Instructions to use VladHong/llama-cpp-K2-FAST2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use VladHong/llama-cpp-K2-FAST2 with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf VladHong/llama-cpp-K2-FAST2 # Run inference directly in the terminal: llama cli -hf VladHong/llama-cpp-K2-FAST2
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf VladHong/llama-cpp-K2-FAST2 # Run inference directly in the terminal: llama cli -hf VladHong/llama-cpp-K2-FAST2
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf VladHong/llama-cpp-K2-FAST2 # Run inference directly in the terminal: ./build/bin/llama cli -hf VladHong/llama-cpp-K2-FAST2
Use Docker
docker model run hf.co/VladHong/llama-cpp-K2-FAST2
- Llama new
- LM Studio
- Jan
- Ollama
How to use VladHong/llama-cpp-K2-FAST2 with Ollama:
ollama run hf.co/VladHong/llama-cpp-K2-FAST2
- Unsloth Desktop
- Docker Model Runner
How to use VladHong/llama-cpp-K2-FAST2 with Docker Model Runner:
docker model run hf.co/VladHong/llama-cpp-K2-FAST2
- Lemonade
How to use VladHong/llama-cpp-K2-FAST2 with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull VladHong/llama-cpp-K2-FAST2
Run and chat with the model
lemonade run user.llama-cpp-K2-FAST2-{{QUANT_TAG}}List all available models
lemonade list
- Atomic Chat
File size: 5,735 Bytes
c59fd93 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 | #include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include "mtmd.h"
#include "mtmd-helper.h"
int main(void) {
printf("\n\nTesting libmtmd C API...\n");
printf("--------\n\n");
struct mtmd_context_params params = mtmd_context_params_default();
printf("Default image marker: %s\n", params.image_marker);
mtmd_input_chunks * chunks = mtmd_test_create_input_chunks();
if (!chunks) {
fprintf(stderr, "Failed to create input chunks\n");
return 1;
}
// simple test for the helper
size_t n_tokens_total = mtmd_helper_get_n_tokens(chunks);
printf("Total tokens in chunks: %zu\n", n_tokens_total);
assert(n_tokens_total > 0);
size_t n_chunks = mtmd_input_chunks_size(chunks);
printf("Number of chunks: %zu\n", n_chunks);
assert(n_chunks > 0);
for (size_t i = 0; i < n_chunks; i++) {
const mtmd_input_chunk * chunk = mtmd_input_chunks_get(chunks, i);
assert(chunk != NULL);
enum mtmd_input_chunk_type type = mtmd_input_chunk_get_type(chunk);
printf("Chunk %zu type: %d\n", i, type);
if (type == MTMD_INPUT_CHUNK_TYPE_TEXT) {
size_t n_tokens;
const llama_token * tokens = mtmd_input_chunk_get_tokens_text(chunk, &n_tokens);
printf(" Text chunk with %zu tokens\n", n_tokens);
assert(tokens != NULL);
assert(n_tokens > 0);
for (size_t j = 0; j < n_tokens; j++) {
assert(tokens[j] >= 0);
printf(" > Token %zu: %d\n", j, tokens[j]);
}
} else if (type == MTMD_INPUT_CHUNK_TYPE_IMAGE) {
const mtmd_image_tokens * image_tokens = mtmd_input_chunk_get_tokens_image(chunk);
size_t n_tokens = mtmd_image_tokens_get_n_tokens(image_tokens);
// get position of the last token, which should be (nx - 1, ny - 1)
struct mtmd_decoder_pos pos = mtmd_image_tokens_get_decoder_pos(image_tokens, 0, n_tokens - 1);
size_t nx = pos.x + 1;
size_t ny = pos.y + 1;
const char * id = mtmd_image_tokens_get_id(image_tokens);
assert(n_tokens > 0);
assert(nx > 0);
assert(ny > 0);
assert(id != NULL);
printf(" Image chunk with %zu tokens\n", n_tokens);
printf(" Image size: %zu x %zu\n", nx, ny);
printf(" Image ID: %s\n", id);
}
}
// test chunk save/load round-trip
for (size_t i = 0; i < n_chunks; i++) {
const mtmd_input_chunk * chunk = mtmd_input_chunks_get(chunks, i);
assert(chunk != NULL);
enum mtmd_input_chunk_type type = mtmd_input_chunk_get_type(chunk);
// query the required buffer size (out_buf == NULL)
size_t expected_len = 0;
int32_t rc = mtmd_input_chunk_save(chunk, NULL, 0, &expected_len);
printf(" Chunk %zu: save query rc = %d, expected_len = %zu\n", i, rc, expected_len);
assert(rc == 0);
assert(expected_len > 0);
// saving into a too-small buffer must fail, not crash
char tiny_buf[1];
rc = mtmd_input_chunk_save(chunk, tiny_buf, sizeof(tiny_buf), NULL);
printf(" Chunk %zu: save into too-small buffer rc = %d (expect non-zero)\n", i, rc);
assert(rc != 0);
// save into a properly-sized buffer
char * buf = (char *) malloc(expected_len);
assert(buf != NULL);
rc = mtmd_input_chunk_save(chunk, buf, expected_len, NULL);
assert(rc == 0);
// loading from a truncated buffer must fail gracefully, not crash
if (expected_len > 1) {
mtmd_input_chunk * bad = mtmd_input_chunk_load(buf, expected_len - 1);
printf(" Chunk %zu: load from truncated buffer = %p (expect NULL)\n", i, (void *) bad);
assert(bad == NULL);
}
// load it back
mtmd_input_chunk * loaded = mtmd_input_chunk_load(buf, expected_len);
assert(loaded != NULL);
// metadata must match the original chunk
assert(mtmd_input_chunk_get_type(loaded) == type);
assert(mtmd_input_chunk_get_n_tokens(loaded) == mtmd_input_chunk_get_n_tokens(chunk));
assert(mtmd_input_chunk_get_n_pos(loaded) == mtmd_input_chunk_get_n_pos(chunk));
if (type == MTMD_INPUT_CHUNK_TYPE_TEXT) {
size_t n_tok_orig, n_tok_loaded;
const llama_token * tok_orig = mtmd_input_chunk_get_tokens_text(chunk, &n_tok_orig);
const llama_token * tok_loaded = mtmd_input_chunk_get_tokens_text(loaded, &n_tok_loaded);
printf(" Chunk %zu: loaded %zu text tokens (orig %zu), first token %d (orig %d)\n",
i, n_tok_loaded, n_tok_orig,
n_tok_loaded > 0 ? tok_loaded[0] : -1,
n_tok_orig > 0 ? tok_orig[0] : -1);
assert(n_tok_orig == n_tok_loaded);
for (size_t j = 0; j < n_tok_orig; j++) {
assert(tok_orig[j] == tok_loaded[j]);
}
} else if (type == MTMD_INPUT_CHUNK_TYPE_IMAGE || type == MTMD_INPUT_CHUNK_TYPE_AUDIO) {
const char * id_orig = mtmd_input_chunk_get_id(chunk);
const char * id_loaded = mtmd_input_chunk_get_id(loaded);
printf(" Chunk %zu: loaded id '%s' (orig '%s')\n", i, id_loaded, id_orig);
assert(id_orig != NULL && id_loaded != NULL);
assert(strcmp(id_orig, id_loaded) == 0);
}
mtmd_input_chunk_free(loaded);
free(buf);
}
printf("Chunk save/load round-trip OK\n");
// Free the chunks
mtmd_input_chunks_free(chunks);
printf("\n\nDONE: test libmtmd C API...\n");
return 0;
}
|