TensorRT-LLMs/examples/bloom/README.md

199 lines
8.1 KiB
Markdown

# BLOOM
This document shows how to build and run a BLOOM model in TensorRT-LLM on both single GPU, single node multi-GPU and multi-node multi-GPU.
## Overview
The TensorRT-LLM BLOOM implementation can be found in [tensorrt_llm/models/bloom/model.py](../../tensorrt_llm/models/bloom/model.py). The TensorRT-LLM BLOOM example code is located in [`examples/bloom`](./). There are three main files in that folder::
* [`build.py`](./build.py) to build the [TensorRT](https://developer.nvidia.com/tensorrt) engine(s) needed to run the BLOOM model,
* [`run.py`](./run.py) to run the inference on an input text,
* [`summarize.py`](./summarize.py) to summarize the articles in the [cnn_dailymail](https://huggingface.co/datasets/cnn_dailymail) dataset using the model.
## Support Matrix
* FP16
* INT8 & INT4 Weight-Only
* INT8 KV CACHE
* Smooth Quant
* Tensor Parallel
## Usage
The TensorRT-LLM BLOOM example code locates at [examples/bloom](./). It takes HF weights as input, and builds the corresponding TensorRT engines. The number of TensorRT engines depends on the number of GPUs used to run inference.
### Build TensorRT engine(s)
Need to prepare the HF BLOOM checkpoint first by following the guides here https://huggingface.co/docs/transformers/main/en/model_doc/bloom.
e.g. To install BLOOM-560M
```bash
# Setup git-lfs
git lfs install
rm -rf ./bloom/560M
mkdir -p ./bloom/560M && git clone https://huggingface.co/bigscience/bloom-560m ./bloom/560M
```
TensorRT-LLM BLOOM builds TensorRT engine(s) from HF checkpoint. If no checkpoint directory is specified, TensorRT-LLM will build engine(s) with dummy weights.
Normally `build.py` only requires single GPU, but if you've already got all the GPUs needed for inference, you could enable parallel building to make the engine building process faster by adding `--parallel_build` argument. Please note that currently `parallel_build` feature only supports single node.
Here're some examples:
```bash
# Build a single-GPU float16 engine from HF weights.
# Try use_gemm_plugin to prevent accuracy issue. TODO check this holds for BLOOM
# Single GPU on BLOOM 560M
python build.py --model_dir ./bloom/560M/ \
--dtype float16 \
--use_gemm_plugin float16 \
--use_gpt_attention_plugin float16 \
--output_dir ./bloom/560M/trt_engines/fp16/1-gpu/
# Build the BLOOM 560M using a single GPU and apply INT8 weight-only quantization.
python build.py --model_dir ./bloom/560M/ \
--dtype float16 \
--use_gemm_plugin float16 \
--use_gpt_attention_plugin float16 \
--use_weight_only \
--output_dir ./bloom/560M/trt_engines/int8_weight_only/1-gpu/
# Use 2-way tensor parallelism on BLOOM 560M
python build.py --model_dir ./bloom/560M/ \
--dtype float16 \
--use_gemm_plugin float16 \
--use_gpt_attention_plugin float16 \
--output_dir ./bloom/560M/trt_engines/fp16/2-gpu/ \
--world_size 2
# Use 8-way tensor parallelism on BLOOM 176B
# Currently, TensorRT does not support tensors with more than 2^31-1 elements,
# so we have to shard the embedding table to multi-GPUs.
# sharding embedding table in the vocab dimension (the lookup plugin is optional)
python build.py --model_dir ./bloom/176B/ \
--dtype float16 \
--use_gemm_plugin float16 \
--use_gpt_attention_plugin float16 \
--output_dir ./bloom/176B/trt_engines/fp16/8-gpu/ \
--world_size 8 \
--use_parallel_embedding \
--embedding_sharding_dim 0 \
--use_lookup_plugin float16
# sharding embedding table in the hidden dimension
python build.py --model_dir ./bloom/176B/ \
--dtype float16 \
--use_gemm_plugin float16 \
--use_gpt_attention_plugin float16 \
--output_dir ./bloom/176B/trt_engines/fp16/8-gpu/ \
--world_size 8 \
--use_parallel_embedding \
--embedding_sharding_dim 1
# share embedding table between embedding() and lm_head() layers
# To reduce the generated engine size, we has to use gemm and lookup plugin (--use_gemm_plugin --use_lookup_plugin) and must shard the embedding table in the vocab dimension.
python build.py --model_dir ./bloom/176B/ \
--dtype float16 \
--use_gemm_plugin float16 \
--use_gpt_attention_plugin float16 \
--output_dir ./bloom/176B/trt_engines/fp16/8-gpu/ \
--world_size 8 \
--use_parallel_embedding \
--embedding_sharding_dim 0 \
--use_lookup_plugin float16 \
--use_embedding_sharing
```
#### INT8 weight only + INT8 KV cache
For INT8 KV cache, [`hf_bloom_convert.py`](./hf_bloom_convert.py) features a
`--calibrate-kv-cache, -kv` option. Setting `-kv` will calibrate the model,
and then export the scaling factors needed for INT8 KV cache inference.
Example:
```bash
python3 hf_bloom_convert.py -i bloom/560M -o ./c-model/bloom/int8_kv_cache/560M --calibrate-kv-cache -t float16
```
[`build.py`](./build.py) add new options for the support of INT8 KV cache.
`--int8_kv_cache` is the command-line option to enable INT8 KV cache.
In addition, it could be combined with INT8 weight-only quantization, as follows:
Examples of INT8 weight-only quantization + INT8 KV cache
```bash
# Build model with both INT8 weight-only and INT8 KV cache enabled
python build.py --bin_model_dir=./c-model/bloom/int8_kv_cache/560M/1-gpu \
--dtype float16 \
--use_gpt_attention_plugin float16 \
--use_gemm_plugin float16 \
--use_layernorm_plugin \
--int8_kv_cache \
--use_weight_only
```
#### SmoothQuant
Unlike the FP16 build where the HF weights are processed and loaded into the TensorRT-LLM directly, the SmoothQuant needs to load INT8 weights which should be pre-processed before building an engine.
Example:
```bash
python3 hf_bloom_convert.py -i bloom/560M -o ./c-model/bloom-smooth/560M --smoothquant 0.5 --tensor-parallelism 1 --storage-type float16
```
[`build.py`](./build.py) add new options for the support of INT8 inference of SmoothQuant models.
`--use_smooth_quant` is the starting point of INT8 inference. By default, it
will run the model in the _per-tensor_ mode.
Then, you can add any combination of `--per-token` and `--per-channel` to get the corresponding behaviors.
Examples of build invocations:
```bash
# Build model for SmoothQuant in the _per_tensor_ mode.
python3 build.py --bin_model_dir=./c-model/bloom-smooth/560M/1-gpu \
--use_smooth_quant --use_gpt_attention_plugin float16
# Build model for SmoothQuant in the _per_token_ + _per_channel_ mode
python3 build.py --bin_model_dir=./c-model/bloom-smooth/560M/1-gpu \
--use_smooth_quant --use_gpt_attention_plugin float16 \
--per_token \
--per_channel
```
Note that GPT attention plugin is required to be enabled for SmoothQuant for now.
Note we use `--bin_model_dir` instead of `--model_dir` since SmoothQuant model needs INT8 weights and various scales from the binary files.
### 4. Run
```bash
python summarize.py --test_trt_llm \
--hf_model_location ./bloom/560M/ \
--data_type fp16 \
--engine_dir ./bloom/560M/trt_engines/fp16/1-gpu/
python summarize.py --test_trt_llm \
--hf_model_location ./bloom/560M/ \
--data_type fp16 \
--engine_dir ./bloom/560M/trt_engines/int8_weight_only/1-gpu/
mpirun -n 2 --allow-run-as-root \
python summarize.py --test_trt_llm \
--hf_model_location ./bloom/560M/ \
--data_type fp16 \
--engine_dir ./bloom/560M/trt_engines/fp16/2-gpu/
mpirun -n 8 --allow-run-as-root \
python summarize.py --test_trt_llm \
--hf_model_location ./bloom/176B/ \
--data_type fp16 \
--engine_dir ./bloom/176B/trt_engines/fp16/8-gpu/
```