diff --git a/.gitignore b/.gitignore index 5eb1ff1b873f1..5838a12c3b17b 100644 --- a/.gitignore +++ b/.gitignore @@ -1,5 +1,7 @@ *.o *.a +*.so +*.dylib .cache/ .vs/ .vscode/ @@ -18,6 +20,7 @@ models/* /main /quantize +/chat arm_neon.h compile_commands.json diff --git a/Makefile b/Makefile index 8388c290d75ce..a430a37af2e4d 100644 --- a/Makefile +++ b/Makefile @@ -16,6 +16,7 @@ CXXV := $(shell $(CXX) --version | head -n 1) # Mac OS + Arm can report x86_64 # ref: https://github.com/ggerganov/whisper.cpp/issues/66#issuecomment-1282546789 ifeq ($(UNAME_S),Darwin) + SHARED := dylib ifneq ($(UNAME_P),arm) SYSCTL_M := $(shell sysctl -n hw.optional.arm64) ifeq ($(SYSCTL_M),1) @@ -24,6 +25,8 @@ ifeq ($(UNAME_S),Darwin) warn := $(warning Your arch is announced as x86_64, but it seems to actually be ARM64. Not fixing that can lead to bad performance. For more info see: https://github.com/ggerganov/whisper.cpp/issues/66\#issuecomment-1282546789) endif endif +else + SHARED := so endif # @@ -172,7 +175,7 @@ $(info I CC: $(CCV)) $(info I CXX: $(CXXV)) $(info ) -default: main quantize +default: chat libllama.$(SHARED) libllama.a main quantize # # Build library @@ -184,13 +187,25 @@ ggml.o: ggml.c ggml.h utils.o: utils.cpp utils.h $(CXX) $(CXXFLAGS) -c utils.cpp -o utils.o +libllama.o: libllama.cpp libllama.h + $(CXX) $(CXXFLAGS) -c libllama.cpp -o libllama.o + clean: - rm -f *.o main quantize + rm -f *.o main quantize chat libllama.$(SHARED) libllama.a main: main.cpp ggml.o utils.o $(CXX) $(CXXFLAGS) main.cpp ggml.o utils.o -o main $(LDFLAGS) ./main -h +libllama.$(SHARED): libllama.o ggml.o utils.o + $(CXX) -shared $(CXXFLAGS) libllama.o ggml.o utils.o -o libllama.$(SHARED) $(LDFLAGS) + +libllama.a: libllama.o ggml.o utils.o + ar -crs libllama.a libllama.o ggml.o utils.o + +chat: chat.cpp libllama.a + $(CXX) $(CXXFLAGS) chat.cpp libllama.a -o chat $(LDFLAGS) + quantize: quantize.cpp ggml.o utils.o $(CXX) $(CXXFLAGS) quantize.cpp ggml.o utils.o -o quantize $(LDFLAGS) diff --git a/chat.cpp b/chat.cpp new file mode 100644 index 0000000000000..8bcfe00b6cc2f --- /dev/null +++ b/chat.cpp @@ -0,0 +1,19 @@ +#include "libllama.h" + +#include + +int main() +{ + void *session = llama_init("./models/7B/ggml-model-q4_0.bin", -1, 10, 128, 40, 0.95f, 64, 1.30f, 0.80f, 8); + if (!session) { + return 1; + } + + char buf[1024*16]; + while (true) { + llama_query(session, "Hey, how are you? ", buf, sizeof(buf)); + std::cout << buf << std::endl; + } + + llama_free(session); +} diff --git a/libllama.cpp b/libllama.cpp new file mode 100644 index 0000000000000..4cd2b765e7324 --- /dev/null +++ b/libllama.cpp @@ -0,0 +1,918 @@ +#include "libllama.h" +#include "ggml.h" + +#include "utils.h" + +#include +#include +#include +#include +#include +#include +#include +#include + +// determine number of model parts based on the dimension +static const std::map LLAMA_N_PARTS = { + { 4096, 1 }, + { 5120, 2 }, + { 6656, 4 }, + { 8192, 8 }, +}; + +// default hparams (LLaMA 7B) +struct llama_hparams { + int32_t n_vocab = 32000; + int32_t n_ctx = 512; // this is provided as user input? + int32_t n_embd = 4096; + int32_t n_mult = 256; + int32_t n_head = 32; + int32_t n_layer = 32; + int32_t n_rot = 64; + int32_t f16 = 1; +}; + +struct llama_layer { + // normalization + struct ggml_tensor * attention_norm; + + // attention + struct ggml_tensor * wq; + struct ggml_tensor * wk; + struct ggml_tensor * wv; + struct ggml_tensor * wo; + + // normalization + struct ggml_tensor * ffn_norm; + + // ff + struct ggml_tensor * w1; + struct ggml_tensor * w2; + struct ggml_tensor * w3; +}; + +struct llama_model { + llama_hparams hparams; + + struct ggml_tensor * tok_embeddings; + + struct ggml_tensor * norm; + struct ggml_tensor * output; + + std::vector layers; + + // key + value memory + struct ggml_tensor * memory_k; + struct ggml_tensor * memory_v; + + // + struct ggml_context * ctx; + std::map tensors; +}; + +struct llama { + std::mt19937 rng; + gpt_params params; + gpt_vocab vocab; + struct llama_model model; +}; + +// load the model's weights from a file +bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab & vocab, int n_ctx) { + //printf("%s: loading model from '%s' - please wait ...