Update README.md

README.md

@@ -17,11 +17,11 @@ base_model:
 pipeline_tag: text-generation
 ---
 
-[Phi4-mini](https://huggingface.co/microsoft/Phi-4-mini-instruct) model quantized with [torchao](https://huggingface.co/docs/transformers/main/en/quantization/torchao) int4 weight only quantization, by PyTorch team.
+[Phi4-mini](https://huggingface.co/microsoft/Phi-4-mini-instruct) quantized with [torchao](https://huggingface.co/docs/transformers/main/en/quantization/torchao) int4 weight only quantization, by PyTorch team. Use it directly or serve using [vLLM](https://docs.vllm.ai/en/latest/). Get 67% VRAM reduction and 12-20% speedup on A100 GPUs.
 
 # Quantization Recipe
 
-First need to install the required packages:
+Install the required packages:
 ```
 pip install git+https://github.com/huggingface/transformers@main
 pip install --pre torchao --index-url https://download.pytorch.org/whl/nightly/cu126
@@ -29,7 +29,7 @@ pip install torch
 pip install accelerate
 ```
 
-We used following code to get the quantized model:
+Use the following code to get the quantized model:
 ```
 import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer, TorchAoConfig
@@ -144,7 +144,6 @@ lm_eval --model hf --model_args pretrained=pytorch/Phi-4-mini-instruct-int4wo-hq
 
 # Peak Memory Usage
 
-We can use the following code to get a sense of peak memory usage during inference:
 
 ## Results
 
@@ -155,6 +154,7 @@ We can use the following code to get a sense of peak memory usage during inferen
 
 
 ## Benchmark Peak Memory
+We can use the following code to get a sense of peak memory usage during inference:
 
 ```
 import torch
@@ -198,8 +198,8 @@ print(f"Peak Memory Usage: {mem:.02f} GB")
 
 # Model Performance
 
-Our int4wo is only optimized for batch size 1, so we'll see slowdown in larger batch sizes, we expect this to be used in local server deployment for single or a few users
-and decode tokens per second will be more important than time to first token.
+Our int4wo is only optimized for batch size 1, so expect some slowdown with larger batch sizes, we expect this to be used in local server deployment for single or a few users
+where the decode tokens per second will matters more than the time to first token.
 
 
 ## Results (A100 machine)