llama : refactor model loading code (#2620)

* llama : style formatting + remove helper methods

* llama : fix quantization using gguf tool

* llama : simplify gguf_file_saver

* llama : fix method names

* llama : simplify write_header()

* llama : no need to pass full file loader to the file saver

just gguf_ctx

* llama : gguf_file_saver write I32

* llama : refactor tensor names (#2622)

* gguf: update tensor names searched in quantization

* gguf : define tensor names as constants

* gguf : initial write API (not tested yet)

* gguf : write to file API (not tested)

* gguf : initial write API ready + example

* gguf : fix header write

* gguf : fixes + simplify example + add ggml_nbytes_pad()

* gguf : minor

* llama : replace gguf_file_saver with new gguf write API

* gguf : streaming support when writing files

* gguf : remove oboslete write methods

* gguf : remove obosolete gguf_get_arr_xxx API

* llama : simplify gguf_file_loader

* llama : move hparams and vocab from gguf_file_loader to llama_model_loader

* llama : merge gguf-util.h in llama.cpp

* llama : reorder definitions in .cpp to match .h

* llama : minor simplifications

* llama : refactor llama_model_loader (WIP)

wip : remove ggml_ctx from llama_model_loader

wip : merge gguf_file_loader in llama_model_loader

* llama : fix shape prints

* llama : fix Windows build + fix norm_rms_eps key

* llama : throw error on missing KV paris in model meta data

* llama : improve printing + log meta data

* llama : switch print order of meta data

---------

Co-authored-by: M. Yusuf Sarıgöz <[email protected]>

Author: ggerganov

Makefile

@@ -332,7 +332,7 @@ OBJS += ggml-alloc.o
 llama.o: llama.cpp ggml.h ggml-alloc.h ggml-cuda.h ggml-metal.h llama.h llama-util.h
 	$(CXX) $(CXXFLAGS) -c $< -o $@
 
-gguf-llama.o: gguf-llama.cpp ggml.h ggml-alloc.h ggml-cuda.h ggml-metal.h gguf-llama.h gguf-util.h
+gguf-llama.o: gguf-llama.cpp ggml.h ggml-alloc.h ggml-cuda.h ggml-metal.h gguf-llama.h
 	$(CXX) $(CXXFLAGS) -c $< -o $@
 
 common.o: examples/common.cpp examples/common.h

convert-llama-h5-to-gguf.py

@@ -135,7 +135,7 @@ def count_model_parts(dir_model: str) -> int:
         toktype = 1 # defualt to normal token type
         if tokenizer.is_unknown(i): toktype = 2
         if tokenizer.is_control(i): toktype = 3
- 
+
         # TODO: How to determinate if a token is user defined?
         # ref: https://github.com/google/sentencepiece/blob/master/src/sentencepiece_model.proto
         # if tokenizer.is_user_defined(i): toktype = 4

examples/gguf/gguf-llama-simple.cpp

@@ -74,7 +74,9 @@ int main(int argc, char ** argv) {
     // tokens (see "infinite text generation via context swapping" in the main example), but in this minimalist
     // example, we will just stop the loop once this cache is full or once an end of stream is detected.
 
-    while (llama_get_kv_cache_token_count(ctx) < max_context_size) {
+    const int n_gen = std::min(32, max_context_size);
+
+    while (llama_get_kv_cache_token_count(ctx) < n_gen) {
         // evaluate the transformer
 
         if (llama_eval(ctx, tokens_list.data(), int(tokens_list.size()), llama_get_kv_cache_token_count(ctx), params.n_threads)) {
@@ -114,13 +116,14 @@ int main(int argc, char ** argv) {
 
         // push this new token for next evaluation
         tokens_list.push_back(new_token_id);
-
     }
 
     llama_free(ctx);
     llama_free_model(model);
 
     llama_backend_free();
 
+    fprintf(stderr, "\n\n");
+
     return 0;
 }

examples/gguf/gguf.cpp

@@ -1,5 +1,4 @@
 #include "ggml.h"
-#include "gguf-util.h"
 #include "gguf-llama.h"
 
 #include <cstdio>
@@ -21,133 +20,22 @@ static std::string to_string(const T & val) {
     return ss.str();
 }
 
