vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-11 17:44:38 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 3

Merge branch 'master' into concedo_experimental

Browse source 
# Conflicts:
#	.devops/nix/package.nix
#	.github/workflows/docker.yml
#	CMakeLists.txt
This commit is contained in:
Concedo 2024-01-25 00:17:07 +08:00
commit 0a70cc1ba7
11 changed files with 357 additions and 283 deletions
Download patch file Download diff file Expand all files Collapse all files

26
.devops/main-intel.Dockerfile Normal file
View file

@ -0,0 +1,26 @@
ARG ONEAPI_VERSION=2024.0.1-devel-ubuntu22.04
ARG UBUNTU_VERSION=22.04
FROM intel/hpckit:$ONEAPI_VERSION as build
RUN apt-get update && \
apt-get install -y git
WORKDIR /app
COPY . .
# for some reasons, "-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=Intel10_64lp -DLLAMA_NATIVE=ON" give worse performance
RUN mkdir build && \
cd build && \
cmake .. -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx && \
cmake --build . --config Release --target main server
FROM ubuntu:$UBUNTU_VERSION as runtime
COPY --from=build /app/build/bin/main /main
COPY --from=build /app/build/bin/server /server
ENV LC_ALL=C.utf8
ENTRYPOINT [ "/main" ]

10
common/common.cpp
View file

@ -217,12 +217,10 @@ bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params) {
} }
// store the external file name in params // store the external file name in params
params.prompt_file = argv[i]; params.prompt_file = argv[i];
file.seekg(0, std::ios::end); std::ostringstream ss;
size_t size = file.tellg(); ss << file.rdbuf();
file.seekg(0, std::ios::beg); params.prompt = ss.str();
params.prompt.resize(size); fprintf(stderr, "Read %zu bytes from binary file %s\n", params.prompt.size(), argv[i]);
file.read((char *)params.prompt.data(), size);
fprintf(stderr, "Read %zu bytes from binary file %s\n", size, argv[i]);
} else if (arg == "-f" || arg == "--file") { } else if (arg == "-f" || arg == "--file") {
if (++i >= argc) { if (++i >= argc) {
invalid_param = true; invalid_param = true;

5
examples/llama.vim
View file

@ -6,7 +6,7 @@
" Similarly, you could add an insert mode keybind with " Similarly, you could add an insert mode keybind with
" inoremap <C-B> <Cmd>call llama#doLlamaGen()<CR> " inoremap <C-B> <Cmd>call llama#doLlamaGen()<CR>
" "
" g:llama_api_url and g:llama_overrides can be configured in your .vimrc " g:llama_api_url, g:llama_api_key and g:llama_overrides can be configured in your .vimrc
" let g:llama_api_url = "192.168.1.10:8080" " let g:llama_api_url = "192.168.1.10:8080"
" llama_overrides can also be set through buffer/window scopes. For instance " llama_overrides can also be set through buffer/window scopes. For instance
" autocmd filetype python let b:llama_overrides = {"temp": 0.2} " autocmd filetype python let b:llama_overrides = {"temp": 0.2}
@ -82,6 +82,9 @@ func llama#doLlamaGen()
endif endif
let l:querydata.prompt = join(l:buflines, "\n") let l:querydata.prompt = join(l:buflines, "\n")
let l:curlcommand = copy(s:curlcommand) let l:curlcommand = copy(s:curlcommand)
if exists("g:llama_api_key")
call extend(l:curlcommand, ['--header', 'Authorization: Bearer ' .. g:llama_api_key])
endif
let l:curlcommand[2] = json_encode(l:querydata) let l:curlcommand[2] = json_encode(l:querydata)
let b:job = job_start(l:curlcommand, {"callback": function("s:callbackHandler", [l:cbuffer])}) let b:job = job_start(l:curlcommand, {"callback": function("s:callbackHandler", [l:cbuffer])})
endfunction endfunction

47
examples/llava/clip.cpp
View file

@ -2,18 +2,6 @@
// so there might be still unnecessary artifacts hanging around // so there might be still unnecessary artifacts hanging around
// I'll gradually clean and extend it // I'll gradually clean and extend it
#include <cassert>
#include <cmath>
#include <cstdlib>
#include <cstring>
#include <fstream>
#include <iostream>
#include <map>
#include <regex>
#include <stdexcept>
#include <vector>
#include <sstream>
#include "clip.h" #include "clip.h"
#include "ggml.h" #include "ggml.h"
#include "ggml-alloc.h" #include "ggml-alloc.h"
@ -30,6 +18,19 @@
#define STB_IMAGE_IMPLEMENTATION #define STB_IMAGE_IMPLEMENTATION
#include "stb_image.h" #include "stb_image.h"
#include <cassert>
#include <cmath>
#include <cstdlib>
#include <cstring>
#include <fstream>
#include <iostream>
#include <map>
#include <regex>
#include <stdexcept>
#include <vector>
#include <sstream>
#include <cinttypes>
static std::string format(const char * fmt, ...) { static std::string format(const char * fmt, ...) {
va_list ap; va_list ap;
va_list ap2; va_list ap2;
@ -217,9 +218,9 @@ static std::string gguf_kv_to_str(const struct gguf_context * ctx_gguf, int i) {
static void print_tensor_info(const ggml_tensor* tensor, const char* prefix = "") { static void print_tensor_info(const ggml_tensor* tensor, const char* prefix = "") {
size_t tensor_size = ggml_nbytes(tensor); size_t tensor_size = ggml_nbytes(tensor);
printf("%s: n_dims = %d, name = %s, tensor_size=%zu, shape:[%d, %d, %d, %d], type: %d\n", printf("%s: n_dims = %d, name = %s, tensor_size=%zu, shape:[%" PRId64 ", %" PRId64 ", %" PRId64 ", %" PRId64 "], type = %s\n",
prefix, ggml_n_dims(tensor), tensor->name, tensor_size, prefix, ggml_n_dims(tensor), tensor->name, tensor_size,
tensor->ne[0], tensor->ne[1], tensor->ne[2], tensor->ne[3], tensor->type); tensor->ne[0], tensor->ne[1], tensor->ne[2], tensor->ne[3], ggml_type_name(tensor->type));
} }
static projector_type clip_projector_type_from_string(const std::string & name) { static projector_type clip_projector_type_from_string(const std::string & name) {
@ -592,7 +593,7 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32
mlp_3 = ggml_cont(ctx0, ggml_permute(ctx0, mlp_3, 1, 0, 2, 3)); mlp_3 = ggml_cont(ctx0, ggml_permute(ctx0, mlp_3, 1, 0, 2, 3));
mlp_3 = ggml_reshape_4d(ctx0, mlp_3, n_patch, n_patch, mlp_3->ne[1], mlp_3->ne[2]); mlp_3 = ggml_reshape_4d(ctx0, mlp_3, n_patch, n_patch, mlp_3->ne[1], mlp_3->ne[2]);
// stride = 1, padding = 1, bias is nullptr // stride = 1, padding = 1, bias is nullptr
block_1 = ggml_conv_depthwise_2d(ctx0, model.mm_model_block_1_block_0_0_w, mlp_3, nullptr, 1, 1, 1, 1, 1, 1); block_1 = ggml_conv_depthwise_2d(ctx0, model.mm_model_block_1_block_0_0_w, mlp_3, 1, 1, 1, 1, 1, 1);
// layer norm // layer norm
// // block_1 shape = [1, 2048, 24, 24], ne = [24, 24, 2048, 1] // // block_1 shape = [1, 2048, 24, 24], ne = [24, 24, 2048, 1]
@ -640,7 +641,7 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32
// block_2 // block_2
{ {
// stride = 2 // stride = 2
block_1 = ggml_conv_depthwise_2d(ctx0, model.mm_model_block_2_block_0_0_w, block_1, nullptr, 2, 2, 1, 1, 1, 1); block_1 = ggml_conv_depthwise_2d(ctx0, model.mm_model_block_2_block_0_0_w, block_1, 2, 2, 1, 1, 1, 1);
// block_1 shape = [1, 2048, 12, 12], ne = [12, 12, 2048, 1] // block_1 shape = [1, 2048, 12, 12], ne = [12, 12, 2048, 1]
// layer norm // layer norm
@ -741,18 +742,10 @@ struct clip_ctx * clip_model_load(const char * fname, const int verbosity = 1) {
{ {
std::map<enum ggml_type, uint32_t> n_type; std::map<enum ggml_type, uint32_t> n_type;
uint32_t n_type_max = 0;
enum ggml_type type_max = GGML_TYPE_F32;
for (int i = 0; i < n_tensors; i++) { for (int i = 0; i < n_tensors; i++) {
enum ggml_type type = gguf_get_tensor_type(ctx, i); enum ggml_type type = gguf_get_tensor_type(ctx, i);
n_type[type]++; n_type[type]++;
if (n_type_max < n_type[type]) {
n_type_max = n_type[type];
type_max = type;
}
} }
printf("%s: Dumping metadata keys/values. Note: KV overrides do not apply in this output.\n", __func__); printf("%s: Dumping metadata keys/values. Note: KV overrides do not apply in this output.\n", __func__);
@ -795,14 +788,12 @@ struct clip_ctx * clip_model_load(const char * fname, const int verbosity = 1) {
size_t tensor_size = ggml_nbytes(cur); size_t tensor_size = ggml_nbytes(cur);
buffer_size += tensor_size; buffer_size += tensor_size;
if (verbosity >= 3) { if (verbosity >= 3) {
printf("%s: tensor[%d]: n_dims = %d, name = %s, tensor_size=%zu, offset=%zu, shape:[%d, %d, %d, %d], type: %d\n", __func__, i, printf("%s: tensor[%d]: n_dims = %d, name = %s, tensor_size=%zu, offset=%zu, shape:[%" PRIu64 ", %" PRIu64 ", %" PRIu64 ", %" PRIu64 "], type = %s\n",
ggml_n_dims(cur), cur->name, tensor_size, offset, cur->ne[0], cur->ne[1], cur->ne[2], cur->ne[3], type); __func__, i, ggml_n_dims(cur), cur->name, tensor_size, offset, cur->ne[0], cur->ne[1], cur->ne[2], cur->ne[3], ggml_type_name(type));
} }
} }
} }
buffer_size += n_tensors * 128 /* CLIP PADDING */; buffer_size += n_tensors * 128 /* CLIP PADDING */;
clip_ctx * new_clip = new clip_ctx; clip_ctx * new_clip = new clip_ctx;

