mirrors/cosmopolitan
1
0
Fork 0
mirror of https://github.com/jart/cosmopolitan.git synced 2025-09-10 10:43:48 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

Upgrade llama.cpp to e6a46b0ed1884c77267dc70693183e3b7164e0e0

Browse source
This commit is contained in:
Justine Tunney 2023-05-10 03:18:38 -07:00
parent 5a455eaa0b
commit 5f57fc1f59
No known key found for this signature in database
GPG key ID: BE714B4575D6E328
8 changed files with 2001 additions and 820 deletions
Download patch file Download diff file Expand all files Collapse all files

56
third_party/ggml/ggml.h vendored
View file

@ -1,4 +1,3 @@
// clang-format off
#ifndef COSMOPOLITAN_THIRD_PARTY_LLAMA_CPP_GGML_H_
#define COSMOPOLITAN_THIRD_PARTY_LLAMA_CPP_GGML_H_
#if !(__ASSEMBLER__ + __LINKER__ + 0)
@ -198,6 +197,14 @@ COSMOPOLITAN_C_START_
#define GGML_MAX_OPT 4
#define GGML_DEFAULT_N_THREADS 4
#define GGML_ASSERT(x) \
do { \
if (!(x)) { \
fprintf(stderr, "GGML_ASSERT: %s:%d: %s\n", __FILE__, __LINE__, #x); \
abort(); \
} \
} while (0)
#ifdef __ARM_NEON
// we use the built-in 16-bit float type
typedef __fp16 ggml_fp16_t;
@ -209,6 +216,9 @@ COSMOPOLITAN_C_START_
GGML_API float ggml_fp16_to_fp32(ggml_fp16_t x);
GGML_API ggml_fp16_t ggml_fp32_to_fp16(float x);
GGML_API void ggml_fp16_to_fp32_row(const ggml_fp16_t * x, float * y, size_t n);
GGML_API void ggml_fp32_to_fp16_row(const float * x, ggml_fp16_t * y, size_t n);
struct ggml_object;
struct ggml_context;
@ -218,7 +228,7 @@ COSMOPOLITAN_C_START_
GGML_TYPE_Q4_0 = 2,
GGML_TYPE_Q4_1 = 3,
GGML_TYPE_Q4_2 = 4,
GGML_TYPE_Q4_3 = 5,
// GGML_TYPE_Q4_3 (5) support has been removed
GGML_TYPE_Q5_0 = 6,
GGML_TYPE_Q5_1 = 7,
GGML_TYPE_Q8_0 = 8,
@ -229,6 +239,20 @@ COSMOPOLITAN_C_START_
GGML_TYPE_COUNT,
};
// model file types
enum ggml_ftype {
GGML_FTYPE_UNKNOWN = -1,
GGML_FTYPE_ALL_F32 = 0,
GGML_FTYPE_MOSTLY_F16 = 1, // except 1d tensors
GGML_FTYPE_MOSTLY_Q4_0 = 2, // except 1d tensors
GGML_FTYPE_MOSTLY_Q4_1 = 3, // except 1d tensors
GGML_FTYPE_MOSTLY_Q4_1_SOME_F16 = 4, // tok_embeddings.weight and output.weight are F16
GGML_FTYPE_MOSTLY_Q4_2 = 5, // except 1d tensors
GGML_FTYPE_MOSTLY_Q8_0 = 7, // except 1d tensors
GGML_FTYPE_MOSTLY_Q5_0 = 8, // except 1d tensors
GGML_FTYPE_MOSTLY_Q5_1 = 9, // except 1d tensors
};
// available tensor operations:
enum ggml_op {
GGML_OP_NONE = 0,
@ -266,6 +290,7 @@ COSMOPOLITAN_C_START_
GGML_OP_DIAG_MASK_INF,
GGML_OP_SOFT_MAX,
GGML_OP_ROPE,
GGML_OP_ALIBI,
GGML_OP_CONV_1D_1S,
GGML_OP_CONV_1D_2S,
@ -321,7 +346,10 @@ COSMOPOLITAN_C_START_
int64_t perf_time_us;
void * data;
char padding[8];
char name[32];
char padding[8]; // TODO: remove and add padding to name?
};
// computation graph
@ -381,6 +409,9 @@ COSMOPOLITAN_C_START_
GGML_API bool ggml_is_quantized(enum ggml_type type);
// TODO: temporary until model loading of ggml examples is refactored
GGML_API enum ggml_type ggml_ftype_to_ggml_type(enum ggml_ftype ftype);
// main
GGML_API struct ggml_context * ggml_init(struct ggml_init_params params);
@ -441,6 +472,9 @@ COSMOPOLITAN_C_START_
GGML_API void * ggml_get_data (const struct ggml_tensor * tensor);
GGML_API float * ggml_get_data_f32(const struct ggml_tensor * tensor);
GGML_API const char * ggml_get_name(const struct ggml_tensor * tensor);
GGML_API void ggml_set_name(struct ggml_tensor * tensor, const char * name);
//
// operations on tensors with backpropagation
//
@ -659,6 +693,14 @@ COSMOPOLITAN_C_START_
int n_dims,
int mode);
// alibi position embedding
// in-place, returns view(a)
struct ggml_tensor * ggml_alibi(
struct ggml_context * ctx,
struct ggml_tensor * a,
int n_past,
int n_head);
// padding = 1
// TODO: we don't support extra parameters for now
// that's why we are hard-coding the stride, padding, and dilation
@ -689,8 +731,8 @@ COSMOPOLITAN_C_START_
struct ggml_tensor * c1);
// Mapping operations
GGML_API typedef void (*ggml_unary_op_f32_t)(const int, float *, const float *);
GGML_API typedef void (*ggml_binary_op_f32_t)(const int, float *, const float *, const float *);
typedef void (*ggml_unary_op_f32_t)(const int, float *, const float *);
typedef void (*ggml_binary_op_f32_t)(const int, float *, const float *, const float *);
GGML_API struct ggml_tensor * ggml_map_unary_f32(
struct ggml_context * ctx,
@ -831,7 +873,6 @@ COSMOPOLITAN_C_START_
GGML_API size_t ggml_quantize_q4_0(const float * src, void * dst, int n, int k, int64_t * hist);
GGML_API size_t ggml_quantize_q4_1(const float * src, void * dst, int n, int k, int64_t * hist);
GGML_API size_t ggml_quantize_q4_2(const float * src, void * dst, int n, int k, int64_t * hist);
GGML_API size_t ggml_quantize_q4_3(const float * src, void * dst, int n, int k, int64_t * hist);
GGML_API size_t ggml_quantize_q5_0(const float * src, void * dst, int n, int k, int64_t * hist);
GGML_API size_t ggml_quantize_q5_1(const float * src, void * dst, int n, int k, int64_t * hist);
GGML_API size_t ggml_quantize_q8_0(const float * src, void * dst, int n, int k, int64_t * hist);
@ -855,10 +896,11 @@ COSMOPOLITAN_C_START_
GGML_API int ggml_cpu_has_wasm_simd (void);
GGML_API int ggml_cpu_has_blas (void);
GGML_API int ggml_cpu_has_cublas (void);
GGML_API int ggml_cpu_has_clblast (void);
GGML_API int ggml_cpu_has_gpublas (void);
GGML_API int ggml_cpu_has_sse3 (void);
GGML_API int ggml_cpu_has_vsx (void);
//
// Internal types and functions exposed for tests and benchmarks
//