-void gguf_ex_write_str(std::ofstream & fout, const std::string & val) {
-    const int32_t n = val.size();
-    fout.write((const char *) &n, sizeof(n));
-    fout.write(val.c_str(), n);
-}
-
-void gguf_ex_write_i32(std::ofstream & fout, int32_t val) {
-    fout.write((const char *) &val, sizeof(val));
-}
-
-void gguf_ex_write_u64(std::ofstream & fout, size_t val) {
-    fout.write((const char *) &val, sizeof(val));
-}
-
-template<typename T>
-void gguf_ex_write_val(std::ofstream & fout, const std::string & key, enum gguf_type type, const T & val) {
-    gguf_ex_write_str(fout, key);
-    fout.write((const char *) &type, sizeof(type));
-    fout.write((const char *) &val,  sizeof(val));
-
-    fprintf(stdout, "%s: write param: %s = %s\n", __func__, key.c_str(), to_string(val).c_str());
-}
-
-template<>
-void gguf_ex_write_val<std::string>(std::ofstream & fout, const std::string & key, enum gguf_type type, const std::string & val) {
-    gguf_ex_write_str(fout, key);
-    fout.write((const char *) &type, sizeof(type));
-
-    const int32_t n = val.size();
-    fout.write((const char *) &n, sizeof(n));
-    fout.write(val.c_str(), n);
-
-    fprintf(stdout, "%s: write param: %s = %s\n", __func__, key.c_str(), val.c_str());
-}
-
-template<typename T>
-void gguf_ex_write_arr(std::ofstream & fout, const std::string & key, enum gguf_type type, const std::vector<T> & val) {
-    gguf_ex_write_str(fout, key);
-    {
-        const enum gguf_type tarr = GGUF_TYPE_ARRAY;
-        fout.write((const char *) &tarr, sizeof(tarr));
-    }
-
-    const int32_t n = val.size();
-    fout.write((const char *) &type, sizeof(type));
-    fout.write((const char *) &n,    sizeof(n));
-    fout.write((const char *) val.data(), n * sizeof(T));
-
-    fprintf(stdout, "%s: write param: %s = [", __func__, key.c_str());
-    for (int i = 0; i < n; ++i) {
-        fprintf(stdout, "%s", to_string(val[i]).c_str());
-        if (i < n - 1) {
-            fprintf(stdout, ", ");
-        }
-    }
-    fprintf(stdout, "]\n");
-}
-
-template<>
-void gguf_ex_write_arr<std::string>(std::ofstream & fout, const std::string & key, enum gguf_type type, const std::vector<std::string> & val) {
-    gguf_ex_write_str(fout, key);
-    {
-        const enum gguf_type tarr = GGUF_TYPE_ARRAY;
-        fout.write((const char *) &tarr, sizeof(tarr));
-    }
-
-    const int32_t n = val.size();
-    fout.write((const char *) &type, sizeof(type));
-    fout.write((const char *) &n,    sizeof(n));
-    for (int i = 0; i < n; ++i) {
-        const int32_t nstr = val[i].size();
-        fout.write((const char *) &nstr, sizeof(nstr));
-        fout.write(val[i].c_str(), nstr);
-    }
-
-    fprintf(stdout, "%s: write param: %s = [", __func__, key.c_str());
-    for (int i = 0; i < n; ++i) {
-        fprintf(stdout, "%s", val[i].c_str());
-        if (i < n - 1) {
-            fprintf(stdout, ", ");
-        }
-    }
-    fprintf(stdout, "]\n");
-}
-
 bool gguf_ex_write(const std::string & fname) {
-    std::ofstream fout(fname.c_str(), std::ios::binary);
-
-    {
-        const int32_t magic = GGUF_MAGIC;
-        fout.write((const char *) &magic, sizeof(magic));
-    }
-
-    {
-        const int32_t version = GGUF_VERSION;
-        fout.write((const char *) &version, sizeof(version));
-    }
-
-    // NOTE: these have to match the output below!
-    const int n_tensors = 10;
-    const int n_kv      = 12;
-
-    fout.write((const char*) &n_tensors, sizeof(n_tensors));
-    fout.write((const char*) &n_kv, sizeof(n_kv));
-
-    fprintf(stdout, "%s: write header\n", __func__);
-
-    // kv data
-    {
-        gguf_ex_write_val< uint8_t>(fout, "some.parameter.uint8",   GGUF_TYPE_UINT8,   0x12);
-        gguf_ex_write_val<  int8_t>(fout, "some.parameter.int8",    GGUF_TYPE_INT8,   -0x13);
-        gguf_ex_write_val<uint16_t>(fout, "some.parameter.uint16",  GGUF_TYPE_UINT16,  0x1234);
-        gguf_ex_write_val< int16_t>(fout, "some.parameter.int16",   GGUF_TYPE_INT16,  -0x1235);
-        gguf_ex_write_val<uint32_t>(fout, "some.parameter.uint32",  GGUF_TYPE_UINT32,  0x12345678);
-        gguf_ex_write_val< int32_t>(fout, "some.parameter.int32",   GGUF_TYPE_INT32,  -0x12345679);
-
-        gguf_ex_write_val<float>   (fout, "some.parameter.float32", GGUF_TYPE_FLOAT32, 0.123456789f);
-        gguf_ex_write_val<bool>    (fout, "some.parameter.bool",    GGUF_TYPE_BOOL,    true);
-
-        gguf_ex_write_val<std::string>(fout, "some.parameter.string",  GGUF_TYPE_STRING,  "hello world");
-
-        gguf_ex_write_arr<int16_t>    (fout, "some.parameter.arr.i16", GGUF_TYPE_INT16,   { 1, 2, 3, 4, });
-        gguf_ex_write_arr<float>      (fout, "some.parameter.arr.f32", GGUF_TYPE_FLOAT32, { 3.145f, 2.718f, 1.414f, });
-        gguf_ex_write_arr<std::string>(fout, "some.parameter.arr.str", GGUF_TYPE_STRING,  { "hello", "world", "!" });
-    }
-
-    uint64_t offset_tensor = 0;
+    struct gguf_context * ctx = gguf_init_empty();
+
+    gguf_set_val_u8  (ctx, "some.parameter.uint8",    0x12);
+    gguf_set_val_i8  (ctx, "some.parameter.int8",    -0x13);
+    gguf_set_val_u16 (ctx, "some.parameter.uint16",   0x1234);
+    gguf_set_val_i16 (ctx, "some.parameter.int16",   -0x1235);
+    gguf_set_val_u32 (ctx, "some.parameter.uint32",   0x12345678);
+    gguf_set_val_i32 (ctx, "some.parameter.int32",   -0x12345679);
+    gguf_set_val_f32 (ctx, "some.parameter.float32",  0.123456789f);
+    gguf_set_val_bool(ctx, "some.parameter.bool",     true);
+    gguf_set_val_str (ctx, "some.parameter.string",   "hello world");
+
+    gguf_set_arr_data(ctx, "some.parameter.arr.i16", GGUF_TYPE_INT16,   std::vector<int16_t>{ 1, 2, 3, 4, }.data(), 4);
+    gguf_set_arr_data(ctx, "some.parameter.arr.f32", GGUF_TYPE_FLOAT32, std::vector<float>{ 3.145f, 2.718f, 1.414f, }.data(), 3);
+    gguf_set_arr_str (ctx, "some.parameter.arr.str",                    std::vector<const char *>{ "hello", "world", "!" }.data(), 3);
 
