Add LlavaImageProcessor (#33191)

* First draft

* Add equivalence test

* Update docstrings

* Add tests

* Use numpy

* Fix tests

* Improve variable names

* Improve docstring

* Add link

* Remove script

* Add copied from

* Address comment

* Add note in docs

* Add docstring, data format

* Improve test

* Add test

* update

* Update src/transformers/models/llava/image_processing_llava.py

Co-authored-by: Pavel Iakubovskii <qubvel@gmail.com>

* Update src/transformers/models/llava/image_processing_llava.py

Co-authored-by: Pavel Iakubovskii <qubvel@gmail.com>

* loop once only

---------

Co-authored-by: raushan <raushan@huggingface.co>
Co-authored-by: Raushan Turganbay <raushan.turganbay@alumni.nu.edu.kz>
Co-authored-by: Pavel Iakubovskii <qubvel@gmail.com>

docs/source/en/model_doc/llava.md

@@ -162,6 +162,16 @@ For multiple turns conversation:
 "USER: <image>\n<prompt1> ASSISTANT: <answer1></s>USER: <prompt2> ASSISTANT: <answer2></s>USER: <prompt3> ASSISTANT:"
 ```
 
+## Note regarding reproducing original implementation
+
+In order to match the logits of the [original implementation](https://github.com/haotian-liu/LLaVA/tree/main), one needs to additionally specify `do_pad=True` when instantiating `LLavaImageProcessor`:
+
+```python
+from transformers import LLavaImageProcessor
+
+image_processor = LLavaImageProcessor.from_pretrained("https://huggingface.co/llava-hf/llava-1.5-7b-hf", do_pad=True)
+```
+
 ### Using Flash Attention 2
 
 Flash Attention 2 is an even faster, optimized version of the previous optimization, please refer to the [Flash Attention 2 section of performance docs](https://huggingface.co/docs/transformers/perf_infer_gpu_one).
@@ -180,6 +190,11 @@ A list of official Hugging Face and community (indicated by 🌎) resources to h
 
 [[autodoc]] LlavaConfig
 
+## LlavaImageProcessor
+
+[[autodoc]] LlavaImageProcessor
+    - preprocess
+
 ## LlavaProcessor
 
 [[autodoc]] LlavaProcessor