| .. |
|
__init__.py
|
f8dfac6372
chore: attention refactor and upstream sync apr01 (#365)
|
il y a 11 mois |
|
cpu_executor.py
|
99680b2d23
feat: soft prompts (#589)
|
il y a 6 mois |
|
distributed_gpu_executor.py
|
ae04f57ec1
feat: Pipeline Parallel support (#581)
|
il y a 7 mois |
|
executor_base.py
|
99680b2d23
feat: soft prompts (#589)
|
il y a 6 mois |
|
gpu_executor.py
|
99680b2d23
feat: soft prompts (#589)
|
il y a 6 mois |
|
multiproc_gpu_executor.py
|
0061aea5d5
fix: prevent contention amongst shards by setting OMP_NUM_THREADS=1
|
il y a 6 mois |
|
multiproc_worker_utils.py
|
017b42c517
chore: use fork as the default method for mp backend
|
il y a 7 mois |
|
neuron_executor.py
|
4501ae5f15
fix: neuron executor for adapters
|
il y a 6 mois |
|
openvino_executor.py
|
4150b1ea3a
fix: adapter methods for OpenVINO executor
|
il y a 6 mois |
|
ray_gpu_executor.py
|
1562e073c6
fix: ray worker rank assigment
|
il y a 6 mois |
|
ray_utils.py
|
5240c0da23
fix: avoid unnecessary ray import warnings
|
il y a 6 mois |
|
ray_xpu_executor.py
|
99680b2d23
feat: soft prompts (#589)
|
il y a 6 mois |
|
tpu_executor.py
|
4628caeae6
fix: missed these adapter methods for TPU executor
|
il y a 6 mois |
|
xpu_executor.py
|
99680b2d23
feat: soft prompts (#589)
|
il y a 6 mois |
AlpinDale
4501ae5f15
fix: neuron executor for adapters