75
examples/perplexity/perplexity.cpp
View file

@ -223,13 +223,18 @@ struct kl_divergence_result {
double sum_kld2 = 0; double sum_kld2 = 0;
double sum_nll_diff = 0; double sum_nll_diff = 0;
double sum_nll_diff2 = 0; double sum_nll_diff2 = 0;
size_t n_same_top = 0;
size_t count = 0; size_t count = 0;
}; };
static void log_softmax(int n_vocab, const float * logits, const uint16_t * base_log_prob, int tok, kl_divergence_result & kld) { static double log_softmax(int n_vocab, const float * logits, const uint16_t * base_log_prob, int tok, kl_divergence_result & kld) {
float max_logit = logits[0]; float max_logit = logits[0];
int imax = 0;
for (int i = 1; i < n_vocab; ++i) { for (int i = 1; i < n_vocab; ++i) {
max_logit = std::max(max_logit, logits[i]); if (logits[i] > max_logit) {
max_logit = logits[i];
imax = i;
}
} }
double sum_exp = 0.0; double sum_exp = 0.0;
for (int i = 0; i < n_vocab; ++i) { for (int i = 0; i < n_vocab; ++i) {
@ -248,8 +253,14 @@ static void log_softmax(int n_vocab, const float * logits, const uint16_t * base
kld.sum_nll_diff2 += nll*nll; kld.sum_nll_diff2 += nll*nll;
max_logit += log_sum_exp; max_logit += log_sum_exp;
double sum = 0; double sum = 0;
int imax_base = -1;
float p_log_base_max = 0;
for (int i = 0; i < n_vocab; ++i) { for (int i = 0; i < n_vocab; ++i) {
const float p_log_base = scale*base_log_prob[i] + min_log_prob; const float p_log_base = scale*base_log_prob[i] + min_log_prob;
if (i == 0 || p_log_base > p_log_base_max) {
p_log_base_max = p_log_base;
imax_base = i;
}
if (p_log_base > -16.f) { if (p_log_base > -16.f) {
const float p_base = expf(p_log_base); const float p_base = expf(p_log_base);
sum += p_base * (p_log_base - logits[i] + max_logit); sum += p_base * (p_log_base - logits[i] + max_logit);
@ -258,14 +269,17 @@ static void log_softmax(int n_vocab, const float * logits, const uint16_t * base
kld.sum_kld += sum; kld.sum_kld += sum;
kld.sum_kld2 += sum*sum; kld.sum_kld2 += sum*sum;
++kld.count; ++kld.count;
if (imax == imax_base) ++kld.n_same_top;
return sum;
} }
static void process_logits(int n_vocab, const float * logits, const int * tokens, int n_token, static void process_logits(int n_vocab, const float * logits, const int * tokens, int n_token,
std::vector<std::thread> & workers, const std::vector<uint16_t> & base_log_probs, kl_divergence_result & kld) { std::vector<std::thread> & workers, const std::vector<uint16_t> & base_log_probs, kl_divergence_result & kld,
float * kld_values) {
std::mutex mutex; std::mutex mutex;
const int nv = 2*((n_vocab + 1)/2) + 4; const int nv = 2*((n_vocab + 1)/2) + 4;
int counter = 0; int counter = 0;
auto compute = [&mutex, &counter, &base_log_probs, &kld, n_vocab, logits, tokens, n_token, nv] () { auto compute = [&mutex, &counter, &base_log_probs, &kld, n_vocab, logits, tokens, n_token, nv, kld_values] () {
kl_divergence_result local_kld; kl_divergence_result local_kld;
while (true) { while (true) {
std::unique_lock<std::mutex> lock(mutex); std::unique_lock<std::mutex> lock(mutex);
@ -277,11 +291,13 @@ static void process_logits(int n_vocab, const float * logits, const int * tokens
kld.sum_kld2 += local_kld.sum_kld2; kld.sum_kld2 += local_kld.sum_kld2;
kld.sum_nll_diff += local_kld.sum_nll_diff; kld.sum_nll_diff += local_kld.sum_nll_diff;
kld.sum_nll_diff2 += local_kld.sum_nll_diff2; kld.sum_nll_diff2 += local_kld.sum_nll_diff2;
kld.n_same_top += local_kld.n_same_top;
kld.count += local_kld.count; kld.count += local_kld.count;
break; break;
} }
lock.unlock(); lock.unlock();
log_softmax(n_vocab, logits + i*n_vocab, base_log_probs.data() + i*nv, tokens[i+1], local_kld); double v = log_softmax(n_vocab, logits + i*n_vocab, base_log_probs.data() + i*nv, tokens[i+1], local_kld);
kld_values[i] = (float)v;
} }
}; };
for (auto & w : workers) { for (auto & w : workers) {
@ -1203,11 +1219,11 @@ static void winogrande_score(llama_context * ctx, const gpt_params & params) {
printf("Final Winogrande score(%d tasks): %.4lf +/- %.4lf\n", n_done, 100*p, sigma); printf("Final Winogrande score(%d tasks): %.4lf +/- %.4lf\n", n_done, 100*p, sigma);
} }
static bool deserialize_string(std::istream& in, std::string& str) { static bool deserialize_string(std::istream & in, std::string & str) {
uint32_t size; uint32_t size;
if (!in.read((char *)&size, sizeof(size)).fail()) { if (!in.read((char *)&size, sizeof(size)).fail()) {
str.resize(size); str.resize(size);
if (!in.read((char *)str.data(), size).fail()) return true; if (!in.read((char *)&str[0], size).fail()) return true;
} }
return false; return false;
} }
@ -1616,7 +1632,7 @@ static void kl_divergence(llama_context * ctx, const gpt_params & params) {
in.read((char *)&n_vocab, sizeof(n_vocab)); in.read((char *)&n_vocab, sizeof(n_vocab));
in.read((char *)&n_chunk, sizeof(n_chunk)); in.read((char *)&n_chunk, sizeof(n_chunk));
if (in.fail()) { if (in.fail()) {
fprintf(stderr, "%s: failed rwading n_vocab, n_chunk from %s\n", __func__, params.logits_file.c_str()); fprintf(stderr, "%s: failed reading n_vocab, n_chunk from %s\n", __func__, params.logits_file.c_str());
return; return;
} }
if (n_vocab != llama_n_vocab(llama_get_model(ctx))) { if (n_vocab != llama_n_vocab(llama_get_model(ctx))) {
@ -1635,6 +1651,7 @@ static void kl_divergence(llama_context * ctx, const gpt_params & params) {
const bool add_bos = llama_should_add_bos_token(llama_get_model(ctx)); const bool add_bos = llama_should_add_bos_token(llama_get_model(ctx));
std::vector<uint16_t> log_probs_uint16(size_t(n_ctx - 1 - n_ctx/2) * nv); std::vector<uint16_t> log_probs_uint16(size_t(n_ctx - 1 - n_ctx/2) * nv);
std::vector<float> kld_values(size_t(n_ctx - 1 - n_ctx/2)*n_chunk);
std::vector<float> logits; std::vector<float> logits;
if (num_batches > 1) { if (num_batches > 1) {
logits.reserve(n_ctx * n_vocab); logits.reserve(n_ctx * n_vocab);
@ -1653,6 +1670,7 @@ static void kl_divergence(llama_context * ctx, const gpt_params & params) {
}; };
kl_divergence_result kld; kl_divergence_result kld;
auto kld_ptr = kld_values.data();
for (int i = 0; i < n_chunk; ++i) { for (int i = 0; i < n_chunk; ++i) {
const int start = i * n_ctx; const int start = i * n_ctx;
@ -1706,20 +1724,24 @@ static void kl_divergence(llama_context * ctx, const gpt_params & params) {
} }
fprintf(stderr, "%.2f minutes\n", total_seconds / 60.0); fprintf(stderr, "%.2f minutes\n", total_seconds / 60.0);
printf("\nchunk PPL ln(PPL(Q)/PPL(base)) KL-Divergence\n"); printf("\nchunk PPL ln(PPL(Q)/PPL(base)) KL-Divergence Same top\n");
} }
const int first = n_ctx/2; const int first = n_ctx/2;
const float * all_logits = num_batches > 1 ? logits.data() : llama_get_logits(ctx); const float * all_logits = num_batches > 1 ? logits.data() : llama_get_logits(ctx);
process_logits(n_vocab, all_logits + first*n_vocab, tokens.data() + start + first, n_ctx - 1 - first, process_logits(n_vocab, all_logits + first*n_vocab, tokens.data() + start + first, n_ctx - 1 - first,
workers, log_probs_uint16, kld); workers, log_probs_uint16, kld, kld_ptr);
kld_ptr += n_ctx - 1 - first;
auto ppl = mean_and_uncertainty(kld.sum_nll, kld.sum_nll2, kld.count); auto ppl = mean_and_uncertainty(kld.sum_nll, kld.sum_nll2, kld.count);
auto log_ppl_ratio = mean_and_uncertainty(kld.sum_nll_diff, kld.sum_nll_diff2, kld.count); auto log_ppl_ratio = mean_and_uncertainty(kld.sum_nll_diff, kld.sum_nll_diff2, kld.count);
auto kl_div = mean_and_uncertainty(kld.sum_kld, kld.sum_kld2, kld.count); auto kl_div = mean_and_uncertainty(kld.sum_kld, kld.sum_kld2, kld.count);
auto p_top = 1.*kld.n_same_top/kld.count;
auto d_p_top = sqrt(p_top*(1 - p_top)/(kld.count - 1));
printf("%4d %10.4lf %10.5lf ± %10.5f %10.5f ± %10.5lf\n", i+1, exp(ppl.first), printf("%4d %10.4lf %10.5lf ± %10.5f %10.5f ± %10.5lf %.5f ± %.5f\n", i+1, exp(ppl.first),
log_ppl_ratio.first, log_ppl_ratio.second, kl_div.first, kl_div.second); log_ppl_ratio.first, log_ppl_ratio.second, kl_div.first, kl_div.second,
p_top, d_p_top);
fflush(stdout); fflush(stdout);
@ -1727,6 +1749,35 @@ static void kl_divergence(llama_context * ctx, const gpt_params & params) {
} }
printf("\n"); printf("\n");
if (kld.count < 100) return; // we do not wish to do statistics on so few values
std::sort(kld_values.begin(), kld_values.end());
printf("===== KL-divergence statistics\n");
auto kl_div = mean_and_uncertainty(kld.sum_kld, kld.sum_kld2, kld.count);
printf("Average: %10.6f ±%10.6lf\n", kl_div.first, kl_div.second);
auto kld_median = kld_values.size()%2 == 0 ? 0.5f*(kld_values[kld_values.size()/2] + kld_values[kld_values.size()/2-1])
: kld_values[kld_values.size()/2];
printf("Median : %10.6f\n", kld_median);
auto percentile = [&kld_values] (float fraction) {
if (fraction <= 0) return kld_values.front();
if (fraction >= 1) return kld_values.back();
float p = fraction*(kld_values.size() - 1);
size_t ip = size_t(p); p -= ip;
return (1 - p)*kld_values[ip] + p*kld_values[std::min(ip+1, kld_values.size()-1)];
};
printf("Maximum: %10.6f\n", kld_values.back());
printf("KLD_99 : %10.6f\n", percentile(0.99f));
printf("KLD_95 : %10.6f\n", percentile(0.95f));
printf("KLD_90 : %10.6f\n", percentile(0.90f));
printf("Minimum: %10.6f\n", kld_values.front());
printf("KLD_01 : %10.6f\n", percentile(0.01f));
printf("KLD_05 : %10.6f\n", percentile(0.05f));
printf("KLD_10 : %10.6f\n", percentile(0.10f));
} }
int main(int argc, char ** argv) { int main(int argc, char ** argv) {

