Instructions to use isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev with Transformers:

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

pipe = pipeline("text-generation", model="isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev

SGLang

How to use isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev 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 "isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev" \
    --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": "isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev",
		"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 "isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev" \
        --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": "isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev with Docker Model Runner:
```
docker model run hf.co/isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev
```

Model Card for Model ID

Model Details

Model Description

This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.

Developed by: Isaac Chung
License: [Apache 2.0]
Finetuned from model [optional]: openbmb/MiniCPM-2B-sft-bf16

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.

How to Get Started with the Model

Use the code below to get started with the model.

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("isaacchung/MiniCPM-2B-RAFT-lora-hotpotqa-dev", trust_remote_code=True)

Training Details

Training Data

isaacchung/hotpotqa-dev-raft-subset

Training Procedure

Training Hyperparameters

-->

See https://github.com/isaac-chung/MiniCPM/commit/213282b679eb8eb054bb13f02af71b9d71ad3721.

Speeds, Sizes, Times [optional]

train_runtime: 4607.6477
train_samples_per_second: 5.209
train_steps_per_second: 0.651
train_loss: 0.5153028686841329
epoch: 9.52

Training Loss

From the last epoch:

'loss': 0.4504, 'grad_norm': 2.259155507591921, 'learning_rate': 2.7586206896551725e-06, 'epoch': 9.02}                                   
{'loss': 0.431, 'grad_norm': 1.7071411656099411, 'learning_rate': 2.586206896551724e-06, 'epoch': 9.05}                                    
{'loss': 0.4627, 'grad_norm': 1.7915555416805786, 'learning_rate': 2.413793103448276e-06, 'epoch': 9.08}                                   
{'loss': 0.4528, 'grad_norm': 1.9988269942330565, 'learning_rate': 2.2413793103448275e-06, 'epoch': 9.11}                                  
{'loss': 0.445, 'grad_norm': 1.8423666856380017, 'learning_rate': 2.0689655172413796e-06, 'epoch': 9.14}                                   
{'loss': 0.4424, 'grad_norm': 1.7539963730934427, 'learning_rate': 1.896551724137931e-06, 'epoch': 9.17}                                   
{'loss': 0.3817, 'grad_norm': 1.755668315740134, 'learning_rate': 1.724137931034483e-06, 'epoch': 9.21}                                    
{'loss': 0.4012, 'grad_norm': 1.8214703589809635, 'learning_rate': 1.5517241379310346e-06, 'epoch': 9.24}                                  
{'loss': 0.4567, 'grad_norm': 1.6490771602855827, 'learning_rate': 1.3793103448275862e-06, 'epoch': 9.27}                                  
{'loss': 0.491, 'grad_norm': 1.5838108179327266, 'learning_rate': 1.206896551724138e-06, 'epoch': 9.3}                                     
{'loss': 0.516, 'grad_norm': 1.7848893180960532, 'learning_rate': 1.0344827586206898e-06, 'epoch': 9.33}                                   
{'loss': 0.3674, 'grad_norm': 1.6589815898285354, 'learning_rate': 8.620689655172415e-07, 'epoch': 9.37}                                   
{'loss': 0.455, 'grad_norm': 1.6377170040397837, 'learning_rate': 6.896551724137931e-07, 'epoch': 9.4}                                     
{'loss': 0.4322, 'grad_norm': 1.7061632686271986, 'learning_rate': 5.172413793103449e-07, 'epoch': 9.43}                                   
{'loss': 0.3934, 'grad_norm': 1.784527156508834, 'learning_rate': 3.4482758620689656e-07, 'epoch': 9.46}                                   
{'loss': 0.4457, 'grad_norm': 1.5131773700813846, 'learning_rate': 1.7241379310344828e-07, 'epoch': 9.49}                                  
{'loss': 0.4026, 'grad_norm': 1.8239453129182908, 'learning_rate': 0.0, 'epoch': 9.52}