| | --- |
| | library_name: pytorch |
| | license: unknown |
| | tags: |
| | - llm |
| | - generative_ai |
| | - android |
| | pipeline_tag: text-generation |
| |
|
| | --- |
| | |
| |  |
| |
|
| | # Ministral-3B: Optimized for Mobile Deployment |
| | ## State-of-the-art large language model useful on a variety of language understanding and generation tasks |
| |
|
| |
|
| | Ministraux are Mistral AI's first premier, commercial AI model designed specifically for on-device use. |
| |
|
| | This model is an implementation of Ministral-3B found [here](https://github.com/mistralai/mistral-inference). |
| |
|
| |
|
| | Please contact us to purchase this model. More details on model performance across various devices, can be found [here](https://aihub.qualcomm.com/models/ministral_3b). |
| |
|
| | **WARNING**: The model assets are not readily available for download due to licensing restrictions. |
| |
|
| | ### Model Details |
| |
|
| | - **Model Type:** Model_use_case.text_generation |
| | - **Model Stats:** |
| | - Input sequence length for Prompt Processor: 128 |
| | - Max context length: 4096 |
| | - Number of parameters: 3B |
| | - Precision: w4a16 + w8a16 (few layers) |
| | - Use: Initiate conversation with prompt-processor and then token generator for subsequent iterations. |
| | - Minimum QNN SDK version required: 2.29.0 |
| | - Supported languages: English. |
| | - TTFT: Time To First Token is the time it takes to generate the first response token. This is expressed as a range because it varies based on the length of the prompt. The lower bound is for a short prompt (up to 128 tokens, i.e., one iteration of the prompt processor) and the upper bound is for a prompt using the full context length (4096 tokens). |
| | - Response Rate: Rate of response generation after the first response token. |
| | |
| | | Model | Precision | Device | Chipset | Target Runtime | Response Rate (tokens per second) | Time To First Token (range, seconds) |
| | |---|---|---|---|---|---| |
| | | Ministral-3B | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_CONTEXT_BINARY | 18.79867 | 0.10237900000000001 - 3.2761280000000004 | -- | -- | |
| | |
| | ## Deploying Ministral 3B on-device |
| | |
| | Please follow the [LLM on-device deployment](https://github.com/quic/ai-hub-apps/tree/main/tutorials/llm_on_genie) tutorial. |
| | |
| | |
| | |
| | |
| | ## References |
| | * [Source Model Implementation](https://github.com/mistralai/mistral-inference) |
| | |
| | |
| | |
| | ## Community |
| | * Join [our AI Hub Slack community](https://qualcomm-ai-hub.slack.com/join/shared_invite/zt-2d5zsmas3-Sj0Q9TzslueCjS31eXG2UA#/shared-invite/email) to collaborate, post questions and learn more about on-device AI. |
| | * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com). |
| | |
| | ## Usage and Limitations |
| | |
| | Model may not be used for or in connection with any of the following applications: |
| | |
| | - Accessing essential private and public services and benefits; |
| | - Administration of justice and democratic processes; |
| | - Assessing or recognizing the emotional state of a person; |
| | - Biometric and biometrics-based systems, including categorization of persons based on sensitive characteristics; |
| | - Education and vocational training; |
| | - Employment and workers management; |
| | - Exploitation of the vulnerabilities of persons resulting in harmful behavior; |
| | - General purpose social scoring; |
| | - Law enforcement; |
| | - Management and operation of critical infrastructure; |
| | - Migration, asylum and border control management; |
| | - Predictive policing; |
| | - Real-time remote biometric identification in public spaces; |
| | - Recommender systems of social media platforms; |
| | - Scraping of facial images (from the internet or otherwise); and/or |
| | - Subliminal manipulation |
| | |
