yixinsong commited on
Commit
3ab53ca
1 Parent(s): 6ee8174

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -0
README.md CHANGED
@@ -27,6 +27,11 @@ We take ChatML as our chat template:
27
 
28
  As we merged the predictors for FFN neurons in models, you can finetune TurboSparse-Mixtral with any framework and algorithm.
29
 
 
 
 
 
 
30
  ## License
31
 
32
  The model is licensed under Apache-2.0, while model weights are fully open for academic research and also allow **free** commercial usage.
 
27
 
28
  As we merged the predictors for FFN neurons in models, you can finetune TurboSparse-Mixtral with any framework and algorithm.
29
 
30
+ ## Limitations
31
+ * TurboSparse, having just undergone training with 150B tokens, may still exhibit performance gaps in certain tasks.
32
+ * The TurboSparse model has only been trained on English-language datasets, hence its capabilities in other languages are still lacking.
33
+ * The model may produce unexpected outputs due to its small size and probabilistic generation paradigm.
34
+
35
  ## License
36
 
37
  The model is licensed under Apache-2.0, while model weights are fully open for academic research and also allow **free** commercial usage.