We introduce Bamba-9B-v2, a decoder-only language model based on the Mamba-2 architecture and is designed to handle a wide range of text generation tasks. Bamba v2 is trained for an additional 1T tokens that significantly improves on Bamba v1. The L1 and L2 leaderboard scores outperform Llama 3.1 8B, which was trained with nearly 5x the amount of data.
Model
Params
# Layers
Hidden Dim.
Attention Heads
GQA
KV Heads
Context Length
Tied Embeddings
Bamba
9B (9.78B)
32
4096
32
Yes
8
4096
False
The current release includes the following models:
We found some of these dependencies picky on PyTorch versions when doing pip install, so
the best way is to build from source for all Mamba dependencies if you hit dependency
issue with your env:
For users using our HF versions of the model, you would need to install the latest transformers which includes our newly merged implementation for our Bamba models:
You can utilize our newly contributed HF integration to run inference on our Bamba models:
python
1from transformers import AutoModelForCausalLM, AutoTokenizer
23model = AutoModelForCausalLM.from_pretrained("ibm-ai-platform/Bamba-9B-v2")4tokenizer = AutoTokenizer.from_pretrained("ibm-ai-platform/Bamba-9B-v2")56message =["Mamba is a snake with following properties "]7inputs = tokenizer(message, return_tensors='pt', return_token_type_ids=False)8response = model.generate(**inputs, max_new_tokens=64)9print(tokenizer.batch_decode(response, skip_special_tokens=True)[0])10
Training
We trained our Bamba model with FSDP using our training repo here.
Note that this training effort was started before FSDP2 and also long before we contributed
Mamba2-Hybrid to HF, so we were doing FSDP1 training with official Mamba implementation.
For users trying to reproduce the training you now have much more options with our newly
contributed HF-version of Mamba2-Hybrid.
Benchmark scores
Base pretrained models
Category
Benchmark
Bamba 9B (3.1T)
General
MMLU (5-shot)
67.92
ARC-C (25-shot)
63.57
GSM8K (5-shot)
41.70
Hellaswag (10-shot)
83.85
OpenbookQA (5-shot)
51.0
Piqa (5-shot)
83.62
TruthfulQA (0-shot)
50.86
Winogrande (5-shot)
79.48
HF OpenLLM- V2*
MMLU-PRO (5-shot)
25.41
BBH (3-shot)
24.78
GPQA (0-shot)
5.93
IFEval (0-shot)
19.0
MATH Lvl 5 (4-shot)
6.42
MuSR (0-shot)
9.28
*For the v2 leaderboard results, we perform normalization and report the normalized results.
Further details on our evaluation and normalization detailes along with run and analysis scripts can be found here.
Fine-tuning
This example shows how to fine tune the bamba model for a specific task using SFT Trainer.
Quantization
We can create a (FP8) quantized model using fms-model-optimizer, which will make the storage and inference even more efficient.
More details about fms-model-optimizer can be found here.
Llama.cpp
There is preliminary work to enable running Bamba architecture models using llama.cpp. This is work-in-progress, so should only be used as a guide for the adventurous!
Known Limitations
Currently, inference is only supported on CPUs
Models quantized with llama-quantize exhibit bad performance
Setup
To enable Bamba support, you'll need to build from source using Gabe's fork.
sh
1git clone --branch BambaArchitecture git@github.com:gabe-l-hart/llama.cpp.git
2cd llama.cpp
3mkdir build
4cd build
5# NOTE: To build with debug symbols and extra logging, use CMAKE_BUILD_TYPE=Debug
6cmake .. -DCMAKE_BUILD_TYPE=Release
7make -j
Conversion to GGUF
You can use a pre-converted GGUF file from Huggingface (e.g. bamba-9b.gguf). If one doesn't exist, you can use the convert_hf_to_gguf.py script from Gabe's fork to perform the conversion manually.
sh
1# Install the python dependencies
2cd /path/to/llama.cpp
3pip install -r requirements/requirements-convert_hf_to_gguf.txt
45# Perform the conversion
6./convert_hf_to_gguf.py /path/to/bamba-model --outfile /path/to/bamba-model/bamba-model.gguf
Run with llama-cli
sh
1# Run the model with no layers on the GPU (CPU-only)
2cd /path/to/llama.cpp
3./bin/llama-cli -ngl 0 -m /path/to/bamba-model/bamba-model.gguf -p "Tell me a story about a developer and their dog"
Quantization with llama-quantize
You can (optionally) quantize the GGUF model using llama.cpp's built in quantizaiton tool llama-quantize.
sh
1# Run the quantization (see llama-quantize --help for all quant types)
2cd /path/to/llama.cpp
3./build/bin/llama-quantize /path/to/bamba-model/bamba-model.gguf Q4_K_M
Contributors
Data collection and curation: We acknowledge and thank AllenAI team for making a high quality open source dataset Dolma as well as Hugging Face data team for making FineWeb-edu and Cosmopedia available. These are tremendous contributions and enable us to create the model today.
Data preprocessing: We thank IBM's internal data preprocessing team, specifically Tuan Hoang Trong, Syed Zawad, Jay Gala, and Ryan Gordon for helping tokenize the data at scale. The code for tokenization is available here.
Model architecture: The model architecture design was jointly done by Princeton, CMU, IBM, and UIUC and involved the following folks: Tri Dao (Princeton), Albert Gu (CMU), Linsong Chu (IBM), Davis Wertheimer (IBM), Minjia Zhang (UIUC), Mudhakar Srivatsa (IBM), and Raghu Ganti (IBM).
Model training: Model training was performed primarily by the IBM team using the Mamba2 kernels and layer implementation from Tri Dao and Albert Gu. The following folks from IBM were primarily involved: Linsong Chu, Divya Kumari, Davis Wertheimer, Raghu Ganti, and Dakshi Agrawal.
Model tuning: Tuning of the model was enabled and verified in TRL by the IBM team, involving Sukriti Sharma and Anh Uong.
Model inference: Model inference in transformers, vLLM, and llama.cpp builds on the kernels written by Princeton and CMU. The IBM team is working with the community to enable it in various ecosystems, the team includes Fabian Lim, Antoni viros i Martin, Adnan Hoque, Jamie Yang, Nelson Nimura Gomez, Joshua Rosenkranz, Nick Hill, and Gabe Goodhart.
Quantization: Quantization is led by the IBM team - Naigang Wang and Charlie Liu.
Evaluations: Evaluations are led by a team in IBM with long context evaluations being performed by UIUC, involving the following folks: Yotam Perlitz, Ofir Arviv, Michal Shmueli-Scheuer (IBM), Haoechen Shen, and Minjia Zhang (UIUC).
Finally, we would like to thank our leadership for their support in this effort - Priya Nagpurkar, David Cox, Sriram Raghavan, Aya Soffer, and Mukesh Khare.
We would also like to thank the community, in particular Pablo Montalvo-Leroux and Vaibhav Srivastav from Hugging Face who provided valuable feedback to this blog and the PRs into transformers. Further, we would like to thank Tyler Michael Smith from Neural Magic, who is shepherding the integration with vLLM.
A huge shoutout to Meta PyTorch, AllenAI, and Hugging Face teams for their contributions to the open initative, FSDP allowed us to smoothly train this model and the data from Dolma and Fineweb/Cosmopedia made this model today!
