 AlpinDale
|
eef647deab
fix: greedy decoding in TPU
|
vor 4 Monaten |
|
AlpinDale
|
8d88814475
chore: reduce XLA compile times
|
vor 4 Monaten |
|
AlpinDale
|
386ad8bef6
feat: tensor parallelism for TPU with ray
|
vor 4 Monaten |
|
AlpinDale
|
f91991f584
fix: f-string fixes
|
vor 4 Monaten |
|
AlpinDale
|
b6c4dfce23
chore: refactor TPU model runner and worker
|
vor 4 Monaten |
|
AlpinDale
|
8c2dd39500
chore: remove multimodal stuff from TPU
|
vor 4 Monaten |
|
AlpinDale
|
e1475fbec7
feat: MoE support with Pallas GMM kernel for TPUs
|
vor 4 Monaten |
|
AlpinDale
|
1cb06835a0
fix: TPU multimodal kwargs and outlines installation in TPU docker
|
vor 4 Monaten |
|
AlpinDale
|
4f7d212b70
feat: remove vision language config
|
vor 4 Monaten |
|
AlpinDale
|
4599c98f99
feat: dynamic image size support for VLMs
|
vor 4 Monaten |
|
AlpinDale
|
301ec7c77d
fix: pad slot id in tpu runner
|
vor 5 Monaten |
|
AlpinDale
|
cdff8e89f9
feat: introduce `DraftModelRunner`
|
vor 5 Monaten |
|
AlpinDale
|
85ef2fe8b1
chore: clean up placeholder symbols
|
vor 5 Monaten |
|
AlpinDale
|
fcd58614f4
feat: support parallel sampling and swapping in TPU
|
vor 5 Monaten |
|
AlpinDale
|
d36b88b301
fix: raise errors if using unsupported samplers on TPU
|
vor 5 Monaten |
|
AlpinDale
|
608e8e1310
chore: refactor TPU backend to make it more similar to GPU backend
|
vor 5 Monaten |
 AlpinDale
|
fe21123a1c
feat: TPU support (#570)
|
vor 5 Monaten |
