llama-7b
This contains the weights for the LLaMA-7b model.
text → text · made by huggyllama
Sold by
Nobody in the catalogue publishes a price for this yet.
Measured 12 standings
| Place | Board | Metric | Score |
|---|---|---|---|
| 26thof 26 | Science qa — Epoch AI | Score | 36.19 |
| 30thof 53 | Lambada — Epoch AI | Score | 73.3 |
| 32ndof 70 | Open book qa — Epoch AI | Accuracy | 57.2 |
| 66thof 115 | Trivia qa — Epoch AI | EM | 0.71 |
| 73rdof 113 | Hella swag — Epoch AI | Overall accuracy | 0.762 |
| 74thof 113 | Piqa — Epoch AI | Score | 79.8 |
| 77thof 88 | Bbh — Epoch AI | Average | 0.335 |
| 83rdof 136 | Bool q — Epoch AI | Score | 76.5 |
| 89thof 134 | Arc ai2 — Epoch AI | Challenge score | 0.476 |
| 89thof 137 | Wino grande — Epoch AI | Accuracy | 70.1 |
| 142ndof 170 | Gsm8k — Epoch AI | EM | 0.11 |
| 183rdof 217 | Mmlu — Epoch AI | EM | 0.356 |
About
llama-7b — a text model from huggyllama, placed 32nd of 70 on Open book qa — Epoch AI.
It takes text and returns text. It was published in April 2023. The catalogue files it under chat. Its weights are published under the other licence, which attaches conditions the plain open licences do not. It has been measured on 12 boards, and stands best at 32nd of 70 on Open book qa — Epoch AI.
Every current figure
- Maker
- huggyllama
- Register
- model
- Takes
- text
- Returns
- text
- Published
- April 2023
- Parameters
- 6.7 billion · read from its own weights
- Licence
- other
- Boards
- 12
- Best place
- 32nd of 70 — Open book qa — Epoch AI
Known as 1 name
huggyllama/llama-7b