# Llama2 model parameters count is half

**URL:** <https://discuss.huggingface.co/t/llama2-model-parameters-count-is-half/59290>\
**Category:** Models\
**Created:** [October 20, 2023, 3:12pm UTC](https://discuss.huggingface.co/t/llama2-model-parameters-count-is-half/59290 "2023-10-20T15:12:57Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![sisilmehta](https://avatars.discourse-cdn.com/v4/letter/s/a4c791/32.png) [@sisilmehta](https://discuss.huggingface.co/u/sisilmehta)\
**Post date:** [October 20, 2023, 3:12pm UTC](https://discuss.huggingface.co/t/llama2-model-parameters-count-is-half/59290/1 "2023-10-20T15:12:58Z")

</div>

I just loaded the meta-llama/llama2-13b-hf model and did a counted the parameters using the numel() fn. It looks like the 13b model has ~6.7b params. I’m see a similar discrepancy with the meta-llama/llama2-7b-hf model as well.

However the model sizes were fine a week ago.

Did the models change?

---

<div class="post-metadata">

**Author:** ![bbergner](https://avatars.discourse-cdn.com/v4/letter/b/5daacb/32.png) [@bbergner](https://discuss.huggingface.co/u/bbergner)\
**Post date:** [November 29, 2023, 10:28am UTC](https://discuss.huggingface.co/t/llama2-model-parameters-count-is-half/59290/2 "2023-11-29T10:28:57Z")

</div>

I noticed the same in 4 bit quantized models (BitsAndBytes). It shows the correct number in the “original” models.