Update README.md

Author: vkuzo

hf_torchao_vllm/README.md

@@ -1,15 +1,40 @@
 # HF -> torchao -> vLLM convenience scripts
 
-## Usage Example
+## hf -> torchao -> vLLM
 
 ```bash
-# save a quantized model ot data/nvfp4-Qwen1.5-MoE-A2.7B
+# save a quantized model to data/torchao/nvfp4-Qwen1.5-MoE-A2.7B
+# models tested: Qwen1.5-MoE-A2.7B (small MoE), facebook/opt-125m (tiny dense)
+# quant_type tested: fp8 (which means fp8 rowwise here), nvfp4 (without any calibration)
 python quantize_hf_model_with_torchao.py --model_name "Qwen/Qwen1.5-MoE-A2.7B" --experts_only_qwen_1_5_moe_a_2_7b True --save_model_to_disk True --quant_type nvfp4
 
 # run the model from above in vLLM
+# requires https://github.com/vllm-project/vllm/pull/26095 (currently not landed)
 python run_quantized_model_in_vllm.py --model_name "data/torchao/nvfp4-Qwen1.5-MoE-A2.7B" --compile False
 ```
 
+## hf -> torchao -> compressed_tensors checkpoint -> vLLM
+
+```bash
+# save a quantized model to data/torchao/fp8-opt-125m
+# models tested: Qwen1.5-MoE-A2.7B (small MoE), facebook/opt-125m (tiny dense)
+# quant_type tested: fp8 (which means fp8 rowwise here), nvfp4 (without any calibration). Note that nvfp4 on the MoE model leads to an error in vLLM.
+python quantize_hf_model_with_torchao.py --model_name "facebook/opt-125m" --experts_only_qwen_1_5_moe_a_2_7b False --save_model_to_disk True --quant_type fp8
+
+# (optional) save a quantized model with llm-compressor to data/llmcompressor/fp8-opt-125m
+python quantize_hf_model_with_llm_compressor.py --model_name facebook/opt-125m --quant_type fp8
+
+# (optional) inspect the torchao and compressed-tensors checkpoints
+python inspect_torchao_output.py --dir_name data/torchao/fp8-opt-125m
+python inspect_llm_compressor_output.py --dir_name data/llmcompressor/fp8-opt-125m
+
+# convert the torchao checkpoint to compressed-tensors format
+python convert_torchao_checkpoint_to_compressed_tensors.py --dir_source data/torchao/fp8-opt-125m --dir_target data/torchao_compressed_tensors/fp8-opt-125m --dir_validation data/llmcompressor/fp8-opt-125m
+
+# run the converted model vLLM
+python run_quantized_model_in_vllm.py --model_name "data/torchao_compressed_tensors/fp8-opt-125m" --compile False
+```
+
 ## Code Quality & Linting
 
 ```bash