picoLLM Inference Engine
C Quick Start

Platforms

• Linux (x86_64)
• macOS (x86_64, arm64)
• Windows (x86_64)
• Raspberry Pi (4, 5)

Requirements

• C99-compatible compiler
• CMake (3.13+)
• For Windows Only: MinGW is required to build the demo

Picovoice Account & AccessKey

Signup or Login to Picovoice Console to get your AccessKey. Make sure to keep your AccessKey secret.

Quick Start

Setup

1. Clone the repository:

git clone https://github.com/Picovoice/picollm.git

2. Download a picoLLM model file (.pllm) from Picovoice Console.

Usage

1. Include the public header files (picovoice.h and pv_picollm.h).
2. Link the project to an appropriate precompiled library for the target platform and load it.
3. Create an instance of the engine:

pv_picollm_t *pllm = NULL;
pv_status_t status = pv_picollm_init(
    "${ACCESS_KEY}",
    "${MODEL_PATH}",
    "best",
    &pllm);
if (status != PV_STATUS_SUCCESS) {
    // error handling logic
}

4. Generate a prompt completion:

pv_picollm_usage_t usage;
pv_picollm_endpoint_t endpoint;
int32_t num_completion_tokens;
pv_picollm_completion_token_t *completion_tokens;
char *output;
pv_picollm_generate(
    pllm,
    "${PROMPT}",
    -1,    // completion_token_limit
    NULL,  // stop_phrases
    0,     // num_stop_phrases
    -1,    // seed
    0.f,   // presence_penalty
    0.f,   // frequency_penalty
    0.f,   // temperature
    1.f,   // top_p
    0,     // num_top_choices
    NULL,  // stream_callback
    NULL,  // stream_callback_context
    &usage,
    &endpoint,
    &completion_tokens,
    &num_completion_tokens,
    &output);
if (status != PV_STATUS_SUCCESS) {
    // error handling logic
}

5. To interrupt completion generation before it has finished:

pv_picollm_interrupt(object);

6. Release resources explicitly when done with picoLLM:

// after each call to pv_picollm_generate
pv_picollm_delete_completion_tokens(completion_tokens, num_completion_tokens);
pv_picollm_delete_completion(output);

// once done with picoLLM
pv_picollm_delete(pllm);

Demo

For the picoLLM C SDK, we offer a demo application that demonstrates how to use it to generate text from a prompt.

Setup

1. Clone the picoLLM repository from GitHub using HTTPS:

git clone https://github.com/Picovoice/picollm.git

2. Build the demo:

cd picollm
cmake -S demo/c/. -B demo/c/build
cmake --build demo/c/build --target picollm_demo_completion

Usage

To see the usage options for the demo:

./demo/c/build/picollm_demo_completion

Run the command corresponding to your platform from the root of the repository:

./demo/c/build/picollm_demo_completion \
-a ${ACCESS_KEY} \
-m ${MODEL_PATH} \
-l ${LIBRARY_PATH} \
-p ${PROMPT}

For more information on our picoLLM demo for C, head over to our GitHub repository.

Resources