Slam Scaled

Slamming: Training a Speech Language Model on One GPU in a Day

The model was presented in the paper Slamming: Training a Speech Language Model on One GPU in a Day.

Paper abstract

We introduce Slam, a recipe for training high-quality Speech Language Models (SLMs) on a single academic GPU in 24 hours. We do so through empirical analysis of model initialisation and architecture, synthetic training data, preference optimisation with synthetic data and tweaking all other components. We empirically demonstrate that this training recipe also scales well with more compute getting results on par with leading SLMs in a fraction of the compute cost. We hope these insights will make SLM training and research more accessible. In the context of SLM scaling laws, our results far outperform predicted compute optimal performance, giving an optimistic view to SLM feasibility. See code, data, models, samples at - https://pages.cs.huji.ac.il/adiyoss-lab/slamming .

Model Card for Model ID

This is a Speech Language Model (SLM) trained for generating speech continuations over discrete Hubert tokens.

Model Details

Model Description

This Speech Language Model, introduced in "Slamming: Training a Speech Language Model on One GPU in a Day", focuses on efficient training. It was fine-tuned from Qwen/Qwen2.5-0.5B over a vocabulary of 500 speech tokens extracted from the 11-th layer of mhubert-25hz.

The model was pre-trained using next-token prediction on a subset of LibriSpeech, Libri-Light and a synthetic dataset sTinyStories. It was subsequently fine-tuned with DPO on SpokenSwag.

• Developed by: SLP-RL
• Model type: SpeechLM
• License: MIT
• Finetuned from model: Qwen/Qwen2.5-0.5B

Model Sources

• Repository: https://github.com/slp-rl/slamkit
• Paper: https://arxiv.org/abs/2502.15814
• Demo: https://pages.cs.huji.ac.il/adiyoss-lab/slamming/

Uses

This base SpeechLM can be used to generate continuations for speech segments, or as a base for further tuning. See the SlamKit codebase for more details on usage, and checkout the demo page for some generation examples

Out-of-Scope Use

This model was trained on curated speech datasets which contain mainly audio-books and stories, as such the outputs should not be treated as factual in any way.

How to Get Started with the Model

We refer users to the official repository for full usage explanations - github.

Training Details

We highly encourage users to read the full paper, for full training details, a brief overview is provided below.

Training Data

This model was trained on a subset of LibriSpeech train, Libri-Light and the synthetic dataset sTinyStories for the pre-training phase. It was also trained with DPO on the synthetic dataset SpokenSwag.

Training Procedure

This model was trained by next token prediction over several datasets, and then trained with DPO over SpokenSwag. Please refer to the paper or code for the full training recipes.

Preprocessing

Speech tokens are extracted from the audio using Hubert-25hz, and quantised using the official kmeans released with the model in textlesslib. Units are de-duplicated. We encourage you to explore the official repository for full details - github.

Evaluation

The paper provides full results, we do give here some results and also refer to the demo page to listen to some samples.

Model	GPUs	Params	Num Tokens	sBLIMP ↑	sStoryCloze ↑	tStoryCloze ↑	GenPPL ↓	Auto-BLEU ↓
Speech only pre-training
GSLM	8×V100	100M	1B	54.2	53.3	66.6	—	—
SyllableLM	4×A40	300M	16B	63.7	—	75.4	—	—
TWIST-350M	8×V100	305M	10.8B	56.2	—	—	137.3	3.46
TWIST-1.3B	32×V100	1B	10.8B	57.0	52.4	70.6	131.8	3.20
TWIST-7B	32×V100	7B	36B	59.0	55.3	74.1	93.74	3.06
TWIST-13B	32×V100	13B	36B	59.2	55.4	76.4	—	—
Scaled Optimal	—	823M	82B	61.3	56.7	78.0	—	—
Moshi	?×H100	7B	?	58.9	58.7	81.8	—	—
SpiritLM	64×A100	7B	100B	58.0	54.8	72.9	—	—
With text / preference optimization
Scaling Interleaving	—	9B	~1T	—	62.4	82.9	—	—
Moshi	?×H100	7B	~720B	58.8	60.8	83.0	—	—
SpiritLM	64×A100	7B	100B	58.3	61.0	82.9	—	—
AlignSLM-1.3B	64×A100	1B	10.8B + ~158B	59.8	55.0	80.0	—	—
AlignSLM-7B	64×A100	7B	36B + ~158B	62.3	61.1	86.8	—	—
Ours (Slam)
Slam (-DPO)	2×A100	358M	16.7B	58.53	58.15	80.71	67.3	3.25
Slam	1×A5000	358M	1.4B + 5M	58.86	58.04	82.04	62.8	3.88
Slam (scaled)	2×A100	358M	16.7B + 9M	61.11	61.30	84.18	46.6	3.75
Slam (large)	2×A100	1.3B	6.1B + 9M	61.43	61.52	85.30	41.2	3.89

Compute Infrastructure

This model was trained as part of "Slamming: Training a Speech Language Model on One GPU in a Day", focusing on efficient training.

Hardware

This model was trained using only 2 Nvidia A100 GPU for 48 hours.

Software

The model was trained using the SlamKit codebase which builds upon 🤗transformers extending it to support easy and efficient training of Speech Language Models.

Citation

BibTeX:

@misc{maimon2025slamming,
      title={Slamming: Training a Speech Language Model on One GPU in a Day}, 
      author={Gallil Maimon and Avishai Elmakies and Yossi Adi},
      year={2025},
      eprint={2502.15814},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2502.15814}, 
}