|
--- |
|
base_model: neuralmagic/Llama-2-7b-pruned50-retrained-ultrachat |
|
inference: false |
|
model_type: llama |
|
pipeline_tag: text-generation |
|
datasets: |
|
- cerebras/SlimPajama-627B |
|
- HuggingFaceH4/ultrachat_200k |
|
tags: |
|
- sparse |
|
- chat |
|
- deepsparse |
|
--- |
|
|
|
# Llama-2-7b-pruned50-retrained-ultrachat-quant-ds |
|
|
|
This repo contains a [50% sparse Llama 2 7B](https://huggingface.co/neuralmagic/Llama-2-7b-pruned50-retrained) finetuned for chat tasks using the [UltraChat 200k](https://huggingface.co/datasets/HuggingFaceH4/ultrachat_200k) dataset. |
|
It was then quantized to 8-bit weights + activations and exported to deploy with [DeepSparse](https://github.com/neuralmagic/deepsparse), a CPU inference runtime for sparse models. |
|
|
|
**Authors**: Neural Magic, Cerebras |
|
|
|
## Usage |
|
|
|
Below we share some code snippets on how to get quickly started with running the model. |
|
|
|
### Sparse Transfer |
|
|
|
By leveraging a pre-sparsified model's structure, you can efficiently fine-tune on new data, leading to reduced hyperparameter tuning, training times, and computational costs. Learn about this process [here](https://neuralmagic.github.io/docs-v2/get-started/transfer). |
|
|
|
### Running the model |
|
|
|
For accelerated inference with sparsity on CPUs, deploy with [deepsparse](https://github.com/neuralmagic/deepsparse). |
|
|
|
```python |
|
# pip install deepsparse[llm] |
|
from deepsparse import TextGeneration |
|
|
|
model = TextGeneration(model_path="hf:neuralmagic/Llama-2-7b-pruned50-retrained-ultrachat-quant-ds") |
|
|
|
input_text = "Write me a poem about Machine Learning." |
|
outputs = model(formatted_prompt, max_new_tokens=100) |
|
print(outputs.generations[0].text) |
|
``` |
|
|
|
## Evaluation Benchmark Results |
|
|
|
Model evaluation metrics and results. |
|
|
|
| Benchmark | Metric | Llama-2-7b-ultrachat | Llama-2-7b-pruned50-retrained-ultrachat-quant-ds | |
|
|------------------------------------------------|---------------|-------------|-------------------------------| |
|
| [MMLU](https://arxiv.org/abs/2009.03300) | 5-shot | 46.1% | 36.9% | |
|
| [HellaSwag](https://arxiv.org/abs/1905.07830) | 0-shot | 75.9% | 69.0% | |
|
| [WinoGrande](https://arxiv.org/abs/1907.10641) | 5-shot | 72.6% | 65.7% | |
|
| [ARC-c](https://arxiv.org/abs/1911.01547) | 25-shot | 52.8% | 45.7% | |
|
| [TruthfulQA](https://arxiv.org/abs/2109.07958) | 5-shot | 44.8% | 40.5% | |
|
| [GSM8K](https://arxiv.org/abs/2110.14168) | 5-shot | 12.4% | 4.5% | |
|
| [AlpacaEval](https://arxiv.org/abs/2107.03374) ([Llama-2-7b-chat-hf](https://huggingface.co/meta-llama/Llama-2-70b-chat-hf) evaluator) | Win rate | 57.6% | 60.6% | |
|
| [AlpacaEval](https://arxiv.org/abs/2107.03374) (GPT-4 Turbo evaluator) | Win rate | 60.6% | 60.6% | |
|
|
|
## Help |
|
|
|
For further support, and discussions on these models and AI in general, join [Neural Magic's Slack Community](https://join.slack.com/t/discuss-neuralmagic/shared_invite/zt-q1a1cnvo-YBoICSIw3L1dmQpjBeDurQ) |