4
ggml-alloc.c
View file

@ -109,8 +109,8 @@ void ggml_tallocr_alloc(ggml_tallocr_t alloc, struct ggml_tensor * tensor) {
if (block->size >= size) { if (block->size >= size) {
best_fit_block = alloc->n_free_blocks - 1; best_fit_block = alloc->n_free_blocks - 1;
} else { } else {
fprintf(stderr, "%s: not enough space in the buffer (needed %zu, largest block available %zu)\n", fprintf(stderr, "%s: not enough space in the buffer to allocate %s (needed %zu, largest block available %zu)\n",
__func__, size, max_avail); __func__, tensor->name, size, max_avail);
GGML_ASSERT(!"not enough space in the buffer"); GGML_ASSERT(!"not enough space in the buffer");
return; return;
} }

89
ggml-cuda.cu
View file

@ -13,6 +13,10 @@
#include <map> #include <map>
#include <array> #include <array>
// stringize macro for converting __CUDA_ARCH_LIST__ (list of integers) to string
#define STRINGIZE_IMPL(...) #__VA_ARGS__
#define STRINGIZE(...) STRINGIZE_IMPL(__VA_ARGS__)
#if defined(GGML_USE_HIPBLAS) #if defined(GGML_USE_HIPBLAS)
#include <hip/hip_runtime.h> #include <hip/hip_runtime.h>
#include <hipblas/hipblas.h> #include <hipblas/hipblas.h>
@ -585,13 +589,28 @@ static cuda_device_capabilities g_device_caps[GGML_CUDA_MAX_DEVICES] = { {0, 0,
static cublasHandle_t g_cublas_handles[GGML_CUDA_MAX_DEVICES] = {nullptr}; static cublasHandle_t g_cublas_handles[GGML_CUDA_MAX_DEVICES] = {nullptr};
[[noreturn]] [[noreturn]]
static __device__ void bad_arch() { static __device__ void no_device_code(
printf("ERROR: ggml-cuda was compiled without support for the current GPU architecture.\n"); const char * file_name, const int line, const char * function_name, const int arch, const char * arch_list) {
#if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)
printf("%s:%d: ERROR: HIP kernel %s has no device code compatible with HIP arch %d.\n",
file_name, line, function_name, arch);
(void) arch_list;
#else
printf("%s:%d: ERROR: CUDA kernel %s has no device code compatible with CUDA arch %d. ggml-cuda.cu was compiled for: %s\n",
file_name, line, function_name, arch, arch_list);
#endif // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)
__trap(); __trap();
(void) bad_arch; // suppress unused function warning (void) no_device_code; // suppress unused function warning
} }
#ifdef __CUDA_ARCH__
#define NO_DEVICE_CODE no_device_code(__FILE__, __LINE__, __FUNCTION__, __CUDA_ARCH__, STRINGIZE(__CUDA_ARCH_LIST__))
#else
#define NO_DEVICE_CODE GGML_ASSERT(false && "NO_DEVICE_CODE not valid in host code.")
#endif // __CUDA_ARCH__
static __device__ __forceinline__ float warp_reduce_sum(float x) { static __device__ __forceinline__ float warp_reduce_sum(float x) {
#pragma unroll #pragma unroll
for (int mask = 16; mask > 0; mask >>= 1) { for (int mask = 16; mask > 0; mask >>= 1) {
@ -618,7 +637,7 @@ static __device__ __forceinline__ half2 warp_reduce_sum(half2 a) {
return a; return a;
#else #else
(void) a; (void) a;
bad_arch(); NO_DEVICE_CODE;
#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_PASCAL #endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_PASCAL
} }
@ -639,7 +658,7 @@ static __device__ __forceinline__ half2 warp_reduce_max(half2 x) {
return x; return x;
#else #else
(void) x; (void) x;
bad_arch(); NO_DEVICE_CODE;
#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_PASCAL && CUDART_VERSION >= CUDART_HMAX #endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_PASCAL && CUDART_VERSION >= CUDART_HMAX
} }
@ -2422,7 +2441,7 @@ static __global__ void dequantize_block_q8_0_f16(const void * __restrict__ vx, h
} }
#else #else
(void) vx; (void) y; (void) k; (void) vx; (void) y; (void) k;
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= CC_PASCAL #endif // __CUDA_ARCH__ >= CC_PASCAL
} }
@ -2453,7 +2472,7 @@ template <int vdr> static __device__ __forceinline__ float vec_dot_q4_0_q8_1_imp
// second part effectively subtracts 8 from each quant value // second part effectively subtracts 8 from each quant value
return d4 * (sumi * ds8f.x - (8*vdr/QI4_0) * ds8f.y); return d4 * (sumi * ds8f.x - (8*vdr/QI4_0) * ds8f.y);
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2490,7 +2509,7 @@ template <int vdr> static __device__ __forceinline__ float vec_dot_q4_1_q8_1_imp
// scale second part of sum by QI8_1/(vdr * QR4_1) to compensate for multiple threads adding it // scale second part of sum by QI8_1/(vdr * QR4_1) to compensate for multiple threads adding it
return sumi * d4d8 + m4s8 / (QI8_1 / (vdr * QR4_1)); return sumi * d4d8 + m4s8 / (QI8_1 / (vdr * QR4_1));
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2525,7 +2544,7 @@ template <int vdr> static __device__ __forceinline__ float vec_dot_q5_0_q8_1_imp
// second part effectively subtracts 16 from each quant value // second part effectively subtracts 16 from each quant value
return d5 * (sumi * ds8f.x - (16*vdr/QI5_0) * ds8f.y); return d5 * (sumi * ds8f.x - (16*vdr/QI5_0) * ds8f.y);
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2570,7 +2589,7 @@ template <int vdr> static __device__ __forceinline__ float vec_dot_q5_1_q8_1_imp
return sumi*d5d8 + m5s8 / (QI5_1 / vdr); return sumi*d5d8 + m5s8 / (QI5_1 / vdr);
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2591,7 +2610,7 @@ template <int vdr> static __device__ __forceinline__ float vec_dot_q8_0_q8_1_imp
return d8_0*d8_1 * sumi; return d8_0*d8_1 * sumi;
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2621,7 +2640,7 @@ template <int vdr> static __device__ __forceinline__ float vec_dot_q8_1_q8_1_imp
// scale second part of sum by QI8_1/ vdr to compensate for multiple threads adding it // scale second part of sum by QI8_1/ vdr to compensate for multiple threads adding it
return sumi*d8d8 + m8s8 / (QI8_1 / vdr); return sumi*d8d8 + m8s8 / (QI8_1 / vdr);
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2656,7 +2675,7 @@ static __device__ __forceinline__ float vec_dot_q2_K_q8_1_impl_mmvq(
return dm2f.x*sumf_d - dm2f.y*sumf_m; return dm2f.x*sumf_d - dm2f.y*sumf_m;
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2693,7 +2712,7 @@ static __device__ __forceinline__ float vec_dot_q2_K_q8_1_impl_mmq(
return d8 * (dm2f.x*sumi_d - dm2f.y*sumi_m); return d8 * (dm2f.x*sumi_d - dm2f.y*sumi_m);
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2733,7 +2752,7 @@ static __device__ __forceinline__ float vec_dot_q3_K_q8_1_impl_mmvq(
return d3 * sumf; return d3 * sumf;
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2758,7 +2777,7 @@ static __device__ __forceinline__ float vec_dot_q3_K_q8_1_impl_mmq(
return d3*d8 * sumi; return d3*d8 * sumi;
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2791,7 +2810,7 @@ static __device__ __forceinline__ float vec_dot_q4_K_q8_1_impl_vmmq(
return dm4f.x*sumf_d - dm4f.y*sumf_m; return dm4f.x*sumf_d - dm4f.y*sumf_m;
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2824,7 +2843,7 @@ static __device__ __forceinline__ float vec_dot_q4_K_q8_1_impl_mmq(
return dm4f.x*sumf_d - dm4f.y*sumf_m; return dm4f.x*sumf_d - dm4f.y*sumf_m;
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2864,7 +2883,7 @@ static __device__ __forceinline__ float vec_dot_q5_K_q8_1_impl_vmmq(
return dm5f.x*sumf_d - dm5f.y*sumf_m; return dm5f.x*sumf_d - dm5f.y*sumf_m;
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2897,7 +2916,7 @@ static __device__ __forceinline__ float vec_dot_q5_K_q8_1_impl_mmq(
return dm4f.x*sumf_d - dm4f.y*sumf_m; return dm4f.x*sumf_d - dm4f.y*sumf_m;
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2927,7 +2946,7 @@ static __device__ __forceinline__ float vec_dot_q6_K_q8_1_impl_mmvq(
return d*sumf; return d*sumf;
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -2958,7 +2977,7 @@ static __device__ __forceinline__ float vec_dot_q6_K_q8_1_impl_mmq(
return d6 * sumf_d; return d6 * sumf_d;
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
} }
@ -3824,7 +3843,7 @@ static __device__ __forceinline__ float vec_dot_q4_K_q8_1(
return dall * sumf_d - dmin * sumf_m; return dall * sumf_d - dmin * sumf_m;
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
#endif #endif
@ -4007,7 +4026,7 @@ static __device__ __forceinline__ float vec_dot_q5_K_q8_1(
return d * sumf_d; return d * sumf_d;
#else #else
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= MIN_CC_DP4A #endif // __CUDA_ARCH__ >= MIN_CC_DP4A
#endif #endif
@ -4502,7 +4521,7 @@ template <bool need_check> static __global__ void
(vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst); (vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst);
#else #else
(void) vec_dot_q4_0_q8_1_mul_mat; (void) vec_dot_q4_0_q8_1_mul_mat;
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= CC_VOLTA #endif // __CUDA_ARCH__ >= CC_VOLTA
} }
@ -4571,7 +4590,7 @@ template <bool need_check> static __global__ void
(vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst); (vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst);
#else #else
(void) vec_dot_q4_1_q8_1_mul_mat; (void) vec_dot_q4_1_q8_1_mul_mat;
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= CC_VOLTA #endif // __CUDA_ARCH__ >= CC_VOLTA
} }
@ -4638,7 +4657,7 @@ template <bool need_check> static __global__ void
(vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst); (vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst);
#else #else
(void) vec_dot_q5_0_q8_1_mul_mat; (void) vec_dot_q5_0_q8_1_mul_mat;
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= CC_VOLTA #endif // __CUDA_ARCH__ >= CC_VOLTA
} }
@ -4705,7 +4724,7 @@ mul_mat_q5_1(
(vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst); (vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst);
#else #else
(void) vec_dot_q5_1_q8_1_mul_mat; (void) vec_dot_q5_1_q8_1_mul_mat;
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= CC_VOLTA #endif // __CUDA_ARCH__ >= CC_VOLTA
} }
@ -4772,7 +4791,7 @@ template <bool need_check> static __global__ void
(vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst); (vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst);
#else #else
(void) vec_dot_q8_0_q8_1_mul_mat; (void) vec_dot_q8_0_q8_1_mul_mat;
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= CC_VOLTA #endif // __CUDA_ARCH__ >= CC_VOLTA
} }
@ -4839,7 +4858,7 @@ mul_mat_q2_K(
(vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst); (vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst);
#else #else
(void) vec_dot_q2_K_q8_1_mul_mat; (void) vec_dot_q2_K_q8_1_mul_mat;
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= CC_VOLTA #endif // __CUDA_ARCH__ >= CC_VOLTA
} }
@ -4908,7 +4927,7 @@ template <bool need_check> static __global__ void
(vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst); (vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst);
#else #else
(void) vec_dot_q3_K_q8_1_mul_mat; (void) vec_dot_q3_K_q8_1_mul_mat;
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= CC_VOLTA #endif // __CUDA_ARCH__ >= CC_VOLTA
} }
@ -4977,7 +4996,7 @@ template <bool need_check> static __global__ void
(vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst); (vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst);
#else #else
(void) vec_dot_q4_K_q8_1_mul_mat; (void) vec_dot_q4_K_q8_1_mul_mat;
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= CC_VOLTA #endif // __CUDA_ARCH__ >= CC_VOLTA
} }
@ -5044,7 +5063,7 @@ mul_mat_q5_K(
(vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst); (vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst);
#else #else
(void) vec_dot_q5_K_q8_1_mul_mat; (void) vec_dot_q5_K_q8_1_mul_mat;
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= CC_VOLTA #endif // __CUDA_ARCH__ >= CC_VOLTA
} }
@ -5113,7 +5132,7 @@ template <bool need_check> static __global__ void
(vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst); (vx, vy, dst, ncols_x, nrows_x, ncols_y, nrows_y, nrows_dst);
#else #else
(void) vec_dot_q6_K_q8_1_mul_mat; (void) vec_dot_q6_K_q8_1_mul_mat;
bad_arch(); NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= CC_VOLTA #endif // __CUDA_ARCH__ >= CC_VOLTA
} }
@ -5836,7 +5855,7 @@ static __global__ void soft_max_f16(const float * x, const float * y, float * ds
} }
#else #else
(void) x; (void) y; (void) dst; (void) ncols_par; (void) nrows_y; (void) scale; (void) x; (void) y; (void) dst; (void) ncols_par; (void) nrows_y; (void) scale;
bad_arch(); NO_DEVICE_CODE;
#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_PASCAL && CUDART_VERSION >= CUDART_HMAX #endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_PASCAL && CUDART_VERSION >= CUDART_HMAX
} }

