TensorRT-LLMs/examples/multimodal/README.md

# Multi-Modal

This document shows how to run multimodal pipelines with TensorRT-LLM, e.g. from image+text input modalities to text output.

Multimodal models' LLM part has an additional parameter `--max_multimodal_len` compared to LLM-only build commands. Under the hood, `max_multimodal_len` and `max_prompt_embedding_table_size` are effectively the same concept, i.e., prepended/concatenated embeddings (either multimodal feature embeddings or prompt tuning embeddings) to the LLM input embeddings. The multimodal features from the visual encoder of shape `[batch_size, num_visual_features, visual_hidden_dim]` is flattened as `[batch_size * num_visual_features, visual_hidden_dim]` and passed like a prompt embedding table.

## BLIP2-T5

1. Download Huggingface weights and convert original checkpoint to TRT-LLM checkpoint format
   following example in `examples/enc_dec/README.md`.

    ```bash
    export MODEL_NAME=flan-t5-xl
    git clone https://huggingface.co/google/${MODEL_NAME} tmp/hf_models/${MODEL_NAME}

    python ../enc_dec/t5/convert.py -i tmp/hf_models/${MODEL_NAME} \
        -o tmp/trt_models/${MODEL_NAME} --weight_data_type float32 \
        --inference_tensor_para_size 1
    ```

2. Build TRT-LLM engine from TRT-LLM checkpoint

    ```bash
    python ../enc_dec/build.py --model_type t5 \
        --weight_dir tmp/trt_models/${MODEL_NAME}/tp1 \
        --output_dir trt_engines/${MODEL_NAME}/1-gpu \
        --engine_name ${MODEL_NAME} \
        --remove_input_padding \
        --use_bert_attention_plugin \
        --use_gpt_attention_plugin \
        --use_gemm_plugin \
        --dtype bfloat16 \
        --max_beam_width 1 \
        --max_batch_size 8 \
        --max_multimodal_len 256 \ # 8 (max_batch_size) * 32 (num_visual_features)
        --max_encoder_input_len 924 \ # change if LLM text input range is known
        --max_output_len 100
    ```

    **NOTE**: `max_multimodal_len = max_batch_size * num_visual_features`, so if you change max_batch_size, max multimodal length **MUST** be changed accordingly.

    The built T5 engines are located in `./trt_engines/${MODEL_NAME}/1-gpu/bfloat16/tp1`.

3.  Build TensorRT engines for visual components

    ```bash
    python build_visual_engine.py --model_name ${MODEL_NAME} --model_path tmp/hf_models/${MODEL_NAME}
    ```

    The built engines are located in `./visual_engines/${MODEL_NAME}`.

4. Assemble everything into BLIP2 pipeline

    ```bash
    python run.py \
        --blip_encoder \
        --max_new_tokens 30 \
        --input_text "Question: which city is this? Answer:" \
        --hf_model_dir tmp/hf_models/${MODEL_NAME} \
        --visual_engine_dir visual_engines/${MODEL_NAME} \
        --llm_engine_dir trt_engines/${MODEL_NAME}/1-gpu/bfloat16/tp1
    ```

## BLIP2-OPT

OPT pipeline needs few minor changes from T5 pipeline

1. Convert Huggingface weights to TRT-LLM checkpoint format following `examples/opt/README.md`.

2. Use `trtllm-build` command to build TRT-LLM engine for OPT.

3. Add `--decoder-llm` argument to inference script, since OPT is a decoder-only LLM.

4. The full list of commands is as follows:

    ```bash
    export MODEL_NAME=opt-2.7b
    git clone https://huggingface.co/facebook/${MODEL_NAME} tmp/hf_models/${MODEL_NAME}

    python ../opt/convert_checkpoint.py \
        --model_dir tmp/hf_models/${MODEL_NAME} \
        --dtype float16 \
        --output_dir tmp/trt_models/${MODEL_NAME}/fp16/1-gpu

    trtllm-build \
        --checkpoint_dir tmp/trt_models/${MODEL_NAME}/fp16/1-gpu \
        --output_dir trt_engines/${MODEL_NAME}/fp16/1-gpu \
        --gpt_attention_plugin float16 \
        --gemm_plugin float16 \
        --max_beam_width 1 \
        --max_batch_size 8 \
        --max_multimodal_len 256 \
        --max_input_len 924 \
        --max_output_len 100

    python build_visual_engine.py --model_name ${MODEL_NAME} --model_path tmp/hf_models/${MODEL_NAME}

    python run.py \
        --blip_encoder \
        --max_new_tokens 30 \
        --input_text "Question: which city is this? Answer:" \
        --hf_model_dir tmp/hf_models/${MODEL_NAME} \
        --visual_engine_dir visual_engines/${MODEL_NAME} \
        --llm_engine_dir trt_engines/${MODEL_NAME}/fp16/1-gpu \
        --decoder_llm
    ```

