leo-pony
acc52e0fe7
[npu] update cann base image and torch 2.4 ( #7061 )
...
* Update base npu container image version:The Python version required for Hugging Face Transformers is >= python3.10
* Fix the bug: arg type of INSTALL_DEEPSPEED shoud been string now.
* Update Ascend CANN, CANN-Kernel and corresponding torch and torch-npu version
* Upgrade torch-npu needs packages' version: torch==2.1.0 and torch-npu==2.4.0.post2
2025-02-25 23:32:01 +08:00
hiyouga
cd14336506
update wechat
2025-01-04 07:59:57 +00:00
shibingli@yeah.net
f1d76786e0
Add ARG HTTP_PROXY in Dockerfile to support HTTP proxy during image building.
2024-12-27 18:31:14 +08:00
shibingli@yeah.net
a3a49b1ea4
Add ARG HTTP_PROXY in Dockerfile to support HTTP proxy during image building.This commit introduces an ARG parameter named HTTP_PROXY in the Dockerfile. This addition allows for the configuration of an HTTP proxy, facilitating image building in environments with network restrictions.
2024-12-27 18:17:17 +08:00
XYZliang
64414905a3
Increase shm_size to 16GB in docker-compose.yml to optimize shared memory allocation for large-scale model fine-tuning tasks.
...
This pull request increases the shm_size parameter in docker-compose.yml to 16GB. The goal is to enhance the LLaMA-Factory framework’s performance for large model fine-tuning tasks by providing sufficient shared memory for efficient data loading and parallel processing.
This PR also addresses the issues discussed in [this comment](https://github.com/hiyouga/LLaMA-Factory/issues/4316#issuecomment-2466270708 ) regarding Shared Memory Limit error.
2024-11-13 10:13:59 +08:00
hoshi-hiyouga
5e8a619d1d
Update Dockerfile
2024-11-02 20:20:26 +08:00
Valerio Mariani
d5b2def7c0
make base image parametric.
...
default `BASE_IMAGE` is nvcr.io/nvidia/pytorch:24.02-py3 for retro-compatibility
2024-10-30 21:53:32 +01:00
hiyouga
3af57795dd
tiny fix
2024-10-11 23:51:54 +08:00
StrangeBytesDev
237e302b5c
Add additional install options to Dockerfiles
2024-09-24 16:54:46 -07:00
HardAndHeavy
b1f9ccb29f
update the ROCm version to 6.2
2024-09-12 23:46:33 +03:00
MengqingCao
b3f4acd1b4
update npu base image
2024-08-21 09:12:38 +00:00
hiyouga
668654b5ad
tiny fix
2024-07-26 11:51:00 +08:00
HardAndHeavy
c8e18a669a
Add ROCm support
2024-07-25 21:29:28 +03:00
hzhaoy
642c6d666f
fix #4780
2024-07-12 00:25:48 +08:00
hzhaoy
a8bf1abf0f
fix #4779
2024-07-12 00:15:15 +08:00
hiyouga
93e6fbb37d
update npu docker
2024-06-30 21:05:31 +08:00
hiyouga
e3141f5f1b
fix docker flashattn
2024-06-28 01:28:59 +08:00
hiyouga
0f421055da
fix docker file
2024-06-27 20:29:16 +08:00
hiyouga
e44a4f07f0
tiny fix
2024-06-27 20:14:48 +08:00
hoshi-hiyouga
64b131dcfa
Merge pull request #4461 from hzhaoy/feature/support-flash-attn
...
support flash-attn in Dockerfile
2024-06-27 20:05:26 +08:00
hoshi-hiyouga
5fcd33fd4f
Update Dockerfile
2024-06-27 19:57:40 +08:00
hoshi-hiyouga
9624af1df3
Update Dockerfile
2024-06-27 19:51:25 +08:00
fanjunliang
bdda0827b3
support docker-npu-[amd64|arm64] build
2024-06-27 15:25:12 +08:00
hzhaoy
e19491b0f0
add flash-attn installation flag in Dockerfile
2024-06-27 00:13:30 +08:00
fanjunliang
8096f94a7d
fix torch-npu dependency
2024-06-26 18:21:42 +08:00
MengqingCao
106647a99d
fix docker-compose path
2024-06-26 02:15:00 +00:00
hzhaoy
c88b1be9f3
support flash-attn in Dockerfile
2024-06-25 15:13:07 +08:00
hiyouga
efb81b25ec
fix #4419
2024-06-25 01:51:29 +08:00
hoshi-hiyouga
721acd8768
Update docker-compose.yml
2024-06-25 00:54:28 +08:00
hoshi-hiyouga
3af936a76d
Update Dockerfile
2024-06-25 00:50:34 +08:00
hoshi-hiyouga
15608d0558
Update docker-compose.yml
2024-06-25 00:46:47 +08:00
hoshi-hiyouga
fce146ab68
Update Dockerfile
2024-06-25 00:46:08 +08:00
hoshi-hiyouga
dcc2e24f5c
Update Dockerfile
2024-06-24 23:41:35 +08:00
MengqingCao
d7207e8ad1
update docker files
...
1. add docker-npu (Dockerfile and docker-compose.yml)
2. move cuda docker to docker-cuda and tiny changes to adapt to the new path
2024-06-24 10:57:36 +00:00