3
ggml-metal.m
View file

@ -668,7 +668,8 @@ static bool ggml_metal_supports_op(const struct ggml_metal_context * ctx, const
return true; return true;
case GGML_OP_MUL_MAT: case GGML_OP_MUL_MAT:
case GGML_OP_MUL_MAT_ID: case GGML_OP_MUL_MAT_ID:
return ctx->support_simdgroup_reduction; return ctx->support_simdgroup_reduction &&
(op->src[0]->type != GGML_TYPE_F32 || op->src[1]->type == GGML_TYPE_F32);
case GGML_OP_CPY: case GGML_OP_CPY:
case GGML_OP_DUP: case GGML_OP_DUP:
case GGML_OP_CONT: case GGML_OP_CONT:

9
ggml.c
View file

@ -5368,14 +5368,12 @@ struct ggml_tensor * ggml_conv_depthwise_2d(
struct ggml_context * ctx, struct ggml_context * ctx,
struct ggml_tensor * a, struct ggml_tensor * a,
struct ggml_tensor * b, struct ggml_tensor * b,
struct ggml_tensor * c,
int s0, int s0,
int s1, int s1,
int p0, int p0,
int p1, int p1,
int d0, int d0,
int d1) { int d1) {
struct ggml_tensor * new_a = ggml_reshape_4d(ctx, a, a->ne[0], a->ne[1], 1, a->ne[2] * a->ne[3]); struct ggml_tensor * new_a = ggml_reshape_4d(ctx, a, a->ne[0], a->ne[1], 1, a->ne[2] * a->ne[3]);
struct ggml_tensor * im2col = ggml_im2col(ctx, new_a, struct ggml_tensor * im2col = ggml_im2col(ctx, new_a,
ggml_reshape_4d(ctx, b, b->ne[0], b->ne[1], 1, b->ne[2] * b->ne[3]), ggml_reshape_4d(ctx, b, b->ne[0], b->ne[1], 1, b->ne[2] * b->ne[3]),
@ -9991,7 +9989,7 @@ static void ggml_compute_forward_mul_mat(
return; return;
} }
const int64_t tgemm0 = ggml_perf_time_us(); //const int64_t tgemm0 = ggml_perf_time_us();
for (int64_t i13 = 0; i13 < ne13; i13++) { for (int64_t i13 = 0; i13 < ne13; i13++) {
for (int64_t i12 = 0; i12 < ne12; i12++) { for (int64_t i12 = 0; i12 < ne12; i12++) {
const int64_t i03 = i13/r3; const int64_t i03 = i13/r3;
@ -16934,7 +16932,10 @@ struct ggml_cplan ggml_graph_plan(const struct ggml_cgraph * cgraph, int n_threa
if (ggml_compute_forward_mul_mat_use_blas(node)) { if (ggml_compute_forward_mul_mat_use_blas(node)) {
if (node->src[0]->type != GGML_TYPE_F32) { if (node->src[0]->type != GGML_TYPE_F32) {
// here we need memory for fully dequantized matrix from src0 // here we need memory for fully dequantized matrix from src0
cur = ggml_type_size(GGML_TYPE_F32)*ggml_nelements(node->src[0]); // take into account that src0 can be broadcasted into src1[2,3]
cur = ggml_type_size(GGML_TYPE_F32)
* node->src[0]->ne[0]*node->src[0]->ne[1]
* node->src[1]->ne[2]*node->src[1]->ne[3];
} }
} else } else
#endif #endif

1
ggml.h
View file

@ -1506,7 +1506,6 @@ extern "C" {
struct ggml_context * ctx, struct ggml_context * ctx,
struct ggml_tensor * a, struct ggml_tensor * a,
struct ggml_tensor * b, struct ggml_tensor * b,
struct ggml_tensor * c,
int s0, int s0,
int s1, int s1,
int p0, int p0,

