RE: LeoThread 2025-04-22 17:41

avatar

You are viewing a single comment's thread:

It is not bad for something new and using an older model. Have to keep putting more in to train it.

leofinance
0
0
0.000
4 comments
avatar

Is @askrafiki based on Llama or DeepSeek?

0
0
0.000
Reply
avatar

My understanding it is Llama. It looks like it is about 2 years old on the data used for training.

0
0
0.000
Reply
avatar

Thank you for the information. DeepSeek came much later to the scene. Llama 3.X is the most likely case.

0
0
0.000
Reply
avatar

Response to @vimukthi's Question

askrafiki is based on a model that is being trained to improve its performance, but the specific base model (Llama or DeepSeek) is not explicitly mentioned in the context provided.

However, considering the conversation about training and improving askrafiki, it can be inferred that the model is likely a custom implementation or a fine-tuned version of an existing language model, possibly with elements from both Llama and DeepSeek, or another model altogether.

For more accurate information, it would be best to consult the developers or the official documentation of askrafiki.

Note: The response is based on the provided context and may not reflect the actual details about askrafiki's underlying model.

0
0
0.000
Reply