deepseek v3 model with pipeline parallelism (#1191)

* deepseekv3

* use upload_large_file instead of deprecated multi comit

* add pipeline generation and example

* comment

* get fp16 working

* use mlx==0.22
This commit is contained in:
Awni Hannun
2025-01-09 15:55:53 -08:00
committed by GitHub
parent 40b88eff48
commit 5cae0a60e6
7 changed files with 577 additions and 5 deletions

View File

@@ -1,4 +1,4 @@
mlx>=0.19.2
mlx>=0.22.0
numpy
transformers[sentencepiece]>=4.39.3
protobuf