metadata

license: apache-2.0
datasets:
  - tiiuae/falcon-refinedweb
pipeline_tag: text-generation
library_name: openlm
tags:
  - mamba
  - linear
language:
  - en
model-index:
  - name: mamba-7b
    results:
      - task:
          type: text-generation
        dataset:
          type: MMLU
          name: MMLU
        metrics:
          - name: accuracy
            type: accuracy
            value: 33.3
            verified: false
      - task:
          type: text-generation
        dataset:
          type: HellaSwag
          name: HellaSwag
        metrics:
          - name: accuracy
            type: accuracy
            value: 77.9
            verified: false
      - task:
          type: text-generation
        dataset:
          type: PIQA
          name: PIQA
        metrics:
          - name: accuracy
            type: accuracy
            value: 81
            verified: false
      - task:
          type: text-generation
        dataset:
          type: Winogrande
          name: Winogrande
        metrics:
          - name: accuracy
            type: accuracy
            value: 71.8
            verified: false
      - task:
          type: text-generation
        dataset:
          type: ai2_arc
          name: ARC-E
        metrics:
          - name: accuracy
            type: accuracy
            value: 77.5
            verified: false
      - task:
          type: text-generation
        dataset:
          type: ai2_arc
          name: ARC-C
        metrics:
          - name: accuracy
            type: accuracy
            value: 46.7
            verified: false

Mamba-7B

(insert cool midjourney pic here?)
This is a 7B parameter model with the Mamba architecture, trained on 1.2T tokens of the RefinedWeb dataset. Mamba is a state-space model that does not use self-attention unlike the standard transformer architecture. It has shown strong performance on various natural language benchmarks. To date, the largest publicly released pure-Mamba pretrain is Mamba-2.8B. We follow their training recipe and release our version of Mamba-7B.

Model Details

Developed by: Toyota Research Institute
Model Type: This is an auto-regressive language model based on the Mamba architecture.
Dataset: Trained on 1.2T tokens of RefinedWeb
Tokenizer: EleutherAI/gpt-neox-20b
Library: OpenLM
License: This model is licensed under Apache License, Version 2.0.

Parameters	Hidden Size	Layers	Vocab Size	Sequence Length
7B	4096	64	50432	2048

Training Details

Mamba-7B was trained using AWS SageMaker on 128 H100 80GB GPUs.
Training began in March 2024 and lasted around 3 weeks (some down time due to crashes and loss spikes)

Hyperparameter Value

Precision bfloat16

Optimizer AdamW

Learning rate 3e-4

LR cooldown end 1e-5

QK-norm False

Warmup steps 2000

Z-loss 1e-4

Batch size 2M

Hyperparameter	Value
Precision	`bfloat16`
Optimizer	AdamW
Learning rate	3e-4
LR cooldown end	1e-5
QK-norm	False
Warmup steps	2000
Z-loss	1e-4
Batch size	2M

Usage

This model was trained using OpenLM. The weights have been converted to be compatible with HuggingFace.

from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("tri-ml/mamba-7b-rw")
model = AutoModelForCausalLM.from_pretrained("tri-ml/mamba-7b-rw").cuda()

inputs = tokenizer(["A beautiful flower"], return_tensors="pt")
gen_kwargs = {"max_length": 128, "top_p": 0.8, "temperature": 0.8, "do_sample": True, "repetition_penalty": 1.1}
output = model.generate(inputs['input_ids'], **gen_kwargs)
output = tokenizer.decode(output[0].tolist(), skip_special_tokens=True)
print(output)
# <s> A beautiful flower box made of white rose wood. It is a perfect gift for weddings, birthdays and anniversaries.
# All the roses are from our farm Roses Flanders. Therefor you know that these flowers last much longer than those in store or online!</s>

Performance Evaluation

Our evaluations were done using the Eleuther LM Eval Harness repo.

Below we report the performance of Mamba 7B compared to other base models.

	HellaSwag	PIQA	Winogrande	ARC-E	ARC-C	MMLU (5-shot)
Mamba-1.4B	59.0	73.9	61.4	65.5	32.9	25.2
Mamba-2.8B	71.0	78.1	65.9	68.2	41.7	26.2
Llama2-7B	76.0	79.1	69.1	76.3	46.3	45.9
Gemma-7B	80.7	81.9	73.7	81.1	53.2	62.9
Mistral-7B	81.0	82.1	74.0	80.9	53.8	62.4
Mamba-7B	77.9	81.0	71.8	77.5	46.7	33.3

How to Cite

If you use this model, please cite our paper on Linearizing Large Language Models.

@article{Mercat2024Linearizing,
  title={Linearizing Large Language Models},
  author={Jean Mercat and Igor Vasiljevic and Sedrick Keh and Kushal Arora and Achal Dave and Adrien Gaidon and Thomas Kollar},
  journal={ArXiv},
  year={2024},
  volume={},
}

Citations

Mamba

@article{mamba,
  title={Mamba: Linear-Time Sequence Modeling with Selective State Spaces},
  author={Gu, Albert and Dao, Tri},
  journal={arXiv preprint arXiv:2312.00752},
  year={2023}
}

OpenLM

@misc{open_lm,
  author = {Gururangan, Suchin and Wortsman, Mitchell and Gadre, Samir Yitzhak and Dave, Achal and Kilian, Maciej and Shi, Weijia and Mercat, Jean and Smyrnis, Georgios and Ilharco, Gabriel and Jordan, Matt and Heckel, Reinhard and Dimakis, Alex and Farhadi, Ali and Shankar, Vaishaal and Schmidt, Ludwig},
  title = {{open_lm}:  a minimal but performative language modeling (LM) repository},
  year = {2023},
  note = {GitHub repository},
  url = {https://github.com/mlfoundations/open_lm/}
}