Name	Quant method	Size
Llama-2-7b-pruned50-retrained.Q2_K.gguf	Q2_K	2.36GB
Llama-2-7b-pruned50-retrained.IQ3_XS.gguf	IQ3_XS	2.6GB
Llama-2-7b-pruned50-retrained.IQ3_S.gguf	IQ3_S	2.75GB
Llama-2-7b-pruned50-retrained.Q3_K_S.gguf	Q3_K_S	2.75GB
Llama-2-7b-pruned50-retrained.IQ3_M.gguf	IQ3_M	2.9GB
Llama-2-7b-pruned50-retrained.Q3_K.gguf	Q3_K	3.07GB
Llama-2-7b-pruned50-retrained.Q3_K_M.gguf	Q3_K_M	3.07GB
Llama-2-7b-pruned50-retrained.Q3_K_L.gguf	Q3_K_L	3.35GB
Llama-2-7b-pruned50-retrained.IQ4_XS.gguf	IQ4_XS	3.4GB
Llama-2-7b-pruned50-retrained.Q4_0.gguf	Q4_0	3.56GB
Llama-2-7b-pruned50-retrained.IQ4_NL.gguf	IQ4_NL	3.58GB
Llama-2-7b-pruned50-retrained.Q4_K_S.gguf	Q4_K_S	3.59GB
Llama-2-7b-pruned50-retrained.Q4_K.gguf	Q4_K	3.8GB
Llama-2-7b-pruned50-retrained.Q4_K_M.gguf	Q4_K_M	3.8GB
Llama-2-7b-pruned50-retrained.Q4_1.gguf	Q4_1	3.95GB
Llama-2-7b-pruned50-retrained.Q5_0.gguf	Q5_0	4.33GB
Llama-2-7b-pruned50-retrained.Q5_K_S.gguf	Q5_K_S	4.33GB
Llama-2-7b-pruned50-retrained.Q5_K.gguf	Q5_K	4.45GB
Llama-2-7b-pruned50-retrained.Q5_K_M.gguf	Q5_K_M	4.45GB
Llama-2-7b-pruned50-retrained.Q5_1.gguf	Q5_1	4.72GB
Llama-2-7b-pruned50-retrained.Q6_K.gguf	Q6_K	5.15GB
Llama-2-7b-pruned50-retrained.Q8_0.gguf	Q8_0	6.67GB

Original model description:

base_model: meta-llama/Llama-2-7b-hf inference: true model_type: llama pipeline_tag: text-generation datasets:

cerebras/SlimPajama-627B tags:
sparse

Llama-2-7b-pruned50-retrained

This repo contains model files for a Llama 2 7B model that has had 50% of the parameters pruned in one-shot with SparseGPT, then retrained by Cerebras with 45B tokens from SlimPajama while maintaining sparsity.

Official model weights from Enabling High-Sparsity Foundational Llama Models with Efficient Pretraining and Deployment.

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.

Running the model

This model has not been fine-tuned for instruction-following but may be run with the transformers library. For accelerated inference with sparsity, deploy with nm-vllm or deepsparse.

python

# pip install transformers accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("neuralmagic/Llama-2-7b-pruned50-retrained")
model = AutoModelForCausalLM.from_pretrained("neuralmagic/Llama-2-7b-pruned50-retrained", device_map="auto")

input_text = "Write me a poem about Machine Learning."
input_ids = tokenizer(input_text, return_tensors="pt").to("cuda")

outputs = model.generate(**input_ids)
print(tokenizer.decode(outputs[0]))

Evaluation Benchmark Results

Model evaluation metrics and results.

Benchmark	Metric	Llama-2-7b	Llama-2-7b-pruned50-retrained
MMLU	5-shot	46.9%	41.3%
HellaSwag	0-shot	78.6%	76.5%
WinoGrande	5-shot	74.0%	72.1%
ARC-c	25-shot	53.1%	49.8%
TruthfulQA	5-shot	38.8%	37.7%
GSM8K	5-shot	14.5%	9.17%
HumanEval	pass@1	13.4%	14.7%

Model Training Details

Coming soon.

Help

For further support, and discussions on these models and AI in general, join Neural Magic's Slack Community

RichardErkhov/neuralmagic_-_Llama-2-7b-pruned50-retrained-gguf