github
/
ColossalAI
mirror of https://github.com/hpcaitech/ColossalAI
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

[doc] Update llama + sp compatibility; fix dist optim table

Browse Source 
Co-authored-by: Edenzzzz <wtan45@wisc.edu>
pull/5878/head
Edenzzzz 5 months ago committed by GitHub
parent 8ab46b4000
commit 936d0b0f7b
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 37 additions and 40 deletions
Show Stats Download Patch File Download Diff File

36
docs/source/en/features/distributed_optimizers.md
Unescape View File

@ -87,44 +87,42 @@ optim = DistGaloreAwamW(
## Plugin compatibility
<table>
<tr>
<th nowrap="nowrap">Model/Feature</th>
<th nowrap="nowrap" align="center" title="Lamb">Lamb</th>
<th nowrap="nowrap" align="center" title="GaLore">GaLore</th>
<th nowrap="nowrap" align="center" title="Adafactor">Adafactor</th>
<th nowrap="nowrap" align="center" title="CAME">CAME</th>
<th nowrap="nowrap">Optimizer/Plugin</th>
<th nowrap="nowrap" align="center">Hybrid Parallel Plugin</th>
<th nowrap="nowrap" align="center">Low Level Zero Plugin</th>
<th nowrap="nowrap" align="center">Torch DDP Plugin</th>
<th nowrap="nowrap" align="center">Gemini Plugin</th>
<th nowrap="nowrap" align="center">Moe Hybrid Plugin</th>
</tr>
<tr>
<td nowrap="nowrap">Hybrid Parallel<br />Plugin</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center" title="Lamb">Lamb</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
</tr>
<tr>
<td nowrap="nowrap">Low Level Zero<br />Plugin</td>
<td nowrap="nowrap" align="center" title="GaLore">GaLore</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
</tr>
<tr>
<td nowrap="nowrap">Torch DDP<br />Plugin</td>
<td nowrap="nowrap" align="center" title="Adafactor">Adafactor</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
</tr>
<tr>
<td nowrap="nowrap">Gemini<br />Plugin</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
</tr>
<tr>
<td nowrap="nowrap">Moe Hybrid<br />Plugin</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center" title="CAME">CAME</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
</tr>

2
docs/source/en/features/shardformer.md
Unescape View File

@ -55,7 +55,7 @@ Model/Feature Compatibility Matrix:
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">❌</td>
</tr>
<tr>

37
docs/source/zh-Hans/features/distributed_optimizers.md
Unescape View File

@ -84,44 +84,42 @@ optim = DistGaloreAwamW(
## 兼容性
<table>
<tr>
<th nowrap="nowrap">Model/Feature</th>
<th nowrap="nowrap" align="center" title="Lamb">Lamb</th>
<th nowrap="nowrap" align="center" title="GaLore">GaLore</th>
<th nowrap="nowrap" align="center" title="Adafactor">Adafactor</th>
<th nowrap="nowrap" align="center" title="CAME">CAME</th>
<th nowrap="nowrap">Optimizer/Plugin</th>
<th nowrap="nowrap" align="center">Hybrid Parallel Plugin</th>
<th nowrap="nowrap" align="center">Low Level Zero Plugin</th>
<th nowrap="nowrap" align="center">Torch DDP Plugin</th>
<th nowrap="nowrap" align="center">Gemini Plugin</th>
<th nowrap="nowrap" align="center">Moe Hybrid Plugin</th>
</tr>
<tr>
<td nowrap="nowrap">Hybrid Parallel<br />Plugin</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center" title="Lamb">Lamb</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
</tr>
<tr>
<td nowrap="nowrap">Low Level Zero<br />Plugin</td>
<td nowrap="nowrap" align="center" title="GaLore">GaLore</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
</tr>
<tr>
<td nowrap="nowrap">Torch DDP<br />Plugin</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center" title="Adafactor">Adafactor</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
</tr>
<tr>
<td nowrap="nowrap">Gemini<br />Plugin</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
</tr>
<tr>
<td nowrap="nowrap">Moe Hybrid<br />Plugin</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center" title="CAME">CAME</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">❌</td>
</tr>
@ -130,6 +128,7 @@ optim = DistGaloreAwamW(
</tr>
</table>
<!-- doc-test-command: colossalai run --nproc_per_node 4 distributed_optimizers.py -->
## API 参考

2
docs/source/zh-Hans/features/shardformer.md
Unescape View File

@ -51,7 +51,7 @@ Author: [Baizhou Zhang](https://github.com/Fridge003), [Bin Jia](https://github.
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">❌</td>
<td nowrap="nowrap" align="center">✔️</td>
<td nowrap="nowrap" align="center">❌</td>
</tr>
<tr>

Write Preview
Loading…
Cancel
Save