Hugging Face's logo

pszemraj
/
jamba-900M-v0.13-KIx2

Text Generation
Transformers
Safetensors
English
jamba
smol MoE
smol
Model card Files Community
pszemraj commited on
Commit
ffb78a7
•
1 Parent(s): c2efc63

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -34,7 +34,7 @@ Details:
34
  - uses Claude3 tokenizer (as hf GPT2 tokenizer)
35
  - hidden size 1024, 12 layers, 8 experts
36
 
37
- achieves the following results on the evaluation set (_ of the latest dataset_):
38
  - Loss: 3.0366
39
  - Accuracy: 0.4514
40
  - Num Input Tokens Seen: 1975517184
 
34
  - uses Claude3 tokenizer (as hf GPT2 tokenizer)
35
  - hidden size 1024, 12 layers, 8 experts
36
 
37
+ achieves the following results on the evaluation set (_most recent dataset_):
38
  - Loss: 3.0366
39
  - Accuracy: 0.4514
40
  - Num Input Tokens Seen: 1975517184