Aquila 2 Models by Beijing Academy of Artificial Intelligence (BAAI)
About
The Aquila 2 family is a series of bilingual large language models designed to proficiently handle Chinese and English languages. These models vary significantly in size, from 7 billion to 70 billion parameters, offering a broad range for different computational needs. They are developed using the advanced HeuriMentor (HM) framework, which enhances the training process by dynamically adjusting data distributions, thereby boosting efficiency and model performance. This framework includes components such as the Adaptive Training Engine (ATE) and Training State Monitor (TSM). The Aquila 2 models consistently perform well on key benchmarks and have been open-sourced to encourage further innovation. Notably, the Aquila2-34B model retains performance standards even when quantized to Int4 format, highlighting its efficiency and accuracy. The AquilaChat2 variants are specially fine-tuned for conversational applications, demonstrating the versatility of the Aquila 2 family 14.
Current Variants
Use-when guidance is derived from seed capabilities, context, release, and replacement fields.
Use when the workload needs 2k context and 70B parameters.
Use when the workload needs 2k context and 34B parameters.
Use when the workload needs 2k context and 7B parameters.
Use when the workload needs 16k context and 34B parameters.
Use when the workload needs 16k context and 7B parameters.
Use when the workload needs 2k context and 70B parameters.
Use when the workload needs 2k context and 34B parameters.
Use when the workload needs 2k context and 7B parameters.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Aquila Chat 2 70B Expressive | Use when the workload needs 2k context and 70B parameters. | 2023-11 | 2k context70B parameters | Current |
| Aquila Chat 2 34B | Use when the workload needs 2k context and 34B parameters. | 2023-11 | 2k context34B parameters | Current |
| Aquila Chat 2 7B | Use when the workload needs 2k context and 7B parameters. | 2023-11 | 2k context7B parameters | Current |
| Aquila Chat 2 34B-16K | Use when the workload needs 16k context and 34B parameters. | 2023-11 | 16k context34B parameters | Current |
| Aquila Chat 2 7B-16K | Use when the workload needs 16k context and 7B parameters. | 2023-11 | 16k context7B parameters | Current |
| Aquila 2 70B Expressive | Use when the workload needs 2k context and 70B parameters. | 2023-11 | 2k context70B parameters | Current |
| Aquila 2 34B | Use when the workload needs 2k context and 34B parameters. | 2023-11 | 2k context34B parameters | Current |
| Aquila 2 7B | Use when the workload needs 2k context and 7B parameters. | 2023-11 | 2k context7B parameters | Current |
Release Timeline
1 release groupSpecifications(8 models)
| Model | Released | Context | Parameters |
|---|---|---|---|
| Aquila Chat 2 70B Expressive | 2023-11 | 2k | 70B |
| Aquila Chat 2 34B | 2023-11 | 2k | 34B |
| Aquila Chat 2 7B | 2023-11 | 2k | 7B |
| Aquila Chat 2 34B-16K | 2023-11 | 16k | 34B |
| Aquila Chat 2 7B-16K | 2023-11 | 16k | 7B |
| Aquila 2 70B Expressive | 2023-11 | 2k | 70B |
| Aquila 2 34B | 2023-11 | 2k | 34B |
| Aquila 2 7B | 2023-11 | 2k | 7B |
Frequently Asked Questions
- What is Aquila 2 used for?
- Aquila 2 is used for coding, chatbot and role-playing use cases, and long-context generation. The family description and listed model capabilities point to those workloads as the best fit.
- How does Aquila 2 compare to BGE?
- Aquila 2 by Beijing Academy of Artificial Intelligence (BAAI) is strongest where you need coding, while BGE by Beijing Academy of Artificial Intelligence (BAAI) is the closest related family to check for embedding. Aquila 2 has 8 listed variants and reaches up to 16k context, while BGE reaches up to 8k context, so compare the specs and pricing tables before choosing a production model.
- Which Aquila 2 model should I use?
- If price is the main constraint, use the pricing table first because Aquila 2 does not have complete provider pricing in the local data. For the most capable/latest local choice, evaluate Aquila Chat 2 34B-16K with 16k context.