371
llama.cpp
View file

@ -1679,6 +1679,9 @@ struct llama_context {
for (ggml_backend_t backend : backends) { for (ggml_backend_t backend : backends) {
ggml_backend_free(backend); ggml_backend_free(backend);
} }
ggml_backend_buffer_free(buf_input);
ggml_free(ctx_input);
} }
llama_cparams cparams; llama_cparams cparams;
@ -1725,8 +1728,14 @@ struct llama_context {
// allocator for the input tensors // allocator for the input tensors
ggml_tallocr * alloc = nullptr; ggml_tallocr * alloc = nullptr;
// temporary buffer for copying data to/from the backend // input tensors
std::vector<no_init<uint8_t>> buf_copy; ggml_backend_buffer_t buf_input = nullptr;
ggml_context * ctx_input = nullptr;
struct ggml_tensor * inp_tokens; // I32 [n_batch]
struct ggml_tensor * inp_embd; // F32 [n_embd, n_batch]
struct ggml_tensor * inp_pos; // I32 [n_batch]
struct ggml_tensor * inp_KQ_mask; // F32 [n_ctx, n_batch]
struct ggml_tensor * inp_K_shift; // I32 [n_ctx]
#ifdef GGML_USE_MPI #ifdef GGML_USE_MPI
ggml_mpi_context * ctx_mpi = NULL; ggml_mpi_context * ctx_mpi = NULL;
@ -2311,18 +2320,18 @@ struct llama_model_loader {
} }
switch (type_max) { switch (type_max) {
case GGML_TYPE_F32: ftype = LLAMA_FTYPE_ALL_F32; break; case GGML_TYPE_F32: ftype = LLAMA_FTYPE_ALL_F32; break;
case GGML_TYPE_F16: ftype = LLAMA_FTYPE_MOSTLY_F16; break; case GGML_TYPE_F16: ftype = LLAMA_FTYPE_MOSTLY_F16; break;
case GGML_TYPE_Q4_0: ftype = LLAMA_FTYPE_MOSTLY_Q4_0; break; case GGML_TYPE_Q4_0: ftype = LLAMA_FTYPE_MOSTLY_Q4_0; break;
case GGML_TYPE_Q4_1: ftype = LLAMA_FTYPE_MOSTLY_Q4_1; break; case GGML_TYPE_Q4_1: ftype = LLAMA_FTYPE_MOSTLY_Q4_1; break;
case GGML_TYPE_Q5_0: ftype = LLAMA_FTYPE_MOSTLY_Q5_0; break; case GGML_TYPE_Q5_0: ftype = LLAMA_FTYPE_MOSTLY_Q5_0; break;
case GGML_TYPE_Q5_1: ftype = LLAMA_FTYPE_MOSTLY_Q5_1; break; case GGML_TYPE_Q5_1: ftype = LLAMA_FTYPE_MOSTLY_Q5_1; break;
case GGML_TYPE_Q8_0: ftype = LLAMA_FTYPE_MOSTLY_Q8_0; break; case GGML_TYPE_Q8_0: ftype = LLAMA_FTYPE_MOSTLY_Q8_0; break;
case GGML_TYPE_Q2_K: ftype = LLAMA_FTYPE_MOSTLY_Q2_K; break; case GGML_TYPE_Q2_K: ftype = LLAMA_FTYPE_MOSTLY_Q2_K; break;
case GGML_TYPE_Q3_K: ftype = LLAMA_FTYPE_MOSTLY_Q3_K_M; break; case GGML_TYPE_Q3_K: ftype = LLAMA_FTYPE_MOSTLY_Q3_K_M; break;
case GGML_TYPE_Q4_K: ftype = LLAMA_FTYPE_MOSTLY_Q4_K_M; break; case GGML_TYPE_Q4_K: ftype = LLAMA_FTYPE_MOSTLY_Q4_K_M; break;
case GGML_TYPE_Q5_K: ftype = LLAMA_FTYPE_MOSTLY_Q5_K_M; break; case GGML_TYPE_Q5_K: ftype = LLAMA_FTYPE_MOSTLY_Q5_K_M; break;
case GGML_TYPE_Q6_K: ftype = LLAMA_FTYPE_MOSTLY_Q6_K; break; case GGML_TYPE_Q6_K: ftype = LLAMA_FTYPE_MOSTLY_Q6_K; break;
case GGML_TYPE_IQ2_XXS: ftype = LLAMA_FTYPE_MOSTLY_IQ2_XXS; break; case GGML_TYPE_IQ2_XXS: ftype = LLAMA_FTYPE_MOSTLY_IQ2_XXS; break;
case GGML_TYPE_IQ2_XS: ftype = LLAMA_FTYPE_MOSTLY_IQ2_XS; break; case GGML_TYPE_IQ2_XS: ftype = LLAMA_FTYPE_MOSTLY_IQ2_XS; break;
default: default:
@ -4119,22 +4128,24 @@ static struct ggml_tensor * llm_build_inp_embd(
const llama_hparams & hparams, const llama_hparams & hparams,
const llama_batch & batch, const llama_batch & batch,
struct ggml_tensor * tok_embd, struct ggml_tensor * tok_embd,
struct ggml_tensor * inp_tokens,
struct ggml_tensor * inp_embd,
const llm_build_cb & cb) { const llm_build_cb & cb) {
const int64_t n_embd = hparams.n_embd; const int64_t n_embd = hparams.n_embd;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
if (batch.token) { if (batch.token) {
struct ggml_tensor * inp_tokens = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, batch.n_tokens); struct ggml_tensor * inp_tokens_v = ggml_view_1d(ctx, inp_tokens, batch.n_tokens, 0);
cb(inp_tokens, "inp_tokens", -1); cb(inp_tokens, "inp_tokens", -1);
inpL = ggml_get_rows(ctx, tok_embd, inp_tokens); inpL = ggml_get_rows(ctx, tok_embd, inp_tokens_v);
} else { } else {
#ifdef GGML_USE_MPI #ifdef GGML_USE_MPI
GGML_ASSERT(false && "not implemented"); GGML_ASSERT(false && "not implemented");
#endif #endif
inpL = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, n_embd, batch.n_tokens); inpL = ggml_view_2d(ctx, inp_embd, n_embd, batch.n_tokens, inp_embd->nb[1], 0);
} }
return inpL; return inpL;
@ -4148,6 +4159,7 @@ static void llm_build_k_shift(
const llama_cparams & cparams, const llama_cparams & cparams,
const llama_kv_cache & kv, const llama_kv_cache & kv,
struct ggml_cgraph * graph, struct ggml_cgraph * graph,
struct ggml_tensor * K_shift,
llm_rope_type type, llm_rope_type type,
int64_t n_ctx, int64_t n_ctx,
float freq_base, float freq_base,
@ -4164,9 +4176,6 @@ static void llm_build_k_shift(
const float beta_fast = cparams.yarn_beta_fast; const float beta_fast = cparams.yarn_beta_fast;
const float beta_slow = cparams.yarn_beta_slow; const float beta_slow = cparams.yarn_beta_slow;
struct ggml_tensor * K_shift = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, n_ctx);
cb(K_shift, "K_shift", -1);
int rope_type = 0; int rope_type = 0;
switch (type) { switch (type) {
@ -4470,9 +4479,9 @@ static struct ggml_tensor * llm_build_kv(
// these nodes are added to the graph together so that they are not reordered // these nodes are added to the graph together so that they are not reordered
// by doing so, the number of splits in the graph is reduced // by doing so, the number of splits in the graph is reduced
ggml_build_forward_expand(graph, q_cur);
ggml_build_forward_expand(graph, k_cur); ggml_build_forward_expand(graph, k_cur);
ggml_build_forward_expand(graph, v_cur); ggml_build_forward_expand(graph, v_cur);
ggml_build_forward_expand(graph, q_cur);
llm_build_kv_store(ctx, hparams, kv, graph, k_cur, v_cur, n_ctx, n_tokens, kv_head, cb, il); llm_build_kv_store(ctx, hparams, kv, graph, k_cur, v_cur, n_ctx, n_tokens, kv_head, cb, il);
@ -4487,6 +4496,7 @@ static struct ggml_tensor * llm_build_kv(
struct llm_build_context { struct llm_build_context {
const llama_model & model; const llama_model & model;
const llama_context & lctx;
const llama_hparams & hparams; const llama_hparams & hparams;
const llama_cparams & cparams; const llama_cparams & cparams;
const llama_batch & batch; const llama_batch & batch;
@ -4533,6 +4543,7 @@ struct llm_build_context {
const llm_build_cb & cb, const llm_build_cb & cb,
bool worst_case) : bool worst_case) :
model (lctx.model), model (lctx.model),
lctx (lctx),
hparams (model.hparams), hparams (model.hparams),
cparams (lctx.cparams), cparams (lctx.cparams),
batch (batch), batch (batch),
@ -4593,20 +4604,20 @@ struct llm_build_context {
struct ggml_tensor * cur; struct ggml_tensor * cur;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// inp_pos - contains the positions // inp_pos - contains the positions