\n", __func__, fname.c_str()); + + auto fin = std::ifstream(fname, std::ios::binary); + if (!fin) { + fprintf(stderr, "%s: failed to open '%s'\n", __func__, fname.c_str()); + return false; + } + + // verify magic + { + uint32_t magic; + fin.read((char *) &magic, sizeof(magic)); + if (magic != 0x67676d6c) { + fprintf(stderr, "%s: invalid model file '%s' (bad magic)\n", __func__, fname.c_str()); + return false; + } + } + + int n_ff = 0; + int n_parts = 0; + + // load hparams + { + auto & hparams = model.hparams; + + fin.read((char *) &hparams.n_vocab, sizeof(hparams.n_vocab)); + //fin.read((char *) &hparams.n_ctx, sizeof(hparams.n_ctx)); + fin.read((char *) &hparams.n_embd, sizeof(hparams.n_embd)); + fin.read((char *) &hparams.n_mult, sizeof(hparams.n_mult)); + fin.read((char *) &hparams.n_head, sizeof(hparams.n_head)); + fin.read((char *) &hparams.n_layer, sizeof(hparams.n_layer)); + fin.read((char *) &hparams.n_rot, sizeof(hparams.n_rot)); + fin.read((char *) &hparams.f16, sizeof(hparams.f16)); + + hparams.n_ctx = n_ctx; + + n_ff = ((2*(4*hparams.n_embd)/3 + hparams.n_mult - 1)/hparams.n_mult)*hparams.n_mult; + n_parts = LLAMA_N_PARTS.at(hparams.n_embd); + + //printf("%s: n_vocab = %d\n", __func__, hparams.n_vocab); + //printf("%s: n_ctx = %d\n", __func__, hparams.n_ctx); + //printf("%s: n_embd = %d\n", __func__, hparams.n_embd); + //printf("%s: n_mult = %d\n", __func__, hparams.n_mult); + //printf("%s: n_head = %d\n", __func__, hparams.n_head); + //printf("%s: n_layer = %d\n", __func__, hparams.n_layer); + //printf("%s: n_rot = %d\n", __func__, hparams.n_rot); + //printf("%s: f16 = %d\n", __func__, hparams.f16); + //printf("%s: n_ff = %d\n", __func__, n_ff); + //printf("%s: n_parts = %d\n", __func__, n_parts); + } + + // load vocab + { + const int32_t n_vocab = model.hparams.n_vocab; + + if (n_vocab != model.hparams.n_vocab) { + fprintf(stderr, "%s: invalid model file '%s' (bad vocab size %d != %d)\n", + __func__, fname.c_str(), n_vocab, model.hparams.n_vocab); + return false; + } + + std::string word; + for (int i = 0; i < n_vocab; i++) { + uint32_t len; + fin.read((char *) &len, sizeof(len)); + + word.resize(len); + fin.read((char *) word.data(), len); + + vocab.token_to_id[word] = i; + vocab.id_to_token[i] = word; + + //if (i < 30000) { + // printf("%s: vocab[%d] = '%s'\n", __func__, i, word.c_str()); + //} + } + } + + // for the big tensors, we have the option to store the data in 16-bit floats or quantized + // in order to save memory and also to speed up the computation + ggml_type wtype = GGML_TYPE_COUNT; + switch (model.hparams.f16) { + case 0: wtype = GGML_TYPE_F32; break; + case 1: wtype = GGML_TYPE_F16; break; + case 2: wtype = GGML_TYPE_Q4_0; break; + case 3: wtype = GGML_TYPE_Q4_1; break; + default: + { + fprintf(stderr, "%s: invalid model file '%s' (bad f16 value %d)\n", + __func__, fname.c_str(), model.hparams.f16); + return false; + } + } + + const ggml_type wtype2 = GGML_TYPE_F32; + + auto & ctx = model.ctx; + + size_t ctx_size = 0; + + { + const auto & hparams = model.hparams; + + const int n_embd = hparams.n_embd; + const int n_layer = hparams.n_layer; + const int n_ctx = hparams.n_ctx; + const int n_vocab = hparams.n_vocab; + + ctx_size += n_embd*n_vocab*ggml_type_sizef(wtype); // tok_embeddings + + ctx_size += n_embd*ggml_type_sizef(GGML_TYPE_F32); // norm + + ctx_size += n_embd*n_vocab*ggml_type_sizef(wtype); // output + + ctx_size += n_layer*(n_embd*ggml_type_sizef(GGML_TYPE_F32)); // attention_norm + + ctx_size += n_layer*(n_embd*n_embd*ggml_type_sizef(wtype)); // wq + ctx_size += n_layer*(n_embd*n_embd*ggml_type_sizef(wtype)); // wk + ctx_size += n_layer*(n_embd*n_embd*ggml_type_sizef(wtype)); // wv + ctx_size += n_layer*(n_embd*n_embd*ggml_type_sizef(wtype)); // wo + + ctx_size += n_layer*(n_embd*ggml_type_sizef(GGML_TYPE_F32)); // ffn_norm + + ctx_size += n_layer*(n_ff*n_embd*ggml_type_sizef(wtype)); // w1 + ctx_size += n_layer*(n_ff*n_embd*ggml_type_sizef(wtype)); // w2 + ctx_size += n_layer*(n_ff*n_embd*ggml_type_sizef(wtype)); // w3 + + ctx_size += n_ctx*n_layer*n_embd*ggml_type_sizef(GGML_TYPE_F32); // memory_k + ctx_size += n_ctx*n_layer*n_embd*ggml_type_sizef(GGML_TYPE_F32); // memory_v + + ctx_size += (5 + 10*n_layer)*256; // object overhead + + //printf("%s: ggml ctx size = %6.2f MB\n", __func__, ctx_size/(1024.0*1024.0)); + } + + // create the ggml context + { + struct ggml_init_params params = { + .mem_size = ctx_size, + .mem_buffer = NULL, + }; + + model.ctx = ggml_init(params); + if (!model.ctx) { + fprintf(stderr, "%s: ggml_init() failed\n", __func__); + return false; + } + } + + // prepare memory for the weights + { + const auto & hparams = model.hparams; + + const int n_embd = hparams.n_embd; + const int n_layer = hparams.n_layer; + const int n_ctx = hparams.n_ctx; + const int n_vocab = hparams.n_vocab; + + model.layers.resize(n_layer); + + model.tok_embeddings = ggml_new_tensor_2d(ctx, wtype, n_embd, n_vocab); + + model.norm = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, n_embd); + model.output = ggml_new_tensor_2d(ctx, wtype, n_embd, n_vocab); + + // map by name + model.tensors["tok_embeddings.weight"] = model.tok_embeddings; + + model.tensors["norm.weight"] = model.norm; + model.tensors["output.weight"] = model.output; + + for (int i = 0; i < n_layer; ++i) { + auto & layer = model.layers[i]; + + layer.attention_norm = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, n_embd); + + layer.wq = ggml_new_tensor_2d(ctx, wtype, n_embd, n_embd); + layer.wk = ggml_new_tensor_2d(ctx, wtype, n_embd, n_embd); + layer.wv = ggml_new_tensor_2d(ctx, wtype, n_embd, n_embd); + layer.wo = ggml_new_tensor_2d(ctx, wtype, n_embd, n_embd); + + layer.ffn_norm = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, n_embd); + + layer.w1 = ggml_new_tensor_2d(ctx, wtype, n_embd, n_ff); + layer.w2 = ggml_new_tensor_2d(ctx, wtype, n_ff, n_embd); + layer.w3 = ggml_new_tensor_2d(ctx, wtype, n_embd, n_ff); + + // map by name + model.tensors["layers." + std::to_string(i) + ".attention_norm.weight"] = layer.attention_norm; + + model.tensors["layers." + std::to_string(i) + ".attention.wq.weight"] = layer.wq; + model.tensors["layers." + std::to_string(i) + ".attention.wk.weight"] = layer.wk; + model.tensors["layers." + std::to_string(i) + ".attention.wv.weight"] = layer.wv; + model.tensors["layers." + std::to_string(i) + ".attention.wo.weight"] = layer.wo; + + model.tensors["layers." + std::to_string(i) + ".ffn_norm.weight"] = layer.ffn_norm; + + model.tensors["layers." + std::to_string(i) + ".feed_forward.w1.weight"] = layer.w1; + model.tensors["layers." + std::to_string(i) + ".feed_forward.w2.weight"] = layer.w2; + model.tensors["layers." + std::to_string(i) + ".feed_forward.w3.weight"] = layer.w3; + } + } + + // key + value memory + { + const auto & hparams = model.hparams; + + const int n_embd = hparams.n_embd; + const int n_layer = hparams.n_layer; + const int n_ctx = hparams.n_ctx; + + const int n_mem = n_layer*n_ctx; + const int n_elements = n_embd*n_mem; + + model.memory_k = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, n_elements); + model.memory_v = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, n_elements); + + const size_t memory_size = ggml_nbytes(model.memory_k) + ggml_nbytes(model.memory_v); + + printf("%s: memory_size = %8.2f MB, n_mem = %d\n", __func__, memory_size/1024.0/1024.0, n_mem); + } + + const size_t file_offset = fin.tellg(); + + fin.close(); + + std::vector tmp; + + for (int i = 0; i < n_parts; ++i) { + const int part_id = i; + //const int part_id = n_parts - i - 1; + + std::string fname_part = fname; + if (i > 0) { + fname_part += "." + std::to_string(i); + } + + //printf("%s: loading model part %d/%d from '%s'\n", __func__, i+1, n_parts, fname_part.c_str()); + + fin = std::ifstream(fname_part, std::ios::binary); + fin.seekg(file_offset); + + // load weights + { + int n_tensors = 0; + size_t total_size = 0; + + //printf("%s: ", __func__); + + while (true) { + int32_t n_dims; + int32_t length; + int32_t ftype; + + fin.read(reinterpret_cast(&n_dims), sizeof(n_dims)); + fin.read(reinterpret_cast(&length), sizeof(length)); + fin.read(reinterpret_cast(&ftype), sizeof(ftype)); + + if (fin.eof()) { + break; + } + + int32_t nelements = 1; + int32_t ne[2] = { 1, 1 }; + for (int i = 0; i < n_dims; ++i) { + fin.read(reinterpret_cast(&ne[i]), sizeof(ne[i])); + nelements *= ne[i]; + } + + std::string name(length, 0); + fin.read(&name[0], length); + + if (model.tensors.find(name.data()) == model.tensors.end()) { + fprintf(stderr, "%s: unknown tensor '%s' in model file\n", __func__, name.data()); + return false; + } + + // split_type = 0: split by columns + // split_type = 1: split by rows + int split_type = 0; + + // split_type = 0: + // regex: + // - tok_embeddings.* + // - layers.*.attention.wo.weight + // - layers.*.feed_forward.w2.weight + + // split_type = 1: + // regex: + // - output.* + // - layers.*.attention.wq.weight + // - layers.*.attention.wk.weight + // - layers.*.attention.wv.weight + // - layers.*.feed_forward.w1.weight + // - layers.*.feed_forward.w3.weight + if (name.find("tok_embeddings") != std::string::npos) { + split_type = 0; + } else if (name.find("layers") != std::string::npos) { + if (name.find("attention.wo.weight") != std::string::npos) { + split_type = 0; + } else if (name.find("feed_forward.w2.weight") != std::string::npos) { + split_type = 0; + } else { + split_type = 1; + } + } else if (name.find("output") != std::string::npos) { + split_type = 1; + } + + auto tensor = model.tensors[name.data()]; + + if (n_dims == 1) { + if (ggml_nelements(tensor) != nelements) { + fprintf(stderr, "%s: tensor '%s' has wrong size in model file\n", __func__, name.data()); + return false; + } + } else { + if (ggml_nelements(tensor)/n_parts != nelements) { + fprintf(stderr, "%s: tensor '%s' has wrong size in model file\n", __func__, name.data()); + return false; + } + } + + if (n_dims == 1) { + if (tensor->ne[0] != ne[0] || tensor->ne[1] != ne[1]) { + fprintf(stderr, "%s: tensor '%s' has wrong shape in model file: got [%d, %d], expected [%d, %d]\n", + __func__, name.data(), tensor->ne[0], tensor->ne[1], ne[0], ne[1]); + return false; + } + } else { + if (split_type == 0) { + if (tensor->ne[0]/n_parts != ne[0] || tensor->ne[1] != ne[1]) { + fprintf(stderr, "%s: tensor '%s' has wrong shape in model file: got [%d, %d], expected [%d, %d]\n", + __func__, name.data(), tensor->ne[0]/n_parts, tensor->ne[1], ne[0], ne[1]); + return false; + } + } else { + if (tensor->ne[0] != ne[0] || tensor->ne[1]/n_parts != ne[1]) { + fprintf(stderr, "%s: tensor '%s' has wrong shape in model file: got [%d, %d], expected [%d, %d]\n", + __func__, name.data(), tensor->ne[0], tensor->ne[1]/n_parts, ne[0], ne[1]); + return false; + } + } + } + + if (0) { + static const char * ftype_str[] = { "f32", "f16", "q4_0", "q4_1", }; + //printf("%24s - [%5d, %5d], type = %6s, split = %d\n", name.data(), ne[0], ne[1], ftype_str[ftype], split_type); + } + + size_t bpe = 0; + + switch (ftype) { + case 0: bpe = ggml_type_size(GGML_TYPE_F32); break; + case 1: bpe = ggml_type_size(GGML_TYPE_F16); break; + case 2: bpe = ggml_type_size(GGML_TYPE_Q4_0); assert(ne[0] % 64 == 