Instructions to use dwightf/BerkshireGPT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use dwightf/BerkshireGPT with Transformers:

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

pipe = pipeline("text-generation", model="dwightf/BerkshireGPT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("dwightf/BerkshireGPT")
model = AutoModelForCausalLM.from_pretrained("dwightf/BerkshireGPT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use dwightf/BerkshireGPT with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/dwightf/BerkshireGPT

SGLang

How to use dwightf/BerkshireGPT 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 "dwightf/BerkshireGPT" \
    --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": "dwightf/BerkshireGPT",
		"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 "dwightf/BerkshireGPT" \
        --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": "dwightf/BerkshireGPT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use dwightf/BerkshireGPT with Docker Model Runner:
```
docker model run hf.co/dwightf/BerkshireGPT
```

dwightf/BerkshireGPT

Model Description

BerkshireGPT is a 7 billion parameter model trained to be a value investor. It is trained on the Berkshire Hathaway annual shareholder meeting transcripts and other value investing material from the web.

It was fine-tuned from the FinGPT/fingpt-forecaster_dow30_llama2-7b_lora model.

It is good at giving advice on stocks as well as answering financial and stock questions.

How to Use

# Use a pipeline as a high-level helper
from transformers import pipeline
pipe = pipeline("text-generation", model="dwightf/BerkshireGPT")
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("dwightf/BerkshireGPT")
model = AutoModelForCausalLM.from_pretrained("dwightf/BerkshireGPT")

More examples at the git repository here.

Input Prompt

A unique input prompt was used to make sure the model focused on value investing

eval_prompt = f"""[INST]<<SYS>>\n You are a value investor giving your advice on stocks. And choosing whether to buy, sell, or hold them. 
    
    <</SYS>>
    
    
    Q -{question}
    
    [/INST]
    
    """

Future Work

We are working on benchmarking the model on common financial benchmarks. We also plan on fine-tuning it on more data, speficially financial data.

Downloads last month: 4

Safetensors

Model size

7B params

Tensor type

F32