Instructions to use openbmb/BitCPM-CANN-8B-unquantized with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use openbmb/BitCPM-CANN-8B-unquantized with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="openbmb/BitCPM-CANN-8B-unquantized", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("openbmb/BitCPM-CANN-8B-unquantized", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use openbmb/BitCPM-CANN-8B-unquantized with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "openbmb/BitCPM-CANN-8B-unquantized"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "openbmb/BitCPM-CANN-8B-unquantized",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/openbmb/BitCPM-CANN-8B-unquantized

SGLang

How to use openbmb/BitCPM-CANN-8B-unquantized with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "openbmb/BitCPM-CANN-8B-unquantized" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "openbmb/BitCPM-CANN-8B-unquantized",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "openbmb/BitCPM-CANN-8B-unquantized" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "openbmb/BitCPM-CANN-8B-unquantized",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use openbmb/BitCPM-CANN-8B-unquantized with Docker Model Runner:
```
docker model run hf.co/openbmb/BitCPM-CANN-8B-unquantized
```

guanwenyu1995 commited on 2 days ago

Commit

be98f84

verified ·

1 Parent(s): 596de3b

Update README.md

Browse files

Files changed (1) hide show

README.md +7 -7

README.md CHANGED Viewed

@@ -20,9 +20,9 @@ library_name: transformers
 ## Overview
-BitCPM4-CANN-1B-unquantized is the **unquantized QAT (Quantization-Aware Training) checkpoint** of BitCPM4-CANN-1B, designed for **continued pre-training and fine-tuning**. It preserves full-precision latent weights with ternary fake quantizers (weights → {-1, 0, 1} with group-wise scaling, trained via STE) defined in `modeling.py`, enabling the model to keep learning under quantization constraints. For technical details, see our [Technical Report](https://github.com/OpenBMB/MiniCPM/blob/main/docs/BitCPM_CANN.pdf).
-> ⚠️ **This model is NOT for direct inference.** For inference, use the pseudo-quantized version: [openbmb/BitCPM4-CANN-1B](https://huggingface.co/openbmb/BitCPM4-CANN-1B).
 ## Continued Pre-training & Fine-tuning
@@ -30,7 +30,7 @@ The **only requirement** is that the forward pass must go through the bundled `m
 ### Option 1: DeepSpeed (Recommended)
-We provide ready-to-use training scripts in the [example](https://huggingface.co/openbmb/BitCPM4-CANN-1B-unquantized/tree/main/example) directory (using the 1B model as an example):
 - **Continued pre-training**: `example/run.sh` + `example/train.py`
 - **SFT (Supervised Fine-tuning)**: `example/run_sft.sh` + `example/train_sft.py`
@@ -52,7 +52,7 @@ Any framework that supports HuggingFace model loading with custom code can be us
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
-path = 'openbmb/BitCPM4-CANN-1B-unquantized'
 tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True)
 model = AutoModelForCausalLM.from_pretrained(
     path,
@@ -76,13 +76,13 @@ python qat-convert.py \
     --group_size -1
 ```
-The converted model can be loaded for inference in the same way as [openbmb/BitCPM4-CANN-1B](https://huggingface.co/openbmb/BitCPM4-CANN-1B)—no special quantization libraries required.
 ## Workflow
 ```
 ┌─────────────────────────────────┐
-│  BitCPM4-CANN-1B-unquantized  │   ← This model (QAT checkpoint + fake quantizer in modeling.py)
 └───────────────┬─────────────────┘
                 │
                 ▼  Train (DeepSpeed / LLaMA Factory / HF Trainer / ...)
@@ -92,7 +92,7 @@ The converted model can be loaded for inference in the same way as [openbmb/BitC
                 │
                 ▼  python qat-convert.py --quant_type ternary --group_size -1
 ┌─────────────────────────────────┐
-│  Pseudo-quantized model          │   ← Ready for inference (same format as BitCPM4-CANN-1B)
 └─────────────────────────────────┘
 ```

 ## Overview
+BitCPM4-CANN-8B-unquantized is the **unquantized QAT (Quantization-Aware Training) checkpoint** of BitCPM4-CANN-8B, designed for **continued pre-training and fine-tuning**. It preserves full-precision latent weights with ternary fake quantizers (weights → {-1, 0, 1} with group-wise scaling, trained via STE) defined in `modeling.py`, enabling the model to keep learning under quantization constraints. For technical details, see our [Technical Report](https://github.com/OpenBMB/MiniCPM/blob/main/docs/BitCPM_CANN.pdf).
+> ⚠️ **This model is NOT for direct inference.** For inference, use the pseudo-quantized version: [openbmb/BitCPM4-CANN-8B](https://huggingface.co/openbmb/BitCPM4-CANN-8B).
 ## Continued Pre-training & Fine-tuning
 ### Option 1: DeepSpeed (Recommended)
+We provide ready-to-use training scripts in the [example](https://huggingface.co/openbmb/BitCPM4-CANN-8B-unquantized/tree/main/example) directory (using the 1B model as an example):
 - **Continued pre-training**: `example/run.sh` + `example/train.py`
 - **SFT (Supervised Fine-tuning)**: `example/run_sft.sh` + `example/train_sft.py`
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
+path = 'openbmb/BitCPM4-CANN-8B-unquantized'
 tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True)
 model = AutoModelForCausalLM.from_pretrained(
     path,
     --group_size -1
 ```
+The converted model can be loaded for inference in the same way as [openbmb/BitCPM4-CANN-8B](https://huggingface.co/openbmb/BitCPM4-CANN-8B)—no special quantization libraries required.
 ## Workflow
 ```
 ┌─────────────────────────────────┐
+│  BitCPM4-CANN-8B-unquantized  │   ← This model (QAT checkpoint + fake quantizer in modeling.py)
 └───────────────┬─────────────────┘
                 │
                 ▼  Train (DeepSpeed / LLaMA Factory / HF Trainer / ...)
                 │
                 ▼  python qat-convert.py --quant_type ternary --group_size -1
 ┌─────────────────────────────────┐
+│  Pseudo-quantized model          │   ← Ready for inference (same format as BitCPM4-CANN-8B)
 └─────────────────────────────────┘
 ```