Skip to content

[doc] complete README of colossal inference #4585

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
tiandiao123 merged 2 commits into from
Sep 1, 2023
Merged

[doc] complete README of colossal inference #4585

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
aeeb71a
complete fig
CjhHa1 Sep 1, 2023
611f358
Update README.md
CjhHa1 Sep 1, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
32 changes: 25 additions & 7 deletions colossalai/inference/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,7 @@ dependencies

```bash
pytorch= 1.13.1 (gpu)
cuda>= 11.6
cuda>= 11.6
transformers= 4.30.2
triton==2.0.0.dev20221202
# for install vllm, please use this branch to install https://github.com/tiandiao123/vllm/tree/setup_branch
Expand All @@ -66,11 +66,11 @@ flash-attention

### Docker

You can use docker run to use docker container to set-up environment
You can use docker run to use docker container to set-up environment

```
# env: python==3.8, cuda 11.6, pytorch == 1.13.1 triton==2.0.0.dev20221202, vllm kernels support, flash-attention-2 kernels support
docker pull hpcaitech/colossalai-inference:v2
# env: python==3.8, cuda 11.6, pytorch == 1.13.1 triton==2.0.0.dev20221202, vllm kernels support, flash-attention-2 kernels support
docker pull hpcaitech/colossalai-inference:v2
docker run -it --gpus all --name ANY_NAME -v $PWD:/workspace -w /workspace hpcaitech/colossalai-inference:v2 /bin/bash

```
Expand All @@ -88,10 +88,28 @@ python xx

### environment:

We conducted [benchmark tests](https://github.com/hpcaitech/ColossalAI/blob/main/colossalai/shardformer/examples/performance_benchmark.py) to evaluate the performance. We compared the inference `latency` and `throughputs` between `colossal-inference` and `torch`.
We conducted multiple benchmark tests to evaluate the performance. We compared the inference `latency` and `throughputs` between `colossal-inference` and original `hugging-face torch fp16`.

We set the batch size to 4, the number of attention heads to 8, and the head dimension to 64. `N_CTX` refers to the sequence length.
For various models, experiments were conducted using multiple batch sizes under the consistent model configuration of `7 billion(7b)` parameters, `1024` input length, and 128 output length. The obtained results are as follows (due to time constraints, the evaluation has currently been performed solely on the `A100` single GPU performance; multi-GPU performance will be addressed in the future):

In the case of using 2 GPUs, the results are as follows.
### Single GPU Performance:

#### Llama

| batch_size | 8 | 16 | 32 |
| :---------------------: | :----: | :----: | :----: |
| hugging-face torch fp16 | 199.12 | 246.56 | 246.56 |
| colossal-inference | 241.12 | 451.84 | 643.52 |

![llama](https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/inference/Infer-llama.png)

###

| batch_size | 4 | 8 |
| :---------------------: | :----: | :----: |
| hugging-face torch fp16 | 145.28 | 189.68 |
| colossal-inference | 187.48 | 323.28 |

![bloom](https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/inference/Infer-bloom.png)

The results of more models are coming soon!