We compared the high-performance open source LLM "Llama-3-Elyza-JP-8B" and GPT-3.5 on ChatStream.

  Рет қаралды 17

株式会社 Qualiteg

株式会社 Qualiteg

Ай бұрын

Hello, Today we tried Llama-3-Elyza-JP-8B.
Llama-3-Elyza-JP-8B, announced yesterday, June 26, 2024 is said to be a Japanese LLM that outperforms "GPT-4" in the 70B version. Llama-3-Elyza-JP-8B is said to be a Japanese LLM that outperforms GPT-4 in the 70B version.
We have installed Llama-3-Elyza-JP-8B in our Playground environment and tried it out.
The 70B (70 billion parameters) version is said to outperform GPT-4, so we decided to compare the 8B (8 billion parameters) version with GPT-3.5.

Пікірлер
The moment we stopped understanding AI [AlexNet]
17:38
Welch Labs
Рет қаралды 789 М.
Iron Chin ✅ Isaih made this look too easy
00:13
Power Slap
Рет қаралды 35 МЛН
A little girl was shy at her first ballet lesson #shorts
00:35
Fabiosa Animated
Рет қаралды 10 МЛН
50 YouTubers Fight For $1,000,000
41:27
MrBeast
Рет қаралды 199 МЛН
Mama vs Son vs Daddy 😭🤣
00:13
DADDYSON SHOW
Рет қаралды 35 МЛН
This NEW Omega-1 Rotary Engine Will Be The End of EVs!
10:25
UltiumTech
Рет қаралды 16 М.
Generative AI in a Nutshell - how to survive and thrive in the age of AI
17:57
SpaceX Finally Finishes Testing! Starship Flight 5 Next!
19:21
What about it!?
Рет қаралды 282 М.
You need to learn AI in 2024! (And here is your roadmap)
45:21
David Bombal
Рет қаралды 677 М.
Iron Chin ✅ Isaih made this look too easy
00:13
Power Slap
Рет қаралды 35 МЛН