0); break; + case 3: bpe = ggml_type_size(GGML_TYPE_Q4_1); assert(ne[0] % 64 == 0); break; + default: + { + fprintf(stderr, "%s: unknown ftype %d in model file\n", __func__, ftype); + return false; + } + }; + + if (n_dims == 1 || n_parts == 1) { + if ((nelements*bpe)/ggml_blck_size(tensor->type) != ggml_nbytes(tensor)) { + fprintf(stderr, "%s: tensor '%s' has wrong size in model file: got %zu, expected %zu\n", + __func__, name.data(), ggml_nbytes(tensor), nelements*bpe); + return false; + } + + if (part_id == 0) { + fin.read(reinterpret_cast(tensor->data), ggml_nbytes(tensor)); + } else { + fin.seekg(ggml_nbytes(tensor), std::ios::cur); + } + + total_size += ggml_nbytes(tensor); + } else { + if ((nelements*bpe)/ggml_blck_size(tensor->type) != ggml_nbytes(tensor)/n_parts) { + fprintf(stderr, "%s: tensor '%s' has wrong size in model file: got %zu, expected %zu\n", + __func__, name.data(), ggml_nbytes(tensor)/n_parts, nelements*bpe); + return false; + } + + if (split_type == 0) { + const int np0 = ne[0]; + + const size_t row_size = (tensor->ne[0]/ggml_blck_size(tensor->type))*ggml_type_size(tensor->type); + assert(row_size == tensor->nb[1]); + + for (int i1 = 0; i1 < ne[1]; ++i1) { + const size_t offset_row = i1*row_size; + const size_t offset = offset_row + ((part_id*np0)/ggml_blck_size(tensor->type))*ggml_type_size(tensor->type); + fin.read(reinterpret_cast(tensor->data) + offset, row_size/n_parts); + } + } else { + const int np1 = ne[1]; + + const size_t row_size = (tensor->ne[0]/ggml_blck_size(tensor->type))*ggml_type_size(tensor->type); + + for (int i1 = 0; i1 < ne[1]; ++i1) { + const size_t offset_row = (i1 + part_id*np1)*row_size; + fin.read(reinterpret_cast(tensor->data) + offset_row, row_size); + } + } + + total_size += ggml_nbytes(tensor)/n_parts; + } + + //printf("%42s - [%5d, %5d], type = %6s, %6.2f MB\n", name.data(), ne[0], ne[1], ftype == 0 ? "float" : "f16", ggml_nbytes(tensor)/1024.0/1024.0); + if (++n_tensors % 8 == 0) { + printf("."); + fflush(stdout); + } + } + + //printf(" done\n"); + + //printf("%s: model size = %8.2f MB / num tensors = %d\n", __func__, total_size/1024.0/1024.0, n_tensors); + } + + fin.close(); + } + + return true; +} + +// evaluate the transformer +// +// - model: the model +// - n_threads: number of threads to use +// - n_past: the context size so far +// - embd_inp: the embeddings of the tokens in the context +// - embd_w: the predicted logits for the next token +// +// The GPT-J model requires about 16MB of memory per input token. +// +bool llama_eval( + const llama_model & model, + const int n_threads, + const int n_past, + const std::vector & embd_inp, + std::vector & embd_w, + size_t & mem_per_token) { + const int N = embd_inp.size(); + + const auto & hparams = model.hparams; + + const int n_embd = hparams.n_embd; + const int n_layer = hparams.n_layer; + const int n_ctx = hparams.n_ctx; + const int n_head = hparams.n_head; + const int n_vocab = hparams.n_vocab; + const int n_rot = hparams.n_embd/hparams.n_head; + + const int d_key = n_embd/n_head; + + static size_t buf_size = 512u*1024*1024; + static void * buf = malloc(buf_size); + + if (mem_per_token > 0 && mem_per_token*N > buf_size) { + const size_t buf_size_new = 1.1*(mem_per_token*N); // add 10% to account for ggml object overhead + //printf("\n%s: reallocating buffer from %zu to %zu bytes\n", __func__, buf_size, buf_size_new); + + // reallocate + buf_size = buf_size_new; + buf = realloc(buf, buf_size); + if (buf == nullptr) { + fprintf(stderr, "%s: failed to allocate %zu bytes\n", __func__, buf_size); + return false; + } + } + + struct ggml_init_params params = { + .mem_size = buf_size, + .mem_buffer = buf, + }; + + struct ggml_context * ctx0 = ggml_init(params); + struct ggml_cgraph gf = { .n_threads = n_threads }; + + struct ggml_tensor * embd = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, N); + memcpy(embd->data, embd_inp.data(), N*ggml_element_size(embd)); + + struct ggml_tensor * inpL = ggml_get_rows(ctx0, model.tok_embeddings, embd); + + for (int il = 0; il < n_layer; ++il) { + struct ggml_tensor * inpSA = inpL; + + struct ggml_tensor * cur; + + // norm + { + cur = ggml_norm(ctx0, inpL); + + // cur = attention_norm*cur + cur = ggml_mul(ctx0, + ggml_repeat(ctx0, model.layers[il].attention_norm, cur), + cur); + } + + // self-attention + { + struct ggml_tensor * Qcur = ggml_mul_mat(ctx0, model.layers[il].wq, cur); + struct ggml_tensor * Kcur = ggml_mul_mat(ctx0, model.layers[il].wk, cur); + struct ggml_tensor * Vcur = ggml_mul_mat(ctx0, model.layers[il].wv, cur); + + // store key and value to memory + if (N >= 1) { + struct ggml_tensor * k = ggml_view_1d(ctx0, model.memory_k, N*n_embd, (ggml_element_size(model.memory_k)*n_embd)*(il*n_ctx + n_past)); + struct ggml_tensor * v = ggml_view_1d(ctx0, model.memory_v, N*n_embd, (ggml_element_size(model.memory_v)*n_embd)*(il*n_ctx + n_past)); + + ggml_build_forward_expand(&gf, ggml_cpy(ctx0, Kcur, k)); + ggml_build_forward_expand(&gf, ggml_cpy(ctx0, Vcur, v)); + } + + // Q = Qcur.contiguous().view(n_embd/n_head, n_head, N).permute(0, 2, 1, 3) + struct ggml_tensor * Q = + ggml_permute(ctx0, + ggml_rope(ctx0, + ggml_cpy(ctx0, + Qcur, + ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_embd/n_head, n_head, N)), + n_past, n_rot, 0), + 0, 2, 1, 3); + + // K = Kmem.view(n_embd/n_head, n_head, n_past + N).permute(0, 2, 1, 3) + struct ggml_tensor * K = + ggml_permute(ctx0, + ggml_rope(ctx0, + ggml_reshape_3d(ctx0, + ggml_view_1d(ctx0, model.memory_k, (n_past + N)*n_embd, il*n_ctx*ggml_element_size(model.memory_k)*n_embd), + n_embd/n_head, n_head, n_past + N), + n_past, n_rot, 1), + 0, 2, 1, 3); + + // K * Q + struct ggml_tensor * KQ = ggml_mul_mat(ctx0, K, Q); + + // KQ_scaled = KQ / sqrt(n_embd/n_head) + struct ggml_tensor * KQ_scaled = + ggml_scale(ctx0, + KQ, + ggml_new_f32(ctx0, 1.0f/sqrt(float(n_embd)/n_head)) + ); + + // KQ_masked = mask_past(KQ_scaled) + struct ggml_tensor * KQ_masked = ggml_diag_mask_inf(ctx0, KQ_scaled, n_past); + + // KQ = soft_max(KQ_masked) + struct ggml_tensor * KQ_soft_max = ggml_soft_max(ctx0, KQ_masked); + + // V_trans = Vmem.view(n_embd/n_head, n_head, n_past + N).permute(1, 2, 0, 3).contiguous() + struct ggml_tensor * V_trans = + ggml_permute(ctx0, + ggml_reshape_3d(ctx0, + ggml_view_1d(ctx0, model.memory_v, (n_past + N)*n_embd, il*n_ctx*ggml_element_size(model.memory_v)*n_embd), + n_embd/n_head, n_head, n_past + N), + 1, 2, 0, 3); + + // KQV = transpose(V) * KQ_soft_max + struct ggml_tensor * KQV = ggml_mul_mat(ctx0, V_trans, KQ_soft_max); + + // KQV_merged = KQV.permute(0, 2, 1, 3) + struct ggml_tensor * KQV_merged = ggml_permute(ctx0, KQV, 0, 2, 1, 3); + + // cur = KQV_merged.contiguous().view(n_embd, N) + cur = ggml_cpy(ctx0, + KQV_merged, + ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, n_embd, N)); + + // projection (no bias) + cur = ggml_mul_mat(ctx0, + model.layers[il].wo, + cur); + } + + struct ggml_tensor * inpFF = ggml_add(ctx0, cur, inpSA); + + // feed-forward network + { + // norm + { + cur = ggml_norm(ctx0, inpFF); + + // cur = ffn_norm*cur + cur = ggml_mul(ctx0, + ggml_repeat(ctx0, model.layers[il].ffn_norm, cur), + cur); + } + + struct ggml_tensor * tmp = ggml_mul_mat(ctx0, + model.layers[il].w3, + cur); + + + cur = ggml_mul_mat(ctx0, + model.layers[il].w1, + cur); + + // SILU activation + cur = ggml_silu(ctx0, cur); + + cur = ggml_mul(ctx0, cur, tmp); + + cur = ggml_mul_mat(ctx0, + model.layers[il].w2, + cur); + } + + cur = ggml_add(ctx0, cur, inpFF); + + // input for next layer + inpL = cur; + } + + // norm + { + inpL = ggml_norm(ctx0, inpL); + + // inpL = norm*inpL + inpL = ggml_mul(ctx0, + ggml_repeat(ctx0, model.norm, inpL), + inpL); + } + + // lm_head + { + inpL = ggml_mul_mat(ctx0, model.output, inpL); + } + + // logits -> probs + //inpL = ggml_soft_max(ctx0, inpL); + + // run the computation + ggml_build_forward_expand(&gf, inpL); + ggml_graph_compute (ctx0, &gf); + + //if (n_past%100 == 0) { + // ggml_graph_print (&gf); + // ggml_graph_dump_dot(&gf, NULL, "gpt-2.dot"); + //} + + //embd_w.resize(n_vocab*N); + //memcpy(embd_w.data(), ggml_get_data(inpL), sizeof(float)*n_vocab*N); + + // return result for just the last token + embd_w.resize(n_vocab); + memcpy(embd_w.data(), (float *) ggml_get_data(inpL) + (n_vocab*(N-1)), sizeof(float)*n_vocab); + + if (mem_per_token == 0) { + mem_per_token = ggml_used_mem(ctx0)/N; + } + //printf("used_mem = %zu\n", ggml_used_mem(ctx0)); + + ggml_free(ctx0); + + return true; +} + +void llama_free(void *session) +{ + auto l = (struct llama *)session; + ggml_free(l->model.ctx); + free(l); +} + +void *llama_init(const char *model_path, int seed, int n_threads, int n_predict, int top_k, float top_p, int repeat_last_n, float repeat_penalty, float temp, int n_batch) +{ + struct llama *l = new struct llama(); + + const int64_t t_main_start_us = ggml_time_us(); + + l->params.model = model_path; + l->params.seed = seed; + l->params.n_threads = n_threads; + l->params.n_predict = n_predict; + l->params.top_k = top_k; + l->params.top_p = top_p; + l->params.temp = temp; + l->params.repeat_last_n = repeat_last_n; + l->params.repeat_penalty = repeat_penalty; + l->params.temp = temp; + l->params.n_batch = n_batch; + + if (l->params.seed < 0) { + l->params.seed = time(NULL); + } + + printf("%s: seed = %d\n", __func__, l->params.seed); + + l->rng = std::mt19937(l->params.seed); + if (l->params.prompt.empty()) { + l->params.prompt = gpt_random_prompt(l->rng); + } + +// l->params.prompt = R"(// this function checks if the number n is prime +//bool is_prime(int n) {)"; + + int64_t t_load_us = 0; + + // load the model + { + const int64_t t_start_us = ggml_time_us(); + + if (!llama_model_load(l->params.model, l->model, l->vocab, 512)) { // TODO: set context from user input ?? + fprintf(stderr, "%s: failed to load model from '%s'\n", __func__, l->params.model.c_str()); + llama_free(l); + return nullptr; + } + + t_load_us = ggml_time_us() - t_start_us; + } + + return l; +} + +bool llama_query(void *session, const char *prompt, char *buf, int buf_size) +{ + if (buf_size <= 0) { + return false; + } + + buf[0] = '\0'; + + auto l = (struct llama *)session; + int n_past = 0; + + int64_t t_sample_us = 0; + int64_t t_predict_us = 0; + + std::vector logits; + + // tokenize the prompt + std::vector embd_inp = ::llama_tokenize(l->vocab, prompt, true); + + l->params.n_predict = std::min(l->params.n_predict, l->model.hparams.n_ctx - (int) embd_inp.size()); + + //printf("\n"); + //printf("%s: prompt: '%s'\n", __func__, l->params.prompt.c_str()); + //printf("%s: number of tokens in prompt = %zu\n", __func__, embd_inp.size()); + //for (int i = 0; i < (int) embd_inp.size(); i++) { + // printf("%6d -> '%s'\n", embd_inp[i], l->vocab.id_to_token.at(embd_inp[i]).c_str()); + //} + //printf("\n"); + //printf("sampling parameters: temp = %f, top_k = %d, top_p = %f, repeat_last_n = %i, repeat_penalty = %f\n", l->params.temp, l->params.top_k, l->params.top_p, l->params.repeat_last_n, l->params.repeat_penalty); + //printf("\n\n"); + + std::vector embd; + + // determine the required inference memory per token: + size_t mem_per_token = 0; + llama_eval(l->model, l->params.n_threads, 0, { 0, 1, 2, 3 }, logits, mem_per_token); + + int last_n_size = l->params.repeat_last_n; + std::vector last_n_tokens(last_n_size); + std::fill(last_n_tokens.begin(), last_n_tokens.end(), 0); + + for (int i = embd.size(); i < embd_inp.size() + l->params.n_predict; i++) { + // predict + if (embd.size() > 0) { + const int64_t t_start_us = ggml_time_us(); + + if (!llama_eval(l->model, l->params.n_threads, n_past, embd, logits, mem_per_token)) { + //printf("Failed to predict\n"); + return false; + } + + t_predict_us += ggml_time_us() - t_start_us; + } + + n_past += embd.size(); + embd.clear(); + + if (i >= embd_inp.size()) { + // sample next token + const float top_p = l->params.top_p; + const float temp = l->params.temp; + const float repeat_penalty = l->params.repeat_penalty; + + const int n_vocab = l->model.hparams.n_vocab; + + gpt_vocab::id id = 0; + + { + const int64_t t_start_sample_us = ggml_time_us(); + + id = llama_sample_top_p(l->vocab, logits.data() + (logits.size() - n_vocab), last_n_tokens, repeat_penalty, top_p, temp, l->rng); + + last_n_tokens.erase(last_n_tokens.begin()); + last_n_tokens.push_back(id); + + t_sample_us += ggml_time_us() - t_start_sample_us; + } + + // add it to the context + embd.push_back(id); + } else { + // if here, it means we are still processing the input prompt + for (int k = i; k < embd_inp.size(); k++) { + embd.push_back(embd_inp[k]); + last_n_tokens.erase(last_n_tokens.begin()); + last_n_tokens.push_back(embd_inp[k]); + if (embd.size() > l->params.n_batch) { + break; + } + } + i += embd.size() - 1; + } + + // display text + for (auto id : embd) { + strncat(buf, l->vocab.id_to_token[id].c_str(), buf_size); + } + + // end of text token + if (embd.back() == 2) { + strncat(buf, " [end of text]", buf_size); + break; + } + } + + // report timing + //{ + // const int64_t t_main_end_us = ggml_time_us(); + + // printf("\n\n"); + // printf("%s: mem per token = %8zu bytes\n", __func__, mem_per_token); + // printf("%s: load time = %8.2f ms\n", __func__, t_load_us/1000.0f); + // printf("%s: sample time = %8.2f ms\n", __func__, t_sample_us/1000.0f); + // printf("%s: predict time = %8.2f ms / %.2f ms per token\n", __func__, t_predict_us/1000.0f, t_predict_us/1000.0f/n_past); + // printf("%s: total time = %8.2f ms\n", __func__, (t_main_end_us - t_main_start_us)/1000.0f); + //} + + return true; +} diff --git a/libllama.h b/libllama.h new file mode 100644 index 0000000000000..98fd21a46b394 --- /dev/null +++ b/libllama.h @@ -0,0 +1,10 @@ +#pragma once +#ifdef __cplusplus +extern "C" { +#endif +__attribute__((visibility("default"))) void *llama_init(const char *model_path, int seed, int n_threads, int n_predict, int top_k, float top_p, int repeat_last_n, float repeat_penalty, float temp, int n_batch); +__attribute__((visibility("default"))) bool llama_query(void *session, const char *prompt, char *buf, int buf_size); +__attribute__((visibility("default"))) void llama_free(void *l); +#ifdef __cplusplus +} +#endif diff --git a/main.cpp b/main.cpp index f02b5ddbde94d..4caac566b2a6b 100644 --- a/main.cpp +++ b/main.cpp @@ -871,16 +871,16 @@ int main(int argc, char ** argv) { } // report timing - { - const int64_t t_main_end_us = ggml_time_us(); - - printf("\n\n"); - printf("%s: mem per token = %8zu bytes\n", __func__, mem_per_token); - printf("%s: load time = %8.2f ms\n", __func__, t_load_us/1000.0f); - printf("%s: sample time = %8.2f ms\n", __func__, t_sample_us/1000.0f); - printf("%s: predict time = %8.2f ms / %.2f ms per token\n", __func__, t_predict_us/1000.0f, t_predict_us/1000.0f/n_past); - printf("%s: total time = %8.2f ms\n", __func__, (t_main_end_us - t_main_start_us)/1000.0f); - } + //{ + // const int64_t t_main_end_us = ggml_time_us(); + + // printf("\n\n"); + // printf("%s: mem per token = %8zu bytes\n", __func__, mem_per_token); + // printf("%s: load time = %8.2f ms\n", __func__, t_load_us/1000.0f); + // printf("%s: sample time = %8.2f ms\n", __func__, t_sample_us/1000.0f); + // printf("%s: predict time = %8.2f ms / %.2f ms per token\n", __func__, t_predict_us/1000.0f, t_predict_us/1000.0f/n_past); + // printf("%s: total time = %8.2f ms\n", __func__, (t_main_end_us - t_main_start_us)/1000.0f); + //} ggml_free(model.ctx);