Upload folder using huggingface_hub
Browse files
README.md
ADDED
@@ -0,0 +1,72 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
tags:
|
3 |
+
- docker
|
4 |
+
- x86
|
5 |
+
- a100
|
6 |
+
- rtx4090
|
7 |
+
- semamba
|
8 |
+
- cuda
|
9 |
+
- pytorch
|
10 |
+
- mamba
|
11 |
+
license: mit
|
12 |
+
library_name: docker
|
13 |
+
datasets: []
|
14 |
+
---
|
15 |
+
|
16 |
+
# x86 SEMamba Docker Image
|
17 |
+
|
18 |
+
This Docker image provides a pre-configured development environment for running [SEMamba](https://github.com/RoyChao19477/SEMamba) models on x86 systems such as NVIDIA A100, RTX 4090, and other CUDA-compatible GPUs. It is optimized for Python 3.12 and PyTorch 2.2.2, built on top of Ubuntu 22.04 with CUDA 12.4.
|
19 |
+
|
20 |
+
---
|
21 |
+
|
22 |
+
## Contents
|
23 |
+
|
24 |
+
- **OS**: Ubuntu 22.04 (x86_64)
|
25 |
+
- **Python**: 3.12 (via Miniconda)
|
26 |
+
- **CUDA**: 12.4 (base image)
|
27 |
+
- **PyTorch**: 2.2.2
|
28 |
+
- **TorchVision**: 0.17.2
|
29 |
+
- **TorchAudio**: 2.2.2
|
30 |
+
- **Mamba-SSM**: 1.2.0
|
31 |
+
- **Essential packages**: git, vim, screen, htop, tmux, openssh, etc.
|
32 |
+
|
33 |
+
---
|
34 |
+
|
35 |
+
## Usage
|
36 |
+
|
37 |
+
### Load Docker Image
|
38 |
+
|
39 |
+
```bash
|
40 |
+
docker load < x86_semamba_py312_pt222_cuda124.tar
|
41 |
+
```
|
42 |
+
|
43 |
+
### Run Container
|
44 |
+
|
45 |
+
```bash
|
46 |
+
docker run --gpus all -it -v $(pwd):/workspace x86_semamba_py312_pt222_cuda124
|
47 |
+
```
|
48 |
+
|
49 |
+
This will mount your current directory into `/workspace` inside the container.
|
50 |
+
|
51 |
+
---
|
52 |
+
|
53 |
+
## Purpose
|
54 |
+
|
55 |
+
- Simplifies setup for high-performance x86 GPU systems
|
56 |
+
- Ideal for SEMamba experiments using FlashAttention and custom CUDA builds
|
57 |
+
- Provides reproducible environment with version-pinned core libraries
|
58 |
+
|
59 |
+
---
|
60 |
+
|
61 |
+
## License & Attribution
|
62 |
+
|
63 |
+
- This Docker image is shared for **non-commercial research purposes**.
|
64 |
+
- All included libraries retain their original licenses.
|
65 |
+
- Based on [PyTorch](https://pytorch.org/), [Miniconda](https://docs.conda.io/en/latest/miniconda.html), and [Mamba](https://github.com/state-spaces/mamba).
|
66 |
+
|
67 |
+
---
|
68 |
+
|
69 |
+
## Maintainer
|
70 |
+
|
71 |
+
For questions or issues, feel free to open a discussion or connect via GitHub.
|
72 |
+
|