root
/
tabby
mirror of https://gitee.com/zhang_1334717033/tabby.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

docs: add blog for v0.1.1 release (#454)

release-0.2
Meng Zhang 2023-09-17 17:31:41 +08:00 committed by GitHub
parent c107c991ff
commit 0b4206e6f8
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 30 additions and 24 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
website/blog/2023-09-13-release-0-1-0-metal/index.md
View File

@ -1,21 +0,0 @@
---
authors: [ meng ]
---
# Highlights of Tabby v0.1.0: Apple M1/M2 Support
We are thrilled to announce the release of Tabby v0.1.0👏🏻.
Thanks to [llama.cpp](https://github.com/ggerganov/llama.cpp), Apple M1/M2 Tabby users can now harness Metal inference support on Apple's M1 and M2 chips by using the `--device metal` flag.
This enhancement leads to a significant inference speed upgrade🚀. It marks a meaningful milestone in Tabby's adoption on Apple devices. Check out our [Model Directory](/docs/models) to discover LLM models with Metal support! 🎁
<center>
![Inference](./inference.png)
*An example inference benchmarking with [CodeLlama-7B](https://huggingface.co/TabbyML/CodeLlama-7B) on Apple M2 Max, takes ~600ms.*
</center>
:::tip
Check out latest Tabby updates on [Linkedin](https://www.linkedin.com/company/tabbyml/) and [Slack community](https://join.slack.com/t/tabbycommunity/shared_invite/zt-1xeiddizp-bciR2RtFTaJ37RBxr8VxpA)! Our Tabby community is eager for your participation. ❤️
:::

3
website/blog/2023-09-13-release-0-1-0-metal/inference.png
View File

@ -1,3 +0,0 @@
version https://git-lfs.github.com/spec/v1
oid sha256:702e9b69b54a0b86731c23d199ffe454a2f03437b25f0fe8c25257e9c71b8877
size 19495

30
website/blog/2023-09-18-release-0-1-1-metal/index.md Normal file
View File

@ -0,0 +1,30 @@
---
authors: [ meng ]
---
# Highlights of Tabby v0.1.1: Apple M1/M2 Support
We are thrilled to announce the release of Tabby [v0.1.1](https://github.com/TabbyML/tabby/releases/tag/v0.1.1) 👏🏻.
Apple M1/M2 Tabby users can now harness Metal inference support on Apple's M1 and M2 chips by using the `--device metal` flag, thanks to [llama.cpp](https://github.com/ggerganov/llama.cpp)'s awesome metal support.
The Tabby team made a [contribution](https://github.com/ggerganov/llama.cpp/pull/3187) by adding support for the StarCoder series models (1B/3B/7B) in llama.cpp, enabling more appropriate model usage on the edge for completion use cases.
<center>
```
llama_print_timings: load time = 105.15 ms
llama_print_timings: sample time = 0.00 ms / 1 runs ( 0.00 ms per token, inf tokens per second)
llama_print_timings: prompt eval time = 25.07 ms / 6 tokens ( 4.18 ms per token, 239.36 tokens per second)
llama_print_timings: eval time = 311.80 ms / 28 runs ( 11.14 ms per token, 89.80 tokens per second)
llama_print_timings: total time = 340.25 ms
```
*Inference benchmarking with [StarCoder-1B](https://huggingface.co/TabbyML/StarCoder-1B) on Apple M2 Max now takes approximately 340ms, compared to the previous time of around 1790ms. This represents a roughly 5x speed improvement.*
</center>
This enhancement leads to a significant inference speed upgrade🚀, for example, It marks a meaningful milestone in Tabby's adoption on Apple devices. Check out our [Model Directory](/docs/models) to discover LLM models with Metal support! 🎁
:::tip
Check out latest Tabby updates on [Linkedin](https://www.linkedin.com/company/tabbyml/) and [Slack community](https://join.slack.com/t/tabbycommunity/shared_invite/zt-1xeiddizp-bciR2RtFTaJ37RBxr8VxpA)! Our Tabby community is eager for your participation. ❤️
:::