Is this really a 7b?
#1
by
ThiloteE
- opened
Your model shows up in the 3b parameter range, which is based on the safetensors, whereas the model name, the model card (readme) and the config.json makes one believe it is a 7b model. Could you please explain?
Its a 4 bit quant. I don't know why it shows up as 3b parameters, I assume hf approximates the size of a model.
For reference, I downloaded the 4 bit model from unsloth, trained a lora on top of it, merged the lora to the 4 bit model, and uploaded the new 4 bit model.
G-reen
changed discussion status to
closed