mirror of
https://github.com/huggingface/text-generation-inference.git
synced 2025-06-03 21:22:08 +00:00
Small update to docs (#2816)
This commit is contained in:
parent
b2fac5d947
commit
a2d878fa0f
Binary file not shown.
Before Width: | Height: | Size: 201 KiB After Width: | Height: | Size: 209 KiB |
@ -80,7 +80,7 @@ Raw results
|
||||
|
||||
| | | | | |
|
||||
|---|---|---|---|---|
|
||||
|2nd run ||**TGI v3**|**vLLM**|**Amount of req**|
|
||||
|2nd run ||**TGI v3** (time in s)|**vLLM** (s)|**Amount of req**|
|
||||
|**Llama 3.1 8b**|Small test - L4 - 8B|17.5|19.9|200|
|
||||
|**Llama 3.1 8b**|Long test* - L4 - 8B|53|57|10|
|
||||
|**Llama 3.1 8b**|Small test - 4xL4 - 8B|4.8|6|200|
|
||||
@ -88,7 +88,7 @@ Raw results
|
||||
|**Llama 3.1 70b**|Small test - 8XH100 - 70B|6.2|7.4|200|
|
||||
|**Llama 3.1 70b**|Long test - 8H100 - 70B|2|27.5|20|
|
||||
||||||
|
||||
|1st run ||TGI|vLLM|Amount of req|
|
||||
|1st run ||TGI (s)|vLLM (s)|Amount of req|
|
||||
|**Llama 3.1 8b**|Small test - L4|19.9|19.9|200|
|
||||
|**Llama 3.1 8b**|Long test (10) - L4|49.8|55|10|
|
||||
|**Llama 3.1 8b**|Small test - 4xL4|13|12.6|200|
|
||||
|
Loading…
Reference in New Issue
Block a user