Hugging Face's logo

MetaIX
/
OpenAssistant-Llama-30b-4bit

Text Generation
Transformers
PyTorch
llama
text-generation-inference
Inference Endpoints
Model card Files Community
5
MetaIX commited on
Commit
208c3a0
•
1 Parent(s): 808ba96

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -3
README.md CHANGED
@@ -30,10 +30,10 @@ This was made using <a href="https://huggingface.co/OpenAssistant/oasst-sft-7-ll
30
 
31
  <p><strong><font size="4">--true-sequential --groupsize 128</font></strong></p>
32
 
33
- <strong>Wikitext2</strong>: 4.711131572723389
34
 
35
- <strong>Ptb-New</strong>: 9.385306358337402
36
 
37
- <strong>C4-New</strong>: 6.990415573120117
38
 
39
  <strong>Note</strong>: This version uses <i>--groupsize 128</i>, resulting in better evaluations. However, it consumes more VRAM.
 
30
 
31
  <p><strong><font size="4">--true-sequential --groupsize 128</font></strong></p>
32
 
33
+ <strong>Wikitext2</strong>: 4.641914367675781
34
 
35
+ <strong>Ptb-New</strong>: 9.117929458618164
36
 
37
+ <strong>C4-New</strong>: 6.867942810058594
38
 
39
  <strong>Note</strong>: This version uses <i>--groupsize 128</i>, resulting in better evaluations. However, it consumes more VRAM.