     struct ggml_init_params params = {
         /*.mem_size   =*/ 128ull*1024ull*1024ull,
@@ -157,6 +45,8 @@ bool gguf_ex_write(const std::string & fname) {
 
     struct ggml_context * ctx_data = ggml_init(params);
 
+    const int n_tensors = 10;
+
     // tensor infos
     for (int i = 0; i < n_tensors; ++i) {
         const std::string name = "tensor_" + to_string(i);
@@ -178,58 +68,15 @@ bool gguf_ex_write(const std::string & fname) {
             }
         }
 
-        fprintf(stdout, "%s: tensor: %s, %d dims, ne = [", __func__, name.c_str(), n_dims);
-        for (int j = 0; j < 4; ++j) {
-            fprintf(stdout, "%s%3d", j == 0 ? "" : ", ", (int) cur->ne[j]);
-        }
-        fprintf(stdout, "], offset_tensor = %6" PRIu64 "\n", offset_tensor);
-
-        gguf_ex_write_str(fout, name);
-        gguf_ex_write_i32(fout, n_dims);
-        for (int j = 0; j < n_dims; ++j) {
-            gguf_ex_write_i32(fout, cur->ne[j]);
-        }
-        gguf_ex_write_i32(fout, cur->type);
-        gguf_ex_write_u64(fout, offset_tensor);
-
-        offset_tensor += GGML_PAD(ggml_nbytes(cur), GGUF_DEFAULT_ALIGNMENT);
-    }
-
-    const uint64_t offset_data = GGML_PAD((uint64_t) fout.tellp(), GGUF_DEFAULT_ALIGNMENT);
-
-    fprintf(stdout, "%s: data offset = %" PRIu64 "\n", __func__, offset_data);
-
-    {
-        const size_t pad = offset_data - fout.tellp();
-
-        for (size_t j = 0; j < pad; ++j) {
-            fout.put(0);
-        }
-    }
-
-    for (int i = 0; i < n_tensors; ++i) {
-        fprintf(stdout, "%s: writing tensor %d data\n", __func__, i);
-
-        const std::string name = "tensor_" + to_string(i);
-
-        struct ggml_tensor * cur = ggml_get_tensor(ctx_data, name.c_str());
-
-        fout.write((const char *) cur->data, ggml_nbytes(cur));
-
-        {
-            const size_t pad = GGML_PAD(ggml_nbytes(cur), GGUF_DEFAULT_ALIGNMENT) - ggml_nbytes(cur);
-
-            for (size_t j = 0; j < pad; ++j) {
-                fout.put(0);
-            }
-        }
+        gguf_add_tensor(ctx, cur);
     }
 
