Shoddy_Vegetable_115

joined 10 months ago
sorted by: new top controversial old
1
Is RAG better with fine tuning on same data or pure RAG FTW? (alien.top)
 

So I have collected a very high quality and large medical QA dataset that I want to use to create a medical knowledge retrieval app. I have heard LLMs perform much better when it is fine tuned on the same data on which RAG is performed. Is it true? And is it worth the hassle of fine-tuning or am I good with pure RAG?

1
For Intel neuralchat or openhermes mistral 7b, what is the max output token size? (alien.top)
 

In fact, what is the total output token size for all model nowadays? Because I seem to have confused context length and output length. I figured that output length is usually fixed to 4k tokens. Is there such a limit to all models even the ones I mentioned? Because their model cards do not have any info on token limit.

Intel's neural-chat-7b-v3-1 dropped in c/localllama@poweruser.forum
[–] Shoddy_Vegetable_115@alien.top 1 points 9 months ago

What's the correct prompt template for this specific model?

NeuralChat 7B: Intel’s Chat Model Trained with DPO in c/localllama@poweruser.forum
[–] Shoddy_Vegetable_115@alien.top 1 points 10 months ago (2 children)

Exactly. It didn't hallucinate even once in my tests. I used RAG and it gave me perfect to-the-point answers. But I know most people want more verbose outputs it's just that it's good for factual retrieval use cases.

Are 7b models useful? in c/localllama@poweruser.forum
[–] Shoddy_Vegetable_115@alien.top 1 points 10 months ago

It returns a JSON with function name and respective arguments which you can parse later in the program and call the function with those arguments given by the model.

1
Creating a lab report analyzer with llama2 or mistral (alien.top)
 

Hi everyone. I am trying to create a lab report analyzer which takes in lab report data of a patient(blood work, cholesterol, sugar level, etc) and gives a detailed analysis of it.

To do this, I want to use llama2 or mistral as the base model. I'm looking for datasets right now. So I have 2 questions: 1)Which base model will be the best for this considering this will require some logical analysis on the model's part. 2)when I get the dataset, should I use RAG or do a fine tune? Which is likely to give me better results? 3)😅well... do you guys know of any such datasets? It would help a ton.

Thanks!

1
Best small model for function calling and decision making task? (alien.top)
 

I want a small model(ideally <=3B params) which is really effecient and good at function calling and logical decision making(after fine tuning). With decision making, I mean for example if it gets a JSON input telling the distance of an approaching object, it should intelligently call a function to steer the motor to a certain angle to avoid collision or something like that. Any such models?

Which model is best for binary text classification? in c/localllama@poweruser.forum
[–] Shoddy_Vegetable_115@alien.top 1 points 10 months ago

My classification task is to classify a given essay into AI generated and human generated. And I need the answer to be between 0 and 1(both included) with 1 being AI generated and 0 being human generated.

Few-shot examples is a good idea for most classification tasks but I don't think Generative LLMs can understand the more intricate semantic patterns to differentiate between the AI and human generated with just a few examples but I'll try it once and let you know!

Btw do you think fine-tuning would be better?

1
Which model is best for binary text classification? (alien.top)
 

Hi. So I am a bit new to NLP and ML as a whole and I am looking to create a text classification model. I have tried it with deBERTa and the results are decent(about 70%) but I need more accuracy. Are Generstive models a better alternative or should I stick to smaller models like Bert or maybe even non-NN classifiers and work on better dataset quality?