Hugging Face· PlatformNew model

Bamba: Inference-Efficient Hybrid Mamba2 Model

IBM, Princeton, CMU, and UIUC announce Bamba-9B, a hybrid Mamba2 model trained on 2.2T tokens with open data, demonstrating 2.5x throughput improvement and 2x latency speedup over standard transformers in vLLM, and released with support in transformers, vLLM, TRL, and llama.cpp.

From the source

We introduce Bamba-9B, an inference-efficient Hybrid Mamba2 model trained by IBM, Princeton, CMU, and UIUC on completely open data.

huggingface.co

Announced

18 Dec 2024 00:00 UTC

Detected

03 Aug 2026 21:12 UTC (+593d after)