Fix rope theta for OpenLlama

[jla524]

What does this PR do?

Fixes #29506 (issue)

Rope theta has been added in the config with the same default value as Llama.

Results:

% python3
Python 3.12.2 (main, Feb  6 2024, 20:19:44) [Clang 15.0.0 (clang-1500.1.0.2.5)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> from transformers import OpenLlamaForCausalLM
>>> model = OpenLlamaForCausalLM.from_pretrained("openlm-research/open_llama_7b")
You are using a model of type llama to instantiate a model of type open-llama. This is not supported for all configurations of models and can yield errors.
Loading checkpoint shards: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 2/2 [00:22<00:00, 11.41s/it]
Some weights of OpenLlamaForCausalLM were not initialized from the model checkpoint at openlm-research/open_llama_7b and are newly initialized: ['model.embed_layer_norm.bias', 'model.embed_layer_norm.weight']
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.

Who can review?

@ArthurZucker and @kevin-guimard-ext

[ArthurZucker]

Thanks!

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.