DiscoPOP-zephyr-7b-gemma

Instructions to use SakanaAI/DiscoPOP-zephyr-7b-gemma with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SakanaAI/DiscoPOP-zephyr-7b-gemma with Transformers:

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

pipe = pipeline("text-generation", model="SakanaAI/DiscoPOP-zephyr-7b-gemma")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("SakanaAI/DiscoPOP-zephyr-7b-gemma")
model = AutoModelForCausalLM.from_pretrained("SakanaAI/DiscoPOP-zephyr-7b-gemma")
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

vLLM

How to use SakanaAI/DiscoPOP-zephyr-7b-gemma with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "SakanaAI/DiscoPOP-zephyr-7b-gemma"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SakanaAI/DiscoPOP-zephyr-7b-gemma",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/SakanaAI/DiscoPOP-zephyr-7b-gemma

SGLang

How to use SakanaAI/DiscoPOP-zephyr-7b-gemma 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 "SakanaAI/DiscoPOP-zephyr-7b-gemma" \
    --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": "SakanaAI/DiscoPOP-zephyr-7b-gemma",
		"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 "SakanaAI/DiscoPOP-zephyr-7b-gemma" \
        --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": "SakanaAI/DiscoPOP-zephyr-7b-gemma",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use SakanaAI/DiscoPOP-zephyr-7b-gemma with Docker Model Runner:
```
docker model run hf.co/SakanaAI/DiscoPOP-zephyr-7b-gemma
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

DiscoPOP-zephyr-7b-gemma

This model is a fine-tuned version of HuggingFaceH4/zephyr-7b-gemma-sft-v0.1 on the argilla/dpo-mix-7k dataset.

This model is from the paper "Discovering Preference Optimization Algorithms with and for Large Language Models"

Read the blog post on it here!

See the codebase to generate it here: https://github.com/SakanaAI/DiscoPOP

Model description

This model is identical in training to HuggingFaceH4/zephyr-7b-gemma-v0.1, except instead of using Direct Preference Optimization (DPO), it uses DiscoPOP.

DiscoPOP is our Discovered Preference Optimization algorithm, which is defined as follows:

def log_ratio_modulated_loss(
    self,
    policy_chosen_logps: torch.FloatTensor,
    policy_rejected_logps: torch.FloatTensor,
    reference_chosen_logps: torch.FloatTensor,
    reference_rejected_logps: torch.FloatTensor,
) -> torch.FloatTensor:
    pi_logratios = policy_chosen_logps - policy_rejected_logps
    ref_logratios = reference_chosen_logps - reference_rejected_logps
    logits = pi_logratios - ref_logratios
    # Modulate the mixing coefficient based on the log ratio magnitudes
    log_ratio_modulation = torch.sigmoid(logits)
    logistic_component = -F.logsigmoid(self.beta * logits)
    exp_component = torch.exp(-self.beta * logits)
    # Blend between logistic and exponential component based on log ratio modulation
    losses = logistic_component * (1 - log_ratio_modulation) + exp_component * log_ratio_modulation
    return losses

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e-07
train_batch_size: 2
eval_batch_size: 4
seed: 42
distributed_type: multi-GPU
num_devices: 8
gradient_accumulation_steps: 8
total_train_batch_size: 128
total_eval_batch_size: 32
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 2

Framework versions

Transformers 4.40.1
Pytorch 2.1.2+cu121
Datasets 2.19.0
Tokenizers 0.19.1

Downloads last month: 22

Safetensors

Model size

9B params

Tensor type

BF16

Model tree for SakanaAI/DiscoPOP-zephyr-7b-gemma

Base model

google/gemma-7b

Finetuned

HuggingFaceH4/zephyr-7b-gemma-sft-v0.1

Finetuned

(33)

this model

Quantizations

1 model

Dataset used to train SakanaAI/DiscoPOP-zephyr-7b-gemma

Space using SakanaAI/DiscoPOP-zephyr-7b-gemma 1

Paper for SakanaAI/DiscoPOP-zephyr-7b-gemma

Discovering Preference Optimization Algorithms with and for Large Language Models

Paper • 2406.08414 • Published Jun 12, 2024 • 16