struct ggml_tensor * inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens); struct ggml_tensor * inp_pos = ggml_view_1d(ctx0, lctx.inp_pos, n_tokens, 0);
cb(inp_pos, "inp_pos", -1); cb(inp_pos, "inp_pos", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
// shift the entire K-cache if needed // shift the entire K-cache if needed
if (do_rope_shift) { if (do_rope_shift) {
llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, LLM_ROPE, n_ctx, freq_base, freq_scale, cb); llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, lctx.inp_K_shift, LLM_ROPE, n_ctx, freq_base, freq_scale, cb);
} }
for (int il = 0; il < n_layer; ++il) { for (int il = 0; il < n_layer; ++il) {
@ -4777,20 +4788,20 @@ struct llm_build_context {
struct ggml_tensor * cur; struct ggml_tensor * cur;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// inp_pos - contains the positions // inp_pos - contains the positions
struct ggml_tensor * inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens); struct ggml_tensor * inp_pos = ggml_view_1d(ctx0, lctx.inp_pos, n_tokens, 0);
cb(inp_pos, "inp_pos", -1); cb(inp_pos, "inp_pos", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
// shift the entire K-cache if needed // shift the entire K-cache if needed
if (do_rope_shift) { if (do_rope_shift) {
llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, LLM_ROPE, n_ctx, freq_base, freq_scale, cb); llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, lctx.inp_K_shift, LLM_ROPE, n_ctx, freq_base, freq_scale, cb);
} }
for (int il = 0; il < n_layer; ++il) { for (int il = 0; il < n_layer; ++il) {
@ -4898,20 +4909,20 @@ struct llm_build_context {
struct ggml_tensor * cur; struct ggml_tensor * cur;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// inp_pos - contains the positions // inp_pos - contains the positions
struct ggml_tensor * inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens); struct ggml_tensor * inp_pos = ggml_view_1d(ctx0, lctx.inp_pos, n_tokens, 0);
cb(inp_pos, "inp_pos", -1); cb(inp_pos, "inp_pos", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
// shift the entire K-cache if needed // shift the entire K-cache if needed
if (do_rope_shift) { if (do_rope_shift) {
llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, LLM_ROPE_NEOX, n_ctx, freq_base, freq_scale, cb); llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, lctx.inp_K_shift, LLM_ROPE_NEOX, n_ctx, freq_base, freq_scale, cb);
} }
for (int il = 0; il < n_layer; ++il) { for (int il = 0; il < n_layer; ++il) {
@ -5020,15 +5031,15 @@ struct llm_build_context {
struct ggml_tensor * pos; struct ggml_tensor * pos;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// inp_pos - contains the positions // inp_pos - contains the positions
struct ggml_tensor * inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens); struct ggml_tensor * inp_pos = ggml_view_1d(ctx0, lctx.inp_pos, n_tokens, 0);
cb(inp_pos, "inp_pos", -1); cb(inp_pos, "inp_pos", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
pos = ggml_get_rows(ctx0, model.pos_embd, inp_pos); pos = ggml_get_rows(ctx0, model.pos_embd, inp_pos);
@ -5117,19 +5128,19 @@ struct llm_build_context {
struct ggml_tensor * cur; struct ggml_tensor * cur;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// inp_pos - contains the positions // inp_pos - contains the positions
struct ggml_tensor * inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens); struct ggml_tensor * inp_pos = ggml_view_1d(ctx0, lctx.inp_pos, n_tokens, 0);
cb(inp_pos, "inp_pos", -1); cb(inp_pos, "inp_pos", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
if (do_rope_shift) { if (do_rope_shift) {
llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, LLM_ROPE_NEOX, n_ctx, freq_base, freq_scale, cb); llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, lctx.inp_K_shift, LLM_ROPE_NEOX, n_ctx, freq_base, freq_scale, cb);
} }
for (int il = 0; il < n_layer; ++il) { for (int il = 0; il < n_layer; ++il) {
@ -5324,11 +5335,11 @@ struct llm_build_context {
struct ggml_tensor * cur; struct ggml_tensor * cur;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
for (int il = 0; il < n_layer; ++il) { for (int il = 0; il < n_layer; ++il) {
@ -5414,11 +5425,11 @@ struct llm_build_context {
struct ggml_tensor * cur; struct ggml_tensor * cur;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
inpL = llm_build_norm(ctx0, inpL, hparams, inpL = llm_build_norm(ctx0, inpL, hparams,
@ -5507,11 +5518,11 @@ struct llm_build_context {
struct ggml_tensor * cur; struct ggml_tensor * cur;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
for (int il = 0; il < n_layer; ++il) { for (int il = 0; il < n_layer; ++il) {
@ -5603,20 +5614,20 @@ struct llm_build_context {
struct ggml_tensor * cur; struct ggml_tensor * cur;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// inp_pos - contains the positions // inp_pos - contains the positions
struct ggml_tensor * inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens); struct ggml_tensor * inp_pos = ggml_view_1d(ctx0, lctx.inp_pos, n_tokens, 0);
cb(inp_pos, "inp_pos", -1); cb(inp_pos, "inp_pos", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
// shift the entire K-cache if needed // shift the entire K-cache if needed
if (do_rope_shift) { if (do_rope_shift) {
llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, LLM_ROPE_NEOX, n_ctx, freq_base, freq_scale, cb); llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, lctx.inp_K_shift, LLM_ROPE_NEOX, n_ctx, freq_base, freq_scale, cb);
} }
for (int il = 0; il < n_layer; ++il) { for (int il = 0; il < n_layer; ++il) {
@ -5726,20 +5737,20 @@ struct llm_build_context {
struct ggml_tensor * cur; struct ggml_tensor * cur;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// inp_pos - contains the positions // inp_pos - contains the positions
struct ggml_tensor * inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens); struct ggml_tensor * inp_pos = ggml_view_1d(ctx0, lctx.inp_pos, n_tokens, 0);
cb(inp_pos, "inp_pos", -1); cb(inp_pos, "inp_pos", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
// shift the entire K-cache if needed // shift the entire K-cache if needed
if (do_rope_shift) { if (do_rope_shift) {
llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, LLM_ROPE_NEOX, n_ctx, freq_base, freq_scale, cb); llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, lctx.inp_K_shift, LLM_ROPE_NEOX, n_ctx, freq_base, freq_scale, cb);
} }
for (int il = 0; il < n_layer; ++il) { for (int il = 0; il < n_layer; ++il) {
@ -5840,20 +5851,20 @@ struct llm_build_context {
