TRT engine docs (#1396)

* updated docs

* use optimum/gpt2

Author: IlyasMoutawwakil

docs/source/onnxruntime/usage_guides/gpu.mdx

@@ -291,29 +291,33 @@ We recommend setting these two provider options when using the TensorRT executio
 ... )
 ```
 
-TensorRT builds its engine depending on specified input shapes. Unfortunately, in the [current ONNX Runtime implementation](https://github.com/microsoft/onnxruntime/blob/613920d6c5f53a8e5e647c5f1dcdecb0a8beef31/onnxruntime/core/providers/tensorrt/tensorrt_execution_provider.cc#L1677-L1688) (references: [1](https://github.com/microsoft/onnxruntime/issues/13559), [2](https://github.com/microsoft/onnxruntime/issues/13851)), the engine is rebuilt every time an input has a shape smaller than the previously smallest encountered shape, and conversely if the input has a shape larger than the previously largest encountered shape. For example, if a model takes `(batch_size, input_ids)` as inputs, and the model takes successively the inputs:
+TensorRT builds its engine depending on specified input shapes. One big issue is that building the engine can be time consuming, especially for large models. Therefore, as a workaround, one recommendation is to build the TensorRT engine with dynamic shapes. This allows to avoid rebuilding the engine for new small and large shapes, which is unwanted once the model is deployed for inference.
 
-1. `input.shape: (4, 5)  --> the engine is built (first input)`
-2. `input.shape: (4, 10) --> engine rebuilt (10 larger than 5)`
-3. `input.shape: (4, 7)  --> no rebuild (5 <= 7 <= 10)`
-4. `input.shape: (4, 12) --> engine rebuilt (10 <= 12)`
-5. `input.shape: (4, 3)  --> engine rebuilt (3 <= 5)`
+To do so we use the provider's options `trt_profile_min_shapes`, `trt_profile_max_shapes` and `trt_profile_opt_shapes` to specify the minimum, maximum and optimal shapes for the engine. For example, for GPT2, we can use the following shapes:
 
-One big issue is that building the engine can be time consuming, especially for large models. Therefore, as a workaround, one recommendation is to **first build the TensorRT engine with an input of small shape, and then with an input of large shape to have an engine valid for all shapes inbetween**. This allows to avoid rebuilding the engine for new small and large shapes, which is unwanted once the model is deployed for inference.
+```python
+provider_options = {
+    "trt_profile_min_shapes": "input_ids:1x1,attention_mask:1x1,position_ids:1x1",
+    "trt_profile_opt_shapes": "input_ids:1x1,attention_mask:1x1,position_ids:1x1",
+    "trt_profile_max_shapes": "input_ids:1x64,attention_mask:1x64,position_ids:1x64",
+}
+```
 
 Passing the engine cache path in the provider options, the engine can therefore be built once for all and used fully for inference thereafter.
 
 For example, for text generation, the engine can be built with:
 
 ```python
 >>> import os
->>> from transformers import AutoTokenizer
 >>> from optimum.onnxruntime import ORTModelForCausalLM
 
 >>> os.makedirs("tmp/trt_cache_gpt2_example", exist_ok=True)
 >>> provider_options = {
 ...     "trt_engine_cache_enable": True,
-...     "trt_engine_cache_path": "tmp/trt_cache_gpt2_example"
+...     "trt_engine_cache_path": "tmp/trt_cache_gpt2_example",
+...     "trt_profile_min_shapes": "input_ids:1x1,attention_mask:1x1,position_ids:1x1",
+...     "trt_profile_opt_shapes": "input_ids:1x1,attention_mask:1x1,position_ids:1x1",
+...     "trt_profile_max_shapes": "input_ids:1x64,attention_mask:1x64,position_ids:1x64",
 ... }
 
 >>> ort_model = ORTModelForCausalLM.from_pretrained(
@@ -322,17 +326,6 @@ For example, for text generation, the engine can be built with:
 ...     provider="TensorrtExecutionProvider",
 ...     provider_options=provider_options,
 ... )
->>> tokenizer = AutoTokenizer.from_pretrained("optimum/gpt2")
-
->>> print("Building engine for a short sequence...")  # doctest: +IGNORE_RESULT
->>> text = ["short"]
->>> encoded_input = tokenizer(text, return_tensors="pt").to("cuda")
->>> output = ort_model(**encoded_input)
-
->>> print("Building engine for a long sequence...")  # doctest: +IGNORE_RESULT
->>> text = [" a very long input just for demo purpose, this is very long" * 10]
->>> encoded_input = tokenizer(text, return_tensors="pt").to("cuda")
->>> output = ort_model(**encoded_input)
 ```
 
 The engine is stored as: