update ckpt name of BitNet integration for vLLM

microsoft · Aug 9, 2024 · 7164521 · 7164521
1 parent 471a10c
commit 7164521
Show file tree

Hide file tree

Showing 4 changed files with 8 additions and 8 deletions.
diff --git a/integration/BitNet/README.md b/integration/BitNet/README.md
@@ -18,14 +18,14 @@ We provide two scripts to make the checkpoints for vLLM. The first script is `ge
 cd /root/to/BitBLAS/integration/BitNet
 # make the checkpoint
 ./maint/generate_bitnet_model_native_format.sh
-# the output ckpy will be saved in the `./models/bitnet_b1_58-3B` directory
+# the output ckpy will be saved in the `./models/ckpt_bitnet_b1_58-3B` directory
 ```
 
 The second script is `generate_bitnet_model_bitblas_format.sh`, which is used to make a checkpoint with BitBLAS compressed metadata, which can avoid the online dequantize sage for the profiling of vLLM, which lead to more efficient memory utilization.
 
 ```bash
-./maint/generate_bitnet_model_bitblas_format.sh ./models/bitnet_b1_58-3B ./models/bitnet_b1_58-3B_bitblas
-# the output ckpy will be saved in the `./models/bitnet_b1_58-3B_bitblas` directory
+./maint/generate_bitnet_model_bitblas_format.sh ./models/ckpt_bitnet_b1_58-3B ./models/ckpt_bitnet_b1_58-3B_bitblas
+# the output ckpy will be saved in the `./models/ckpt_bitnet_b1_58-3B_bitblas` directory
 ```
 
 Finnaly, you can use the ckpt in vLLM with:

diff --git a/integration/BitNet/maint/generate_bitnet_model_native_format.sh b/integration/BitNet/maint/generate_bitnet_model_native_format.sh
@@ -14,13 +14,13 @@ mkdir -p models
 cd models
 
 # download the model
-git clone https://huggingface.co/1bitLLM/bitnet_b1_58-3B bitnet_b1_58-3B --depth 1
+git clone https://huggingface.co/1bitLLM/bitnet_b1_58-3B ckpt_bitnet_b1_58-3B --depth 1
 
 # copy quantized config into the model directory
-cp ../maint/quantize_config.json bitnet_b1_58-3B
+cp ../maint/quantize_config.json ckpt_bitnet_b1_58-3B
 
 # get the realpath of the model directory
-MODEL_DIR=$(realpath bitnet_b1_58-3B)
+MODEL_DIR=$(realpath ckpt_bitnet_b1_58-3B)
 
 cd ..
 

diff --git a/integration/BitNet/vllm_workspace/inference_with_compress_format.py b/integration/BitNet/vllm_workspace/inference_with_compress_format.py
@@ -19,7 +19,7 @@
 current_file_path = os.path.realpath(__file__)
 current_dir = os.path.dirname(current_file_path)
 
-ckpt_path = os.path.join(current_dir, "../models/bitnet_b1_58-3B_bitblas")
+ckpt_path = os.path.join(current_dir, "../models/ckpt_bitnet_b1_58-3B_bitblas")
 parser = argparse.ArgumentParser(description="Inference with BitNet")
 parser.add_argument(
     "--ckpt_path",

diff --git a/integration/BitNet/vllm_workspace/inference_with_native_format.py b/integration/BitNet/vllm_workspace/inference_with_native_format.py
@@ -19,7 +19,7 @@
 # get the path of the current file
 current_file_path = os.path.realpath(__file__)
 current_dir = os.path.dirname(current_file_path)
-ckpt_path = os.path.join(current_dir, "../models/bitnet_b1_58-3B_bitblas")
+ckpt_path = os.path.join(current_dir, "../models/ckpt_bitnet_b1_58-3B_bitblas")
 
 parser = argparse.ArgumentParser(description="Inference with BitNet")
 parser.add_argument(