Gemma 2 is now available to researchers and developers (blog.google)
from simple@lemm.ee to technology@lemmy.world on 27 Jun 17:21
https://lemm.ee/post/35667035

Context: Gemma is a group of free-to-use AI models with a focus on being small. According to benchmarks this outperforms Llama 3.

#technology

threaded - newest

passepartout@feddit.org on 27 Jun 18:02 next collapse

It’s already on Ollama. Exciting times!

ExtravagantEnzyme@lemm.ee on 27 Jun 18:53 next collapse

Just to be clear, Gemma is only partially open sourced in select area’s of the code.

<img alt="" src="https://lemm.ee/pictrs/image/35b4c57f-41b8-46db-af97-dc55515a3f35.png">

eager_eagle@lemmy.world on 28 Jun 00:41 collapse

how are the weights partially open?

ExtravagantEnzyme@lemm.ee on 28 Jun 06:29 collapse

Only portions of the code are published while the rest is kept under wraps. Classic corporate America bs finding a loop hole to use a trendy term.

eager_eagle@lemmy.world on 28 Jun 15:40 collapse

neural network weights are just files, collections of numbers forming matrices; how is a partially open collection of weights of any use

the weights are open

$ docker exec -it ollama ollama show gemma:7b
  Model                              
  	arch            	gemma	             
  	parameters      	9B   	             
  	quantization    	Q4_0 	             
  	context length  	8192 	             
  	embedding length	3072 	             
  	                                  
  Parameters                         
  	stop            	"<start_of_turn>"	 
  	stop            	"<end_of_turn>"  	 
  	penalize_newline	false            	 
  	repeat_penalty  	1                	 
  	                                  
  License                            
  	Gemma Terms of Use              	  
  	Last modified: February 21, 2024
jacksilver@lemmy.world on 28 Jun 21:10 collapse

Since there is a user acceptance policy that restricts what you can do with the model that might be considered “partially” open.

Yeah you can see the weights, but it seems you are limited on what you can do with the weights. How we’ve gotten to the point you can protect these random numbers that I’ve shared with you through a UA is beyond me.

eager_eagle@lemmy.world on 28 Jun 21:59 collapse

the same happens with BloomZ, and that is listed as open

jacksilver@lemmy.world on 28 Jun 23:05 collapse

Wasn’t aware of that, I was just taking a guess.

That being said I wouldn’t consider either open given those restrictions.

BaroqueInMind@lemmy.one on 27 Jun 19:51 collapse

I’m looking at the HuggingFace leader boards and Gemma isn’t even top 10. How does it stack up against Llama3 or Mistral-Open-Orca?

simple@lemm.ee on 27 Jun 21:04 collapse

It’s currently number 11 in the LMSys chatbot leaderboard. It’s sitting above Llama 3 and Claude 3 Sonnet.