
[RFC]Prototype the MPMD GPipe-circular with single-host H100=8.
#1202 opened Jan 27, 2025 by lc5211 Loading…
1 task
2
ProTip! What’s not been updated in a month: updated:<2025-01-07.