5. INT8/INT4 weight-only quantization for OPT can be enabled using commands as follows (take `INT4` as an example, while `INT8` is the default precision for weight-only quantization):
    ```bash
    python ../opt/convert_checkpoint.py \
        --model_dir tmp/hf_models/${MODEL_NAME} \
        --dtype float16 \
        --output_dir tmp/trt_models/${MODEL_NAME}/int4_weightonly/1-gpu \
        --use_weight_only \
        --weight_only_precision int4

    trtllm-build \
        --checkpoint_dir tmp/trt_models/${MODEL_NAME}/int4_weightonly/1-gpu \
        --output_dir trt_engines/${MODEL_NAME}/int4_weightonly/1-gpu \
        --gpt_attention_plugin float16 \
        --gemm_plugin float16 \
        --max_beam_width 1 \
        --max_batch_size 8 \
        --max_multimodal_len 256 \
        --max_input_len 924 \
        --max_output_len 100
    ```

    The built OPT engines lie in `trt_engines/${MODEL_NAME}/int4_weightonly/1-gpu`.
    You should use this directory as `--llm_engine_dir` argument to `run.py`

    **NOTE:** INT8/INT4 option is not supported for BLIP2-T5, because quantization support has not be
          added for encoder-decoder models yet.

## LLaVA

1. Download Huggingface model weights. This model has both LLM and visual components
   unlike BLIP2 example which downloads only LLM components from Huggingface.

    ```bash
    export MODEL_NAME="llava-1.5-7b-hf"
    git clone https://huggingface.co/llava-hf/${MODEL_NAME} tmp/hf_models/${MODEL_NAME}
    ```

2. Generate TRT-LLM engine for LLaMA following example in `examples/llama/README.md`

    ```bash
    python ../llama/build.py \
        --model_dir tmp/hf_models/${MODEL_NAME} \
        --output_dir trt_engines/${MODEL_NAME}/fp16/1-gpu \
        --dtype float16 \
        --gpt_attention_plugin float16 \
        --gemm_plugin float16 \
        --max_batch_size 1 \
        --max_multimodal_len 576 \ # 1 (max_batch_size) * 576 (num_visual_features)
        --max_input_len 2048 \
        --max_output_len 512
    ```

3.  Build TensorRT engines for visual components

    ```bash
    python build_visual_engine.py --model_name ${MODEL_NAME} --model_path tmp/hf_models/${MODEL_NAME}
    ```

4. Add `--decoder-llm` argument to inference script, since LLaMA is a decoder-only LLM.

    ```bash
    python run.py \
        --max_new_tokens 30 \
        --input_text "Question: which city is this? Answer:" \
        --hf_model_dir tmp/hf_models/${MODEL_NAME} \
        --visual_engine_dir visual_engines/${MODEL_NAME} \
        --llm_engine_dir trt_engines/${MODEL_NAME}/fp16/1-gpu \
        --decoder_llm
    ```

## Nougat

1. Download Huggingface weights

    ```bash
    export MODEL_NAME="nougat-base" # or nougat-small
    git clone https://huggingface.co/facebook/${MODEL_NAME} tmp/hf_models/${MODEL_NAME}
    ```

2. Convert Huggingface weights into TRT-LLM checkpoints and build TRT engines using scripts in `examples/enc_dec`

   Nougat uses mBART architecture but replaces the LLM encoder with a Swin Transformer encoder.
   To achieve this, we add an extra `--nougat` flag (over mBART example) to
   `bart/convert.py` and `build.py` in `examples/enc_dec`.

    ```bash
    python ../enc_dec/bart/convert.py -i tmp/hf_models/${MODEL_NAME} \
        -o tmp/trt_models/${MODEL_NAME} --weight_data_type float32 \
        --inference_tensor_para_size 1 --nougat

    python ../enc_dec/build.py \
        --model_type bart \
        --weight_dir tmp/trt_models/${MODEL_NAME}/tp1 \
        -o trt_engines/${MODEL_NAME}/1-gpu \
        --engine_name $MODEL_NAME \
        --bert_attention_plugin \
        --gpt_attention_plugin \
        --use_gemm_plugin \
        --dtype bfloat16 \
        --max_beam_width 1 \
        --max_batch_size 1 \
        --nougat \
        --max_multimodal_len 588 \ # 1 (max_batch_size) * 588 (num_visual_features)
        --max_output_len 100
    ```

3. Generate TensorRT engines for visual components and combine everything into final pipeline.

    ```bash
    python build_visual_engine.py --model_name ${MODEL_NAME} --model_path tmp/hf_models/${MODEL_NAME}

    python run.py \
        --hf_model_dir tmp/hf_models/${MODEL_NAME} \
        --visual_engine_dir visual_engines/${MODEL_NAME} \
        --llm_engine_dir trt_engines/${MODEL_NAME}/1-gpu/bfloat16/tp1 \
        --nougat
    ```