Model Details
This model is a 3B llama3 model pretrained from scratch with torchtitan on fineweb-edu with AdamW optimizer. 20x chinchilla rule for 60B tokens seen.
How to use
import torch
from transformers import pipeline
pipe = pipeline(
"text-generation",
model="kz919/llama3_3b_chinchilla_8142025",
)
print(pipe("The key to life is"))
Downstream Eval
ARC, Hellaswag, Lambda_OpenAI, OpenbookQA, PIQA
lm_eval --model hf --model_args pretrained=kz919/llama3_3b_chinchilla_8142025,dtype="bfloat16",add_bos_token=True --tasks lambada_openai,hellaswag,piqa,arc_easy,arc_challenge,openbookqa --device cuda:7 --batch_size 8
| Tasks |
Version |
Filter |
n-shot |
Metric |
|
Value |
|
Stderr |
| arc_challenge |
1 |
none |
0 |
acc |
↑ |
0.3003 |
± |
0.0134 |
|
|
none |
0 |
acc_norm |
↑ |
0.3106 |
± |
0.0135 |
| arc_easy |
1 |
none |
0 |
acc |
↑ |
0.6246 |
± |
0.0099 |
|
|
none |
0 |
acc_norm |
↑ |
0.5379 |
± |
0.0102 |
| hellaswag |
1 |
none |
0 |
acc |
↑ |
0.3735 |
± |
0.0048 |
|
|
none |
0 |
acc_norm |
↑ |
0.4614 |
± |
0.0050 |
| lambada_openai |
1 |
none |
0 |
acc |
↑ |
0.3685 |
± |
0.0067 |
|
|
none |
0 |
perplexity |
↓ |
32.9840 |
± |
1.3564 |
| openbookqa |
1 |
none |
0 |
acc |
↑ |
0.2560 |
± |
0.0195 |
|
|
none |
0 |
acc_norm |
↑ |
0.3460 |
± |
0.0213 |
| piqa |
1 |
none |
0 |
acc |
↑ |
0.6703 |
± |
0.0110 |
|
|
none |
0 |
acc_norm |
↑ |
0.6844 |
± |
0.0108 |
MMLU
| Groups |
Version |
Filter |
n-shot |
Metric |
|
Value |
|
Stderr |
| mmlu |
2 |
none |
|
acc |
↑ |
0.2421 |
± |
0.0036 |
| - humanities |
2 |
none |
|
acc |
↑ |
0.2497 |
± |
0.0063 |
| - other |
2 |
none |
|
acc |
↑ |
0.2568 |
± |
0.0078 |
| - social sciences |
2 |
none |
|
acc |
↑ |
0.2265 |
± |
0.0075 |
| - stem |
2 |
none |
|
acc |
↑ |
0.2315 |
± |
0.0075 |
