Text Generation
Transformers
Safetensors
llama
text-generation-inference
Inference Endpoints
adamo1139 commited on
Commit
64f1296
1 Parent(s): 2250872

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -8,7 +8,7 @@ datasets:
8
  ---
9
  ## Known Issues
10
 
11
- <b>There's something weird going on with tokenizer. EXL2 quant works fine in ooba but not in exui. BNB 4-bit quant works fine in ooba.</b>
12
 
13
 
14
  <img src="https://cdn-uploads.huggingface.co/production/uploads/630fdd96a119d49bc1e770d5/BZ1TunduCB0xjfeTCObgL.png" width="600" style="float:center" />
 
8
  ---
9
  ## Known Issues
10
 
11
+ <b>There's something weird going on with tokenizer. EXL2 quant works fine in ooba but not in exui. BNB 4-bit quant works fine in ooba. For best results, use ooba with BOS token being inserted, repp 1.05 and probably exllamav2_HF loader over exllamav2</b>
12
 
13
 
14
  <img src="https://cdn-uploads.huggingface.co/production/uploads/630fdd96a119d49bc1e770d5/BZ1TunduCB0xjfeTCObgL.png" width="600" style="float:center" />