-    fout.close();
+    gguf_write_to_file(ctx, fname.c_str(), false);
 
     fprintf(stdout, "%s: wrote file '%s;\n", __func__, fname.c_str());
 
     ggml_free(ctx_data);
+    gguf_free(ctx);
 
     return true;
 }
@@ -345,8 +192,16 @@ bool gguf_ex_read_1(const std::string & fname) {
 
             struct ggml_tensor * cur = ggml_get_tensor(ctx_data, name);
 
-            fprintf(stdout, "%s: tensor[%d]: n_dims = %d, name = %s, data = %p\n",
-                    __func__, i, cur->n_dims, cur->name, cur->data);
+            fprintf(stdout, "%s: tensor[%d]: n_dims = %d, name = %s, data = %p\n", __func__, i, cur->n_dims, cur->name, cur->data);
+
+            // print first 10 elements
+            const float * data = (const float *) cur->data;
+
+            printf("%s data[:10] : ", name);
+            for (int j = 0; j < MIN(10, ggml_nelements(cur)); ++j) {
+                printf("%f ", data[j]);
+            }
+            printf("\n\n");
 
             // check data
             {
@@ -369,48 +224,6 @@ bool gguf_ex_read_1(const std::string & fname) {
     return true;
 }
 
-// read just the tensor info and mmap the data in user code
-bool gguf_ex_read_2(const std::string & fname) {
-    struct ggml_context * ctx_data = NULL;
-
-    struct gguf_init_params params = {
-        /*.no_alloc = */ true,
-        /*.ctx      = */ &ctx_data,
-    };
-
-    struct gguf_context * ctx = gguf_init_from_file(fname.c_str(), params);
-
-    struct gguf_file file(fname.c_str(), "rb");
-    gguf_mmap data_mmap(&file, 0, false);
-
-    const int n_tensors = gguf_get_n_tensors(ctx);
-
-    for (int i = 0; i < n_tensors; ++i) {
-        const char * name   = gguf_get_tensor_name(ctx, i);
-        const size_t offset = gguf_get_data_offset(ctx) + gguf_get_tensor_offset(ctx, i);
-
-        struct ggml_tensor * cur = ggml_get_tensor(ctx_data, name);
-
-        cur->data = static_cast<char *>(data_mmap.addr) + offset;
-
-        // print first 10 elements
-        const float * data = (const float *) cur->data;
-
-        printf("%s data[:10] : ", name);
-        for (int j = 0; j < MIN(10, ggml_nelements(cur)); ++j) {
-            printf("%f ", data[j]);
-        }
-        printf("\n\n");
-    }
-
-    fprintf(stdout, "%s: ctx_data size: %zu\n", __func__, ggml_get_mem_size(ctx_data));
-
-    ggml_free(ctx_data);
-    gguf_free(ctx);
-
-    return true;
-}
-
 int main(int argc, char ** argv) {
     if (argc < 3) {
         fprintf(stdout, "usage: %s data.gguf r|w\n", argv[0]);
@@ -427,7 +240,6 @@ int main(int argc, char ** argv) {
     } else if (mode == "r") {
         GGML_ASSERT(gguf_ex_read_0(fname) && "failed to read gguf file");
         GGML_ASSERT(gguf_ex_read_1(fname) && "failed to read gguf file");
-        GGML_ASSERT(gguf_ex_read_2(fname) && "failed to read gguf file");
     } else if (mode == "q") {
         llama_model_quantize_params params = llama_model_quantize_default_params();
         llama_model_quantize(fname.c_str(), "quant.gguf", &params);