Filip commited on
Commit
6bd03c4
·
1 Parent(s): accf921
Files changed (1) hide show
  1. README.md +1 -2
README.md CHANGED
@@ -37,8 +37,7 @@ Both models used the same hyperparameters during training.\
37
  `weight_decay=0.01`\
38
  `lr_scheduler_type="linear"`
39
 
40
- We chose float16 as the quantization method as it has the fastest conversion and retains 100% accuracy. However, it is slow and memory hungry which is a disadvantage.
41
- Source: https://github.com/unslothai/unsloth/wiki
42
 
43
  ## Judge
44
 
 
37
  `weight_decay=0.01`\
38
  `lr_scheduler_type="linear"`
39
 
40
+ We chose float16 as the quantization method as it according to [Unsloth wiki](https://github.com/unslothai/unsloth/wiki) has the fastest conversion and retains 100% accuracy. However, it is slow and memory hungry which is a disadvantage.
 
41
 
42
  ## Judge
43