diff --git a/common/common.h b/common/common.h index fc1ae619..b5b67986 100644 --- a/common/common.h +++ b/common/common.h @@ -269,6 +269,8 @@ struct gpt_params { bool spm_infill = false; // suffix/prefix/middle pattern for infill std::string lora_outfile = "ggml-lora-merged-f16.gguf"; + + bool sweep_bench_output_jsonl = false; }; void gpt_params_handle_hf_token(gpt_params & params); diff --git a/examples/sweep-bench/sweep-bench.cpp b/examples/sweep-bench/sweep-bench.cpp index 28b11ecf..4e594de5 100644 --- a/examples/sweep-bench/sweep-bench.cpp +++ b/examples/sweep-bench/sweep-bench.cpp @@ -1,7 +1,15 @@ -#include "arg.h" -#include "common.h" -#include "log.h" +#include "ggml.h" #include "llama.h" +#include "common.h" +#include "llama-vocab.h" + +#ifdef _WIN32 +#define WIN32_LEAN_AND_MEAN +#ifndef NOMINMAX +# define NOMINMAX +#endif +#include +#endif #include #include @@ -16,14 +24,14 @@ static void print_usage(int, char ** argv) { } int main(int argc, char ** argv) { - common_params params; - if (!common_params_parse(argc, argv, params, LLAMA_EXAMPLE_BENCH, print_usage)) { + gpt_params params; + + if (!gpt_params_parse(argc, argv, params)) { + print_usage(argc, argv); return 1; } - common_init(); - // init LLM llama_backend_init(); @@ -31,18 +39,18 @@ int main(int argc, char ** argv) { // initialize the model - llama_model_params model_params = common_model_params_to_llama(params); + llama_model_params model_params = llama_model_params_from_gpt_params(params); - llama_model * model = llama_model_load_from_file(params.model.c_str(), model_params); + llama_model * model = llama_load_model_from_file(params.model.c_str(), model_params); if (model == NULL) { fprintf(stderr , "%s: error: unable to load model\n" , __func__); return 1; } - llama_context_params ctx_params = common_context_params_to_llama(params); + llama_context_params ctx_params = llama_context_params_from_gpt_params(params); - llama_context * ctx = llama_init_from_model(model, ctx_params); + llama_context * ctx = llama_new_context_with_model(model, ctx_params); if (ctx == NULL) { fprintf(stderr , "%s: error: failed to create the llama_context\n" , __func__); @@ -50,10 +58,13 @@ int main(int argc, char ** argv) { } const unsigned int n_kv_max = llama_n_ctx(ctx); - const llama_vocab * vocab = llama_model_get_vocab(model); - const unsigned int n_vocab = llama_vocab_n_tokens(vocab); - const llama_token bos = llama_vocab_bos(vocab); - const llama_token eos = llama_vocab_eos(vocab); + + + const llama_vocab * vocab = llama_get_vocab(ctx); + llama_token bos = llama_token_bos_impl(*vocab); + //llama_token eos = llama_token_eos_impl(*vocab); + + const unsigned int n_vocab = llama_n_vocab(model); // decode in batches of ctx_params.n_batch tokens auto decode_helper = [](llama_context * ctx, llama_batch & batch, int32_t n_batch) { @@ -72,7 +83,7 @@ int main(int argc, char ** argv) { const int ret = llama_decode(ctx, batch_view); if (ret != 0) { - LOG_ERR("failed to decode the batch, n_batch = %d, ret = %d\n", n_batch, ret); + LOG("failed to decode the batch, n_batch = %d, ret = %d\n", n_batch, ret); return false; } @@ -85,7 +96,7 @@ int main(int argc, char ** argv) { const unsigned int pp = params.n_ubatch; const unsigned int tg = params.n_ubatch / 4; - if (!params.batched_bench_output_jsonl) { + if (!params.sweep_bench_output_jsonl) { LOG("\n"); LOG("%s: n_kv_max = %d, n_batch = %d, n_ubatch = %d, flash_attn = %d, n_gpu_layers = %d, n_threads = %u, n_threads_batch = %u\n", __func__, n_kv_max, params.n_batch, params.n_ubatch, params.flash_attn, params.n_gpu_layers, ctx_params.n_threads, ctx_params.n_threads_batch); LOG("\n"); @@ -97,16 +108,15 @@ int main(int argc, char ** argv) { // warm up { - common_batch_add(batch, bos, 0, { 0 }, false); - common_batch_add(batch, eos, 1, { 0 }, false); + llama_batch_add(batch, bos, 0, { 0 }, false); if (!decode_helper(ctx, batch, ctx_params.n_batch)) { - LOG_ERR("%s: llama_decode() failed\n", __func__); + LOG("%s: llama_decode() failed\n", __func__); return 1; } } - common_batch_clear(batch); + llama_batch_clear(batch); llama_kv_cache_clear(ctx); for (unsigned int n_kv = 0; n_kv < n_kv_max; n_kv += params.n_ubatch) { @@ -117,11 +127,11 @@ int main(int argc, char ** argv) { const auto t_tg_start = ggml_time_us(); for (unsigned int i = 0; i < tg; ++i) { - common_batch_clear(batch); - common_batch_add(batch, std::rand() % n_vocab, n_kv + i, { 0 }, true); + llama_batch_clear(batch); + llama_batch_add(batch, std::rand() % n_vocab, n_kv + i, { 0 }, true); if (!decode_helper(ctx, batch, ctx_params.n_batch)) { - LOG_ERR("%s: llama_decode() failed\n", __func__); + LOG("%s: llama_decode() failed\n", __func__); return 1; } } @@ -132,10 +142,10 @@ int main(int argc, char ** argv) { llama_kv_cache_seq_rm(ctx, 0, n_kv, -1); // prepare batch of pp size for prompt processing performance measurement - common_batch_clear(batch); + llama_batch_clear(batch); for (unsigned int i = 0; i < pp; ++i) { - common_batch_add(batch, std::rand() % n_vocab, n_kv + i, { 0 }, false); + llama_batch_add(batch, std::rand() % n_vocab, n_kv + i, { 0 }, false); } batch.logits[batch.n_tokens - 1] = true; @@ -143,7 +153,7 @@ int main(int argc, char ** argv) { const auto t_pp_start = ggml_time_us(); if (!decode_helper(ctx, batch, ctx_params.n_batch)) { - LOG_ERR("%s: llama_decode() failed\n", __func__); + LOG("%s: llama_decode() failed\n", __func__); return 1; } @@ -156,7 +166,7 @@ int main(int argc, char ** argv) { const float speed_pp = pp / t_pp; const float speed_tg = tg / t_tg; - if(params.batched_bench_output_jsonl) { + if(params.sweep_bench_output_jsonl) { LOG( "{\"n_kv_max\": %d, \"n_batch\": %d, \"n_ubatch\": %d, \"flash_attn\": %d, \"n_gpu_layers\": %d, \"n_threads\": %u, \"n_threads_batch\": %u, " "\"pp\": %d, \"tg\": %d, \"n_kv\": %d, \"t_pp\": %f, \"speed_pp\": %f, \"t_tg\": %f, \"speed_tg\": %f }\n", @@ -168,12 +178,10 @@ int main(int argc, char ** argv) { } } - llama_perf_context_print(ctx); - llama_batch_free(batch); llama_free(ctx); - llama_model_free(model); + llama_free_model(model); llama_backend_free();