struct ggml_tensor * cur; struct ggml_tensor * cur;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// inp_pos - contains the positions // inp_pos - contains the positions
struct ggml_tensor * inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens); struct ggml_tensor * inp_pos = ggml_view_1d(ctx0, lctx.inp_pos, n_tokens, 0);
cb(inp_pos, "inp_pos", -1); cb(inp_pos, "inp_pos", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
// shift the entire K-cache if needed // shift the entire K-cache if needed
if (do_rope_shift) { if (do_rope_shift) {
llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, LLM_ROPE_NEOX, n_ctx, freq_base, freq_scale, cb); llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, lctx.inp_K_shift, LLM_ROPE_NEOX, n_ctx, freq_base, freq_scale, cb);
} }
for (int il = 0; il < n_layer; ++il) { for (int il = 0; il < n_layer; ++il) {
@ -5961,20 +5972,20 @@ struct llm_build_context {
struct ggml_tensor * ffn_output; struct ggml_tensor * ffn_output;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// inp_pos - contains the positions // inp_pos - contains the positions
struct ggml_tensor * inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens); struct ggml_tensor * inp_pos = ggml_view_1d(ctx0, lctx.inp_pos, n_tokens, 0);
cb(inp_pos, "inp_pos", -1); cb(inp_pos, "inp_pos", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
// shift the entire K-cache if needed // shift the entire K-cache if needed
if (do_rope_shift) { if (do_rope_shift) {
llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, LLM_ROPE_NEOX, n_ctx, freq_base, freq_scale, cb); llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, lctx.inp_K_shift, LLM_ROPE_NEOX, n_ctx, freq_base, freq_scale, cb);
} }
for (int il = 0; il < n_layer; ++il) { for (int il = 0; il < n_layer; ++il) {
@ -6083,20 +6094,20 @@ struct llm_build_context {
struct ggml_tensor * cur; struct ggml_tensor * cur;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// inp_pos - contains the positions // inp_pos - contains the positions
struct ggml_tensor * inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens); struct ggml_tensor * inp_pos = ggml_view_1d(ctx0, lctx.inp_pos, n_tokens, 0);
cb(inp_pos, "inp_pos", -1); cb(inp_pos, "inp_pos", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
// shift the entire K-cache if needed // shift the entire K-cache if needed
if (do_rope_shift) { if (do_rope_shift) {
llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, LLM_ROPE, n_ctx, freq_base, freq_scale, cb); llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, lctx.inp_K_shift, LLM_ROPE, n_ctx, freq_base, freq_scale, cb);
} }
for (int il = 0; il < n_layer; ++il) { for (int il = 0; il < n_layer; ++il) {
@ -6190,15 +6201,15 @@ struct llm_build_context {
struct ggml_tensor * pos; struct ggml_tensor * pos;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// inp_pos - contains the positions // inp_pos - contains the positions
struct ggml_tensor * inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens); struct ggml_tensor * inp_pos = ggml_view_1d(ctx0, lctx.inp_pos, n_tokens, 0);
cb(inp_pos, "inp_pos", -1); cb(inp_pos, "inp_pos", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
pos = ggml_get_rows(ctx0, model.pos_embd, inp_pos); pos = ggml_get_rows(ctx0, model.pos_embd, inp_pos);
@ -6288,20 +6299,20 @@ struct llm_build_context {
struct ggml_tensor * cur; struct ggml_tensor * cur;
struct ggml_tensor * inpL; struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, cb); inpL = llm_build_inp_embd(ctx0, hparams, batch, model.tok_embd, lctx.inp_tokens, lctx.inp_embd, cb);
cb(inpL, "inp_embd", -1); cb(inpL, "inp_embd", -1);
// inp_pos - contains the positions // inp_pos - contains the positions
struct ggml_tensor * inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens); struct ggml_tensor * inp_pos = ggml_view_1d(ctx0, lctx.inp_pos, n_tokens, 0);
cb(inp_pos, "inp_pos", -1); cb(inp_pos, "inp_pos", -1);
// KQ_mask (mask for 1 head, it will be broadcasted to all heads) // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); struct ggml_tensor * KQ_mask = ggml_view_2d(ctx0, lctx.inp_KQ_mask, n_kv, n_tokens, n_kv*ggml_type_size(lctx.inp_KQ_mask->type), 0);
cb(KQ_mask, "KQ_mask", -1); cb(KQ_mask, "KQ_mask", -1);
// shift the entire K-cache if needed // shift the entire K-cache if needed
if (do_rope_shift) { if (do_rope_shift) {
llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, LLM_ROPE, n_ctx, freq_base, freq_scale, cb); llm_build_k_shift(ctx0, hparams, cparams, kv_self, gf, lctx.inp_K_shift, LLM_ROPE, n_ctx, freq_base, freq_scale, cb);
} }
for (int il = 0; il < n_layer; ++il) { for (int il = 0; il < n_layer; ++il) {
@ -6395,15 +6406,7 @@ static struct ggml_cgraph * llama_build_graph(
// check if we should build the worst-case graph (for memory measurement) // check if we should build the worst-case graph (for memory measurement)
const bool worst_case = ggml_tallocr_is_measure(lctx.alloc); const bool worst_case = ggml_tallocr_is_measure(lctx.alloc);
// keep track of the input that has already been allocated
bool alloc_inp_tokens = false;
bool alloc_inp_embd = false;
bool alloc_inp_pos = false;
bool alloc_inp_KQ_mask = false;
bool alloc_inp_K_shift = false;
// this callback allows us to apply custom logic to each tensor (e.g. ggml-alloc, offloading, etc.) // this callback allows us to apply custom logic to each tensor (e.g. ggml-alloc, offloading, etc.)
// TODO: improve handling of input and output tensors, then replace this with ggml_set_name
llm_build_cb cb = [&](struct ggml_tensor * cur, const char * name, int il) { llm_build_cb cb = [&](struct ggml_tensor * cur, const char * name, int il) {
if (il >= 0) { if (il >= 0) {
ggml_format_name(cur, "%s-%d", name, il); ggml_format_name(cur, "%s-%d", name, il);
@ -6411,127 +6414,79 @@ static struct ggml_cgraph * llama_build_graph(
ggml_set_name(cur, name); ggml_set_name(cur, name);
} }
if (!lctx.cparams.offload_kqv) { if (!lctx.cparams.offload_kqv) {
if (strcmp(name, "kqv_merged_cont") == 0) { if (strcmp(name, "kqv_merged_cont") == 0) {
// all nodes between the KV store and the attention output are run on the CPU // all nodes between the KV store and the attention output are run on the CPU
ggml_backend_sched_set_node_backend(lctx.sched, cur, lctx.backend_cpu); ggml_backend_sched_set_node_backend(lctx.sched, cur, lctx.backend_cpu);
} }
} }
//
// allocate input tensors and set input data
//
if (!alloc_inp_tokens && strcmp(name, "inp_tokens") == 0) {
ggml_tallocr_alloc(lctx.alloc, cur);
if (!ggml_tallocr_is_measure(lctx.alloc) && batch.token) {
const int64_t n_tokens = cur->ne[0];
ggml_backend_tensor_set(cur, batch.token, 0, n_tokens*ggml_element_size(cur));
}
alloc_inp_tokens = true;
}
if (!alloc_inp_embd && strcmp(name, "inp_embd") == 0 && batch.embd) {
ggml_tallocr_alloc(lctx.alloc, cur);
if (!ggml_tallocr_is_measure(lctx.alloc) && batch.embd) {
const int64_t n_embd = cur->ne[0];
const int64_t n_tokens = cur->ne[1];
ggml_backend_tensor_set(cur, batch.embd, 0, n_tokens*n_embd*ggml_element_size(cur));
}
alloc_inp_embd = true;
}
if (!alloc_inp_pos && strcmp(name, "inp_pos") == 0) {
ggml_tallocr_alloc(lctx.alloc, cur);
if (!ggml_tallocr_is_measure(lctx.alloc) && batch.pos) {
const int64_t n_tokens = cur->ne[0];
static_assert(std::is_same<llama_pos, int32_t>::value, "llama_pos must be int32_t");
ggml_backend_tensor_set(cur, batch.pos, 0, n_tokens*ggml_element_size(cur));
}
alloc_inp_pos = true;
}
if (!alloc_inp_KQ_mask && strcmp(name, "KQ_mask") == 0) {
ggml_tallocr_alloc(lctx.alloc, cur);
if (!ggml_tallocr_is_measure(lctx.alloc)) {
const int64_t n_kv = cur->ne[0];
const int64_t n_tokens = cur->ne[1];
float * data;
if (ggml_backend_buffer_is_host(cur->buffer)) {
data = (float *) cur->data;
} else {
lctx.buf_copy.resize(ggml_nbytes(cur));
data = (float *) lctx.buf_copy.data();
}
for (int h = 0; h < 1; ++h) {
for (int j = 0; j < n_tokens; ++j) {
const llama_pos pos = batch.pos[j];
const llama_seq_id seq_id = batch.seq_id[j][0];
for (int i = 0; i < n_kv; ++i) {
float f;
if (!lctx.kv_self.cells[i].has_seq_id(seq_id) || lctx.kv_self.cells[i].pos > pos) {
f = -INFINITY;
} else {
f = 0;
}
data[h*(n_kv*n_tokens) + j*n_kv + i] = f;
}
}
}
if (data != cur->data) {
ggml_backend_tensor_set(cur, data, 0, ggml_nbytes(cur));
}
}
alloc_inp_KQ_mask = true;
}
if (!alloc_inp_K_shift && strcmp(name, "K_shift") == 0) {
ggml_tallocr_alloc(lctx.alloc, cur);
if (!ggml_tallocr_is_measure(lctx.alloc)) {
const int64_t n_ctx = cur->ne[0];
int32_t * data;
if (ggml_backend_buffer_is_host(cur->buffer)) {
data = (int32_t *) cur->data;
} else {
lctx.buf_copy.resize(ggml_nbytes(cur));
data = (int32_t *) lctx.buf_copy.data();
}
for (int i = 0; i < n_ctx; ++i) {
data[i] = lctx.kv_self.cells[i].delta;
}
if (data != cur->data) {
ggml_backend_tensor_set(cur, data, 0, ggml_nbytes(cur));
}
}
alloc_inp_K_shift = true;
}
}; };
struct ggml_cgraph * result = NULL; struct ggml_cgraph * result = NULL;
struct llm_build_context llm(lctx, batch, cb, worst_case); struct llm_build_context llm(lctx, batch, cb, worst_case);
//
// set input data
//
if (!ggml_tallocr_is_measure(lctx.alloc)) {
if (batch.token) {
const int64_t n_tokens = batch.n_tokens;
ggml_backend_tensor_set(lctx.inp_tokens, batch.token, 0, n_tokens*ggml_element_size(lctx.inp_tokens));
}
if (batch.embd) {
const int64_t n_embd = llm.n_embd;
const int64_t n_tokens = batch.n_tokens;
ggml_backend_tensor_set(lctx.inp_embd, batch.embd, 0, n_tokens*n_embd*ggml_element_size(lctx.inp_embd));
}
if (batch.pos) {
const int64_t n_tokens = batch.n_tokens;
ggml_backend_tensor_set(lctx.inp_pos, batch.pos, 0, n_tokens*ggml_element_size(lctx.inp_pos));
}
{
const int64_t n_kv = llm.n_kv;
const int64_t n_tokens = batch.n_tokens;
GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_KQ_mask->buffer));
float * data = (float *) lctx.inp_KQ_mask->data;
for (int h = 0; h < 1; ++h) {
for (int j = 0; j < n_tokens; ++j) {
const llama_pos pos = batch.pos[j];
const llama_seq_id seq_id = batch.seq_id[j][0];
for (int i = 0; i < n_kv; ++i) {
float f;
if (!lctx.kv_self.cells[i].has_seq_id(seq_id) || lctx.kv_self.cells[i].pos > pos) {
f = -INFINITY;
} else {
f = 0;
}
data[h*(n_kv*n_tokens) + j*n_kv + i] = f;
}
}
}
}
if (llm.do_rope_shift) {
const int64_t n_ctx = llm.n_ctx;
GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_K_shift->buffer));
int32_t * data = (int32_t *) lctx.inp_K_shift->data;
for (int i = 0; i < n_ctx; ++i) {
data[i] = lctx.kv_self.cells[i].delta;
}
}
}
llm.init(); llm.init();
switch (model.arch) { switch (model.arch) {
@ -10301,6 +10256,35 @@ struct llama_context * llama_new_context_with_model(
ctx->embedding.resize(hparams.n_embd); ctx->embedding.resize(hparams.n_embd);
} }
// graph inputs
{
ggml_init_params init_params = {
/* .mem_size */ ggml_tensor_overhead()*5,
/* .mem_buffer */ nullptr,
/* .no_alloc */ true,
};
ctx->ctx_input = ggml_init(init_params);
ctx->inp_tokens = ggml_new_tensor_1d(ctx->ctx_input, GGML_TYPE_I32, cparams.n_batch);
ctx->inp_embd = ggml_new_tensor_2d(ctx->ctx_input, GGML_TYPE_F32, hparams.n_embd, cparams.n_batch);
ctx->inp_pos = ggml_new_tensor_1d(ctx->ctx_input, GGML_TYPE_I32, cparams.n_batch);
ctx->inp_KQ_mask = ggml_new_tensor_2d(ctx->ctx_input, GGML_TYPE_F32, cparams.n_ctx, cparams.n_batch);
ctx->inp_K_shift = ggml_new_tensor_1d(ctx->ctx_input, GGML_TYPE_I32, cparams.n_ctx);
ggml_set_name(ctx->inp_tokens, "inp_tokens");
ggml_set_name(ctx->inp_embd, "inp_embd");
ggml_set_name(ctx->inp_pos, "inp_pos");
ggml_set_name(ctx->inp_KQ_mask, "inp_KQ_mask");
ggml_set_name(ctx->inp_K_shift, "inp_K_shift");
ctx->buf_input = ggml_backend_alloc_ctx_tensors_from_buft(ctx->ctx_input, llama_default_buffer_type_cpu(true));
LLAMA_LOG_INFO("%s: %10s input buffer size = %8.2f MiB\n", __func__,
ggml_backend_buffer_name(ctx->buf_input),
ggml_backend_buffer_get_size(ctx->buf_input) / 1024.0 / 1024.0);
}
// scheduler and compute buffers
{ {
// buffer types used for the compute buffer of each backend // buffer types used for the compute buffer of each backend
std::vector<ggml_backend_buffer_type_t> backend_buft; std::vector<ggml_backend_buffer_type_t> backend_buft;
@ -10327,9 +10311,6 @@ struct llama_context * llama_new_context_with_model(
// initialize scheduler with the worst-case graph // initialize scheduler with the worst-case graph
ggml_backend_sched_init_measure(ctx->sched, gf); ggml_backend_sched_init_measure(ctx->sched, gf);
// note: the number of splits during measure is higher than during inference due to the kv shift
int n_splits = ggml_backend_sched_get_n_splits(ctx->sched);
LLAMA_LOG_INFO("%s: graph splits (measure): %d\n", __func__, n_splits);
ctx->alloc = ggml_backend_sched_get_tallocr(ctx->sched, ctx->backend_cpu); ctx->alloc = ggml_backend_sched_get_tallocr(ctx->sched, ctx->backend_cpu);
for (ggml_backend_t backend : ctx->backends) { for (ggml_backend_t backend : ctx->backends) {
@ -10338,6 +10319,10 @@ struct llama_context * llama_new_context_with_model(
ggml_backend_buffer_name(buf), ggml_backend_buffer_name(buf),
ggml_backend_buffer_get_size(buf) / 1024.0 / 1024.0); ggml_backend_buffer_get_size(buf) / 1024.0 / 1024.0);
} }
// note: the number of splits during measure is higher than during inference due to the kv shift
int n_splits = ggml_backend_sched_get_n_splits(ctx->sched);
LLAMA_LOG_INFO("%s: graph splits (measure): %d\n", __func__, n_splits);
} }
} }