Hugging Face's logo

gradientai
/
Llama-3-8B-Instruct-Gradient-4194k

Text Generation
Transformers
Safetensors
English
llama
meta
llama-3
conversational
text-generation-inference
Model card Files Community
6
michaelfeil commited on
Commit
d5aa9c0
·
verified ·
1 Parent(s): 9ac556b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -45,7 +45,7 @@ For training data, we generate long contexts by augmenting [SlimPajama](https://
45
  | Initialize From | LLaMA-3 8B| 65K | 262K | 524k | 1048k |
46
  | Sequence Length 2^N | 16 | 18 | 19 | 20 | 22 |
47
  | RoPE theta | 15.3 M | 207.1 M | 1.06B | 2.80B | 45.2B |
48
- | Batch Size | 1 | 1 | 16 | 16 | 2 |
49
  | Gradient Accumulation Steps | 32 | 16 | 1 | 1 | 2 |
50
  | Steps | 30 | 24 | 50 | 50 | 12 (stopped early) |
51
  | Total Tokens | 62914560 | 100663296 | 419430400 | 838860800 | 201326592 |
 
45
  | Initialize From | LLaMA-3 8B| 65K | 262K | 524k | 1048k |
46
  | Sequence Length 2^N | 16 | 18 | 19 | 20 | 22 |
47
  | RoPE theta | 15.3 M | 207.1 M | 1.06B | 2.80B | 45.2B |
48
+ | Batch Size | 1 | 1 | 16 | 8 | 2 |
49
  | Gradient Accumulation Steps | 32 | 16 | 1 | 1 | 2 |
50
  | Steps | 30 | 24 | 50 | 50 | 12 (stopped early) |
51
  | Total Tokens | 62914560 | 100663296 | 419430400 | 838860800